2023-02-18 04:23:24 +00:00
{
2023-03-08 03:30:51 +00:00
"accessibility" : {
2024-01-18 13:30:56 +00:00
"about" : "About" ,
2023-11-26 09:44:39 +00:00
"createIssue" : "Create Issue" ,
2024-05-22 15:05:18 +00:00
"submitSupportTicket" : "Submit Support Ticket" ,
2023-09-13 07:31:34 +00:00
"invokeProgressBar" : "Invoke progress bar" ,
"menu" : "Menu" ,
2023-11-20 06:24:03 +00:00
"mode" : "Mode" ,
2023-09-13 07:31:34 +00:00
"nextImage" : "Next Image" ,
"previousImage" : "Previous Image" ,
"reset" : "Reset" ,
2023-11-26 09:44:39 +00:00
"resetUI" : "$t(accessibility.reset) UI" ,
feat: queued generation (#4502)
* fix(config): fix typing issues in `config/`
`config/invokeai_config.py`:
- use `Optional` for things that are optional
- fix typing of `ram_cache_size()` and `vram_cache_size()`
- remove unused and incorrectly typed method `autoconvert_path`
- fix types and logic for `parse_args()`, in which `InvokeAIAppConfig.initconf` *must* be a `DictConfig`, but function would allow it to be set as a `ListConfig`, which presumably would cause issues elsewhere
`config/base.py`:
- use `cls` for first arg of class methods
- use `Optional` for things that are optional
- fix minor type issue related to setting of `env_prefix`
- remove unused `add_subparser()` method, which calls `add_parser()` on an `ArgumentParser` (method only available on the `_SubParsersAction` object, which is returned from ArgumentParser.add_subparsers()`)
* feat: queued generation and batches
Due to a very messy branch with broad addition of `isort` on `main` alongside it, some git surgery was needed to get an agreeable git history. This commit represents all of the work on queued generation. See PR for notes.
* chore: flake8, isort, black
* fix(nodes): fix incorrect service stop() method
* fix(nodes): improve names of a few variables
* fix(tests): fix up tests after changes to batches/queue
* feat(tests): add unit tests for session queue helper functions
* feat(ui): dynamic prompts is always enabled
* feat(queue): add queue_status_changed event
* feat(ui): wip queue graphs
* feat(nodes): move cleanup til after invoker startup
* feat(nodes): add cancel_by_batch_ids
* feat(ui): wip batch graphs & UI
* fix(nodes): remove `Batch.batch_id` from required
* fix(ui): cleanup and use fixedCacheKey for all mutations
* fix(ui): remove orphaned nodes from canvas graphs
* fix(nodes): fix cancel_by_batch_ids result count
* fix(ui): only show cancel batch tooltip when batches were canceled
* chore: isort
* fix(api): return `[""]` when dynamic prompts generates no prompts
Just a simple fallback so we always have a prompt.
* feat(ui): dynamicPrompts.combinatorial is always on
There seems to be little purpose in using the combinatorial generation for dynamic prompts. I've disabled it by hiding it from the UI and defaulting combinatorial to true. If we want to enable it again in the future it's straightforward to do so.
* feat: add queue_id & support logic
* feat(ui): fix upscale button
It prepends the upscale operation to queue
* feat(nodes): return queue item when enqueuing a single graph
This facilitates one-off graph async workflows in the client.
* feat(ui): move controlnet autoprocess to queue
* fix(ui): fix non-serializable DOMRect in redux state
* feat(ui): QueueTable performance tweaks
* feat(ui): update queue list
Queue items expand to show the full queue item. Just as JSON for now.
* wip threaded session_processor
* feat(nodes,ui): fully migrate queue to session_processor
* feat(nodes,ui): add processor events
* feat(ui): ui tweaks
* feat(nodes,ui): consolidate events, reduce network requests
* feat(ui): cleanup & abstract queue hooks
* feat(nodes): optimize batch permutation
Use a generator to do only as much work as is needed.
Previously, though we only ended up creating exactly as many queue items as was needed, there was still some intermediary work that calculated *all* permutations. When that number was very high, the system had a very hard time and used a lot of memory.
The logic has been refactored to use a generator. Additionally, the batch validators are optimized to return early and use less memory.
* feat(ui): add seed behaviour parameter
This dynamic prompts parameter allows the seed to be randomized per prompt or per iteration:
- Per iteration: Use the same seed for all prompts in a single dynamic prompt expansion
- Per prompt: Use a different seed for every single prompt
"Per iteration" is appropriate for exploring a the latents space with a stable starting noise, while "Per prompt" provides more variation.
* fix(ui): remove extraneous random seed nodes from linear graphs
* fix(ui): fix controlnet autoprocess not working when queue is running
* feat(queue): add timestamps to queue status updates
Also show execution time in queue list
* feat(queue): change all execution-related events to use the `queue_id` as the room, also include `queue_item_id` in InvocationQueueItem
This allows for much simpler handling of queue items.
* feat(api): deprecate sessions router
* chore(backend): tidy logging in `dependencies.py`
* fix(backend): respect `use_memory_db`
* feat(backend): add `config.log_sql` (enables sql trace logging)
* feat: add invocation cache
Supersedes #4574
The invocation cache provides simple node memoization functionality. Nodes that use the cache are memoized and not re-executed if their inputs haven't changed. Instead, the stored output is returned.
## Results
This feature provides anywhere some significant to massive performance improvement.
The improvement is most marked on large batches of generations where you only change a couple things (e.g. different seed or prompt for each iteration) and low-VRAM systems, where skipping an extraneous model load is a big deal.
## Overview
A new `invocation_cache` service is added to handle the caching. There's not much to it.
All nodes now inherit a boolean `use_cache` field from `BaseInvocation`. This is a node field and not a class attribute, because specific instances of nodes may want to opt in or out of caching.
The recently-added `invoke_internal()` method on `BaseInvocation` is used as an entrypoint for the cache logic.
To create a cache key, the invocation is first serialized using pydantic's provided `json()` method, skipping the unique `id` field. Then python's very fast builtin `hash()` is used to create an integer key. All implementations of `InvocationCacheBase` must provide a class method `create_key()` which accepts an invocation and outputs a string or integer key.
## In-Memory Implementation
An in-memory implementation is provided. In this implementation, the node outputs are stored in memory as python classes. The in-memory cache does not persist application restarts.
Max node cache size is added as `node_cache_size` under the `Generation` config category.
It defaults to 512 - this number is up for discussion, but given that these are relatively lightweight pydantic models, I think it's safe to up this even higher.
Note that the cache isn't storing the big stuff - tensors and images are store on disk, and outputs include only references to them.
## Node Definition
The default for all nodes is to use the cache. The `@invocation` decorator now accepts an optional `use_cache: bool` argument to override the default of `True`.
Non-deterministic nodes, however, should set this to `False`. Currently, all random-stuff nodes, including `dynamic_prompt`, are set to `False`.
The field name `use_cache` is now effectively a reserved field name and possibly a breaking change if any community nodes use this as a field name. In hindsight, all our reserved field names should have been prefixed with underscores or something.
## One Gotcha
Leaf nodes probably want to opt out of the cache, because if they are not cached, their outputs are not saved again.
If you run the same graph multiple times, you only end up with a single image output, because the image storage side-effects are in the `invoke()` method, which is bypassed if we have a cache hit.
## Linear UI
The linear graphs _almost_ just work, but due to the gotcha, we need to be careful about the final image-outputting node. To resolve this, a `SaveImageInvocation` node is added and used in the linear graphs.
This node is similar to `ImagePrimitive`, except it saves a copy of its input image, and has `use_cache` set to `False` by default.
This is now the leaf node in all linear graphs, and is the only node in those graphs with `use_cache == False` _and_ the only node with `is_intermedate == False`.
## Workflow Editor
All nodes now have a footer with a new `Use Cache [ ]` checkbox. It defaults to the value set by the invocation in its python definition, but can be changed by the user.
The workflow/node validation logic has been updated to migrate old workflows to use the new default values for `use_cache`. Users may still want to review the settings that have been chosen. In the event of catastrophic failure when running this migration, the default value of `True` is applied, as this is correct for most nodes.
Users should consider saving their workflows after loading them in and having them updated.
## Future Enhancements - Callback
A future enhancement would be to provide a callback to the `use_cache` flag that would be run as the node is executed to determine, based on its own internal state, if the cache should be used or not.
This would be useful for `DynamicPromptInvocation`, where the deterministic behaviour is determined by the `combinatorial: bool` field.
## Future Enhancements - Persisted Cache
Similar to how the latents storage is backed by disk, the invocation cache could be persisted to the database or disk. We'd need to be very careful about deserializing outputs, but it's perhaps worth exploring in the future.
* fix(ui): fix queue list item width
* feat(nodes): do not send the whole node on every generator progress
* feat(ui): strip out old logic related to sessions
Things like `isProcessing` are no longer relevant with queue. Removed them all & updated everything be appropriate for queue. May be a few little quirks I've missed...
* feat(ui): fix up param collapse labels
* feat(ui): click queue count to go to queue tab
* tidy(queue): update comment, query format
* feat(ui): fix progress bar when canceling
* fix(ui): fix circular dependency
* feat(nodes): bail on node caching logic if `node_cache_size == 0`
* feat(nodes): handle KeyError on node cache pop
* feat(nodes): bypass cache codepath if caches is disabled
more better no do thing
* fix(ui): reset api cache on connect/disconnect
* feat(ui): prevent enqueue when no prompts generated
* feat(ui): add queue controls to workflow editor
* feat(ui): update floating buttons & other incidental UI tweaks
* fix(ui): fix missing/incorrect translation keys
* fix(tests): add config service to mock invocation services
invoking needs access to `node_cache_size` to occur
* optionally remove pause/resume buttons from queue UI
* option to disable prepending
* chore(ui): remove unused file
* feat(queue): remove `order_id` entirely, `item_id` is now an autoinc pk
---------
Co-authored-by: Mary Hipp <maryhipp@Marys-MacBook-Air.local>
2023-09-20 05:09:24 +00:00
"showGalleryPanel" : "Show Gallery Panel" ,
2023-08-23 04:17:18 +00:00
"showOptionsPanel" : "Show Side Panel" ,
2023-09-13 07:31:34 +00:00
"uploadImage" : "Upload Image" ,
feat: queued generation (#4502)
* fix(config): fix typing issues in `config/`
`config/invokeai_config.py`:
- use `Optional` for things that are optional
- fix typing of `ram_cache_size()` and `vram_cache_size()`
- remove unused and incorrectly typed method `autoconvert_path`
- fix types and logic for `parse_args()`, in which `InvokeAIAppConfig.initconf` *must* be a `DictConfig`, but function would allow it to be set as a `ListConfig`, which presumably would cause issues elsewhere
`config/base.py`:
- use `cls` for first arg of class methods
- use `Optional` for things that are optional
- fix minor type issue related to setting of `env_prefix`
- remove unused `add_subparser()` method, which calls `add_parser()` on an `ArgumentParser` (method only available on the `_SubParsersAction` object, which is returned from ArgumentParser.add_subparsers()`)
* feat: queued generation and batches
Due to a very messy branch with broad addition of `isort` on `main` alongside it, some git surgery was needed to get an agreeable git history. This commit represents all of the work on queued generation. See PR for notes.
* chore: flake8, isort, black
* fix(nodes): fix incorrect service stop() method
* fix(nodes): improve names of a few variables
* fix(tests): fix up tests after changes to batches/queue
* feat(tests): add unit tests for session queue helper functions
* feat(ui): dynamic prompts is always enabled
* feat(queue): add queue_status_changed event
* feat(ui): wip queue graphs
* feat(nodes): move cleanup til after invoker startup
* feat(nodes): add cancel_by_batch_ids
* feat(ui): wip batch graphs & UI
* fix(nodes): remove `Batch.batch_id` from required
* fix(ui): cleanup and use fixedCacheKey for all mutations
* fix(ui): remove orphaned nodes from canvas graphs
* fix(nodes): fix cancel_by_batch_ids result count
* fix(ui): only show cancel batch tooltip when batches were canceled
* chore: isort
* fix(api): return `[""]` when dynamic prompts generates no prompts
Just a simple fallback so we always have a prompt.
* feat(ui): dynamicPrompts.combinatorial is always on
There seems to be little purpose in using the combinatorial generation for dynamic prompts. I've disabled it by hiding it from the UI and defaulting combinatorial to true. If we want to enable it again in the future it's straightforward to do so.
* feat: add queue_id & support logic
* feat(ui): fix upscale button
It prepends the upscale operation to queue
* feat(nodes): return queue item when enqueuing a single graph
This facilitates one-off graph async workflows in the client.
* feat(ui): move controlnet autoprocess to queue
* fix(ui): fix non-serializable DOMRect in redux state
* feat(ui): QueueTable performance tweaks
* feat(ui): update queue list
Queue items expand to show the full queue item. Just as JSON for now.
* wip threaded session_processor
* feat(nodes,ui): fully migrate queue to session_processor
* feat(nodes,ui): add processor events
* feat(ui): ui tweaks
* feat(nodes,ui): consolidate events, reduce network requests
* feat(ui): cleanup & abstract queue hooks
* feat(nodes): optimize batch permutation
Use a generator to do only as much work as is needed.
Previously, though we only ended up creating exactly as many queue items as was needed, there was still some intermediary work that calculated *all* permutations. When that number was very high, the system had a very hard time and used a lot of memory.
The logic has been refactored to use a generator. Additionally, the batch validators are optimized to return early and use less memory.
* feat(ui): add seed behaviour parameter
This dynamic prompts parameter allows the seed to be randomized per prompt or per iteration:
- Per iteration: Use the same seed for all prompts in a single dynamic prompt expansion
- Per prompt: Use a different seed for every single prompt
"Per iteration" is appropriate for exploring a the latents space with a stable starting noise, while "Per prompt" provides more variation.
* fix(ui): remove extraneous random seed nodes from linear graphs
* fix(ui): fix controlnet autoprocess not working when queue is running
* feat(queue): add timestamps to queue status updates
Also show execution time in queue list
* feat(queue): change all execution-related events to use the `queue_id` as the room, also include `queue_item_id` in InvocationQueueItem
This allows for much simpler handling of queue items.
* feat(api): deprecate sessions router
* chore(backend): tidy logging in `dependencies.py`
* fix(backend): respect `use_memory_db`
* feat(backend): add `config.log_sql` (enables sql trace logging)
* feat: add invocation cache
Supersedes #4574
The invocation cache provides simple node memoization functionality. Nodes that use the cache are memoized and not re-executed if their inputs haven't changed. Instead, the stored output is returned.
## Results
This feature provides anywhere some significant to massive performance improvement.
The improvement is most marked on large batches of generations where you only change a couple things (e.g. different seed or prompt for each iteration) and low-VRAM systems, where skipping an extraneous model load is a big deal.
## Overview
A new `invocation_cache` service is added to handle the caching. There's not much to it.
All nodes now inherit a boolean `use_cache` field from `BaseInvocation`. This is a node field and not a class attribute, because specific instances of nodes may want to opt in or out of caching.
The recently-added `invoke_internal()` method on `BaseInvocation` is used as an entrypoint for the cache logic.
To create a cache key, the invocation is first serialized using pydantic's provided `json()` method, skipping the unique `id` field. Then python's very fast builtin `hash()` is used to create an integer key. All implementations of `InvocationCacheBase` must provide a class method `create_key()` which accepts an invocation and outputs a string or integer key.
## In-Memory Implementation
An in-memory implementation is provided. In this implementation, the node outputs are stored in memory as python classes. The in-memory cache does not persist application restarts.
Max node cache size is added as `node_cache_size` under the `Generation` config category.
It defaults to 512 - this number is up for discussion, but given that these are relatively lightweight pydantic models, I think it's safe to up this even higher.
Note that the cache isn't storing the big stuff - tensors and images are store on disk, and outputs include only references to them.
## Node Definition
The default for all nodes is to use the cache. The `@invocation` decorator now accepts an optional `use_cache: bool` argument to override the default of `True`.
Non-deterministic nodes, however, should set this to `False`. Currently, all random-stuff nodes, including `dynamic_prompt`, are set to `False`.
The field name `use_cache` is now effectively a reserved field name and possibly a breaking change if any community nodes use this as a field name. In hindsight, all our reserved field names should have been prefixed with underscores or something.
## One Gotcha
Leaf nodes probably want to opt out of the cache, because if they are not cached, their outputs are not saved again.
If you run the same graph multiple times, you only end up with a single image output, because the image storage side-effects are in the `invoke()` method, which is bypassed if we have a cache hit.
## Linear UI
The linear graphs _almost_ just work, but due to the gotcha, we need to be careful about the final image-outputting node. To resolve this, a `SaveImageInvocation` node is added and used in the linear graphs.
This node is similar to `ImagePrimitive`, except it saves a copy of its input image, and has `use_cache` set to `False` by default.
This is now the leaf node in all linear graphs, and is the only node in those graphs with `use_cache == False` _and_ the only node with `is_intermedate == False`.
## Workflow Editor
All nodes now have a footer with a new `Use Cache [ ]` checkbox. It defaults to the value set by the invocation in its python definition, but can be changed by the user.
The workflow/node validation logic has been updated to migrate old workflows to use the new default values for `use_cache`. Users may still want to review the settings that have been chosen. In the event of catastrophic failure when running this migration, the default value of `True` is applied, as this is correct for most nodes.
Users should consider saving their workflows after loading them in and having them updated.
## Future Enhancements - Callback
A future enhancement would be to provide a callback to the `use_cache` flag that would be run as the node is executed to determine, based on its own internal state, if the cache should be used or not.
This would be useful for `DynamicPromptInvocation`, where the deterministic behaviour is determined by the `combinatorial: bool` field.
## Future Enhancements - Persisted Cache
Similar to how the latents storage is backed by disk, the invocation cache could be persisted to the database or disk. We'd need to be very careful about deserializing outputs, but it's perhaps worth exploring in the future.
* fix(ui): fix queue list item width
* feat(nodes): do not send the whole node on every generator progress
* feat(ui): strip out old logic related to sessions
Things like `isProcessing` are no longer relevant with queue. Removed them all & updated everything be appropriate for queue. May be a few little quirks I've missed...
* feat(ui): fix up param collapse labels
* feat(ui): click queue count to go to queue tab
* tidy(queue): update comment, query format
* feat(ui): fix progress bar when canceling
* fix(ui): fix circular dependency
* feat(nodes): bail on node caching logic if `node_cache_size == 0`
* feat(nodes): handle KeyError on node cache pop
* feat(nodes): bypass cache codepath if caches is disabled
more better no do thing
* fix(ui): reset api cache on connect/disconnect
* feat(ui): prevent enqueue when no prompts generated
* feat(ui): add queue controls to workflow editor
* feat(ui): update floating buttons & other incidental UI tweaks
* fix(ui): fix missing/incorrect translation keys
* fix(tests): add config service to mock invocation services
invoking needs access to `node_cache_size` to occur
* optionally remove pause/resume buttons from queue UI
* option to disable prepending
* chore(ui): remove unused file
* feat(queue): remove `order_id` entirely, `item_id` is now an autoinc pk
---------
Co-authored-by: Mary Hipp <maryhipp@Marys-MacBook-Air.local>
2023-09-20 05:09:24 +00:00
"loadMore" : "Load More"
2023-09-13 07:31:34 +00:00
} ,
"boards" : {
"addBoard" : "Add Board" ,
2024-07-09 12:51:08 +00:00
"addPrivateBoard" : "Add Private Board" ,
"addSharedBoard" : "Add Shared Board" ,
2024-06-26 18:04:10 +00:00
"archiveBoard" : "Archive Board" ,
"archived" : "Archived" ,
2023-09-13 07:31:34 +00:00
"autoAddBoard" : "Auto-Add Board" ,
2024-07-08 17:29:53 +00:00
"boards" : "Boards" ,
2024-07-01 16:48:15 +00:00
"selectedForAutoAdd" : "Selected for Auto-Add" ,
2023-09-13 07:31:34 +00:00
"bottomMessage" : "Deleting this board and its images will reset any features currently using them." ,
"cancel" : "Cancel" ,
"changeBoard" : "Change Board" ,
"clearSearch" : "Clear Search" ,
2023-11-20 06:24:03 +00:00
"deleteBoard" : "Delete Board" ,
"deleteBoardAndImages" : "Delete Board and Images" ,
"deleteBoardOnly" : "Delete Board Only" ,
2024-07-29 13:46:02 +00:00
"deletedBoardsCannotbeRestored" : "Deleted boards cannot be restored. Selecting 'Delete Board Only' will move images to an uncategorized state." ,
"deletedPrivateBoardsCannotbeRestored" : "Deleted boards cannot be restored. Selecting 'Delete Board Only' will move images to a private uncategorized state for the image's creator." ,
2024-07-23 22:51:48 +00:00
"hideBoards" : "Hide Boards" ,
2023-09-13 07:31:34 +00:00
"loading" : "Loading..." ,
"menuItemAutoAdd" : "Auto-add to this Board" ,
"move" : "Move" ,
2023-11-25 03:46:19 +00:00
"movingImagesToBoard_one" : "Moving {{count}} image to board:" ,
"movingImagesToBoard_other" : "Moving {{count}} images to board:" ,
2023-09-13 07:31:34 +00:00
"myBoard" : "My Board" ,
2024-07-23 16:28:31 +00:00
"noBoards" : "No {{boardType}} Boards" ,
2023-09-13 07:31:34 +00:00
"noMatching" : "No matching Boards" ,
2024-07-10 08:38:17 +00:00
"private" : "Private Boards" ,
2023-09-13 07:31:34 +00:00
"searchBoard" : "Search Boards..." ,
"selectBoard" : "Select a Board" ,
2024-07-10 08:38:17 +00:00
"shared" : "Shared Boards" ,
2023-09-13 07:31:34 +00:00
"topMessage" : "This board contains images used in the following features:" ,
2024-06-26 18:04:10 +00:00
"unarchiveBoard" : "Unarchive Board" ,
2023-10-11 23:27:22 +00:00
"uncategorized" : "Uncategorized" ,
2024-07-23 22:51:48 +00:00
"viewBoards" : "View Boards" ,
2024-06-21 08:50:50 +00:00
"downloadBoard" : "Download Board" ,
"imagesWithCount_one" : "{{count}} image" ,
"imagesWithCount_other" : "{{count}} images" ,
"assetsWithCount_one" : "{{count}} asset" ,
"assetsWithCount_other" : "{{count}} assets"
2023-03-08 03:30:51 +00:00
} ,
2023-12-28 13:03:21 +00:00
"accordions" : {
"generation" : {
2024-03-13 11:37:47 +00:00
"title" : "Generation"
2023-12-28 13:03:21 +00:00
} ,
"image" : {
"title" : "Image"
} ,
"advanced" : {
2024-03-22 17:53:32 +00:00
"title" : "Advanced" ,
"options" : "$t(accordions.advanced.title) Options"
2023-12-28 13:03:21 +00:00
} ,
"control" : {
2024-03-13 11:37:47 +00:00
"title" : "Control"
2023-12-28 13:03:21 +00:00
} ,
"compositing" : {
"title" : "Compositing" ,
"coherenceTab" : "Coherence Pass" ,
2023-12-29 00:59:21 +00:00
"infillTab" : "Infill"
2023-12-28 13:03:21 +00:00
}
} ,
2023-02-18 04:23:24 +00:00
"common" : {
2024-01-18 13:30:56 +00:00
"aboutDesc" : "Using Invoke for work? Check out:" ,
"aboutHeading" : "Own Your Creative Power" ,
2023-09-13 07:31:34 +00:00
"accept" : "Accept" ,
2024-03-04 14:39:03 +00:00
"add" : "Add" ,
2023-09-13 07:31:34 +00:00
"advanced" : "Advanced" ,
2023-11-25 03:46:19 +00:00
"ai" : "ai" ,
2023-09-13 07:31:34 +00:00
"areYouSure" : "Are you sure?" ,
2023-10-12 11:46:47 +00:00
"auto" : "Auto" ,
2023-09-13 07:31:34 +00:00
"back" : "Back" ,
"batch" : "Batch Manager" ,
2024-04-16 20:21:12 +00:00
"beta" : "Beta" ,
2023-09-13 07:31:34 +00:00
"cancel" : "Cancel" ,
2024-01-29 02:32:04 +00:00
"copy" : "Copy" ,
2023-11-26 09:44:39 +00:00
"copyError" : "$t(gallery.copy) Error" ,
2023-10-12 11:46:47 +00:00
"on" : "On" ,
2024-01-04 05:24:31 +00:00
"or" : "or" ,
2023-11-20 06:24:03 +00:00
"checkpoint" : "Checkpoint" ,
2023-07-26 05:33:02 +00:00
"communityLabel" : "Community" ,
2023-10-07 06:07:32 +00:00
"controlNet" : "ControlNet" ,
2023-11-20 06:24:03 +00:00
"data" : "Data" ,
feat: workflow library (#5148)
* chore: bump pydantic to 2.5.2
This release fixes pydantic/pydantic#8175 and allows us to use `JsonValue`
* fix(ui): exclude public/en.json from prettier config
* fix(workflow_records): fix SQLite workflow insertion to ignore duplicates
* feat(backend): update workflows handling
Update workflows handling for Workflow Library.
**Updated Workflow Storage**
"Embedded Workflows" are workflows associated with images, and are now only stored in the image files. "Library Workflows" are not associated with images, and are stored only in DB.
This works out nicely. We have always saved workflows to files, but recently began saving them to the DB in addition to in image files. When that happened, we stopped reading workflows from files, so all the workflows that only existed in images were inaccessible. With this change, access to those workflows is restored, and no workflows are lost.
**Updated Workflow Handling in Nodes**
Prior to this change, workflows were embedded in images by passing the whole workflow JSON to a special workflow field on a node. In the node's `invoke()` function, the node was able to access this workflow and save it with the image. This (inaccurately) models workflows as a property of an image and is rather awkward technically.
A workflow is now a property of a batch/session queue item. It is available in the InvocationContext and therefore available to all nodes during `invoke()`.
**Database Migrations**
Added a `SQLiteMigrator` class to handle database migrations. Migrations were needed to accomodate the DB-related changes in this PR. See the code for details.
The `images`, `workflows` and `session_queue` tables required migrations for this PR, and are using the new migrator. Other tables/services are still creating tables themselves. A followup PR will adapt them to use the migrator.
**Other/Support Changes**
- Add a `has_workflow` column to `images` table to indicate that the image has an embedded workflow.
- Add handling for retrieving the workflow from an image in python. The image file must be fetched, the workflow extracted, and then sent to client, avoiding needing the browser to parse the image file. With the `has_workflow` column, the UI knows if there is a workflow to be fetched, and only fetches when the user requests to load the workflow.
- Add route to get the workflow from an image
- Add CRUD service/routes for the library workflows
- `workflow_images` table and services removed (no longer needed now that embedded workflows are not in the DB)
* feat(ui): updated workflow handling (WIP)
Clientside updates for the backend workflow changes.
Includes roughed-out workflow library UI.
* feat: revert SQLiteMigrator class
Will pursue this in a separate PR.
* feat(nodes): do not overwrite custom node module names
Use a different, simpler method to detect if a node is custom.
* feat(nodes): restore WithWorkflow as no-op class
This class is deprecated and no longer needed. Set its workflow attr value to None (meaning it is now a no-op), and issue a warning when an invocation subclasses it.
* fix(nodes): fix get_workflow from queue item dict func
* feat(backend): add WorkflowRecordListItemDTO
This is the id, name, description, created at and updated at workflow columns/attrs. Used to display lists of workflowsl
* chore(ui): typegen
* feat(ui): add workflow loading, deleting to workflow library UI
* feat(ui): workflow library pagination button styles
* wip
* feat: workflow library WIP
- Save to library
- Duplicate
- Filter/sort
- UI/queries
* feat: workflow library - system graphs - wip
* feat(backend): sync system workflows to db
* fix: merge conflicts
* feat: simplify default workflows
- Rename "system" -> "default"
- Simplify syncing logic
- Update UI to match
* feat(workflows): update default workflows
- Update TextToImage_SD15
- Add TextToImage_SDXL
- Add README
* feat(ui): refine workflow list UI
* fix(workflow_records): typo
* fix(tests): fix tests
* feat(ui): clean up workflow library hooks
* fix(db): fix mis-ordered db cleanup step
It was happening before pruning queue items - should happen afterwards, else you have to restart the app again to free disk space made available by the pruning.
* feat(ui): tweak reset workflow editor translations
* feat(ui): split out workflow redux state
The `nodes` slice is a rather complicated slice. Removing `workflow` makes it a bit more reasonable.
Also helps to flatten state out a bit.
* docs: update default workflows README
* fix: tidy up unused files, unrelated changes
* fix(backend): revert unrelated service organisational changes
* feat(backend): workflow_records.get_many arg "filter_text" -> "query"
* feat(ui): use custom hook in current image buttons
Already in use elsewhere, forgot to use it here.
* fix(ui): remove commented out property
* fix(ui): fix workflow loading
- Different handling for loading from library vs external
- Fix bug where only nodes and edges loaded
* fix(ui): fix save/save-as workflow naming
* fix(ui): fix circular dependency
* fix(db): fix bug with releasing without lock in db.clean()
* fix(db): remove extraneous lock
* chore: bump ruff
* fix(workflow_records): default `category` to `WorkflowCategory.User`
This allows old workflows to validate when reading them from the db or image files.
* hide workflow library buttons if feature is disabled
---------
Co-authored-by: Mary Hipp <maryhipp@Marys-MacBook-Air.local>
2023-12-08 22:48:38 +00:00
"delete" : "Delete" ,
2023-11-20 06:24:03 +00:00
"details" : "Details" ,
feat: workflow library (#5148)
* chore: bump pydantic to 2.5.2
This release fixes pydantic/pydantic#8175 and allows us to use `JsonValue`
* fix(ui): exclude public/en.json from prettier config
* fix(workflow_records): fix SQLite workflow insertion to ignore duplicates
* feat(backend): update workflows handling
Update workflows handling for Workflow Library.
**Updated Workflow Storage**
"Embedded Workflows" are workflows associated with images, and are now only stored in the image files. "Library Workflows" are not associated with images, and are stored only in DB.
This works out nicely. We have always saved workflows to files, but recently began saving them to the DB in addition to in image files. When that happened, we stopped reading workflows from files, so all the workflows that only existed in images were inaccessible. With this change, access to those workflows is restored, and no workflows are lost.
**Updated Workflow Handling in Nodes**
Prior to this change, workflows were embedded in images by passing the whole workflow JSON to a special workflow field on a node. In the node's `invoke()` function, the node was able to access this workflow and save it with the image. This (inaccurately) models workflows as a property of an image and is rather awkward technically.
A workflow is now a property of a batch/session queue item. It is available in the InvocationContext and therefore available to all nodes during `invoke()`.
**Database Migrations**
Added a `SQLiteMigrator` class to handle database migrations. Migrations were needed to accomodate the DB-related changes in this PR. See the code for details.
The `images`, `workflows` and `session_queue` tables required migrations for this PR, and are using the new migrator. Other tables/services are still creating tables themselves. A followup PR will adapt them to use the migrator.
**Other/Support Changes**
- Add a `has_workflow` column to `images` table to indicate that the image has an embedded workflow.
- Add handling for retrieving the workflow from an image in python. The image file must be fetched, the workflow extracted, and then sent to client, avoiding needing the browser to parse the image file. With the `has_workflow` column, the UI knows if there is a workflow to be fetched, and only fetches when the user requests to load the workflow.
- Add route to get the workflow from an image
- Add CRUD service/routes for the library workflows
- `workflow_images` table and services removed (no longer needed now that embedded workflows are not in the DB)
* feat(ui): updated workflow handling (WIP)
Clientside updates for the backend workflow changes.
Includes roughed-out workflow library UI.
* feat: revert SQLiteMigrator class
Will pursue this in a separate PR.
* feat(nodes): do not overwrite custom node module names
Use a different, simpler method to detect if a node is custom.
* feat(nodes): restore WithWorkflow as no-op class
This class is deprecated and no longer needed. Set its workflow attr value to None (meaning it is now a no-op), and issue a warning when an invocation subclasses it.
* fix(nodes): fix get_workflow from queue item dict func
* feat(backend): add WorkflowRecordListItemDTO
This is the id, name, description, created at and updated at workflow columns/attrs. Used to display lists of workflowsl
* chore(ui): typegen
* feat(ui): add workflow loading, deleting to workflow library UI
* feat(ui): workflow library pagination button styles
* wip
* feat: workflow library WIP
- Save to library
- Duplicate
- Filter/sort
- UI/queries
* feat: workflow library - system graphs - wip
* feat(backend): sync system workflows to db
* fix: merge conflicts
* feat: simplify default workflows
- Rename "system" -> "default"
- Simplify syncing logic
- Update UI to match
* feat(workflows): update default workflows
- Update TextToImage_SD15
- Add TextToImage_SDXL
- Add README
* feat(ui): refine workflow list UI
* fix(workflow_records): typo
* fix(tests): fix tests
* feat(ui): clean up workflow library hooks
* fix(db): fix mis-ordered db cleanup step
It was happening before pruning queue items - should happen afterwards, else you have to restart the app again to free disk space made available by the pruning.
* feat(ui): tweak reset workflow editor translations
* feat(ui): split out workflow redux state
The `nodes` slice is a rather complicated slice. Removing `workflow` makes it a bit more reasonable.
Also helps to flatten state out a bit.
* docs: update default workflows README
* fix: tidy up unused files, unrelated changes
* fix(backend): revert unrelated service organisational changes
* feat(backend): workflow_records.get_many arg "filter_text" -> "query"
* feat(ui): use custom hook in current image buttons
Already in use elsewhere, forgot to use it here.
* fix(ui): remove commented out property
* fix(ui): fix workflow loading
- Different handling for loading from library vs external
- Fix bug where only nodes and edges loaded
* fix(ui): fix save/save-as workflow naming
* fix(ui): fix circular dependency
* fix(db): fix bug with releasing without lock in db.clean()
* fix(db): remove extraneous lock
* chore: bump ruff
* fix(workflow_records): default `category` to `WorkflowCategory.User`
This allows old workflows to validate when reading them from the db or image files.
* hide workflow library buttons if feature is disabled
---------
Co-authored-by: Mary Hipp <maryhipp@Marys-MacBook-Air.local>
2023-12-08 22:48:38 +00:00
"direction" : "Direction" ,
2023-09-16 00:50:30 +00:00
"ipAdapter" : "IP Adapter" ,
2023-10-07 06:07:32 +00:00
"t2iAdapter" : "T2I Adapter" ,
2024-04-22 12:12:29 +00:00
"positivePrompt" : "Positive Prompt" ,
"negativePrompt" : "Negative Prompt" ,
2023-02-18 04:23:24 +00:00
"discordLabel" : "Discord" ,
2023-09-13 07:31:34 +00:00
"dontAskMeAgain" : "Don't ask me again" ,
2024-07-25 05:33:39 +00:00
"dontShowMeThese" : "Don't show me these" ,
2024-05-03 02:13:35 +00:00
"editor" : "Editor" ,
2023-11-25 03:46:19 +00:00
"error" : "Error" ,
"file" : "File" ,
"folder" : "Folder" ,
2023-11-26 09:44:39 +00:00
"format" : "format" ,
2023-09-13 07:31:34 +00:00
"githubLabel" : "Github" ,
2024-05-02 21:04:14 +00:00
"goTo" : "Go to" ,
2023-09-13 07:31:34 +00:00
"hotkeysLabel" : "Hotkeys" ,
2023-09-26 15:24:53 +00:00
"imageFailedToLoad" : "Unable to Load Image" ,
2023-09-13 07:31:34 +00:00
"img2img" : "Image To Image" ,
2023-11-20 06:24:03 +00:00
"inpaint" : "inpaint" ,
2023-11-25 03:46:19 +00:00
"input" : "Input" ,
"installed" : "Installed" ,
2023-09-13 07:31:34 +00:00
"languagePickerLabel" : "Language" ,
Partial migration of UI to nodes API (#3195)
* feat(ui): add axios client generator and simple example
* fix(ui): update client & nodes test code w/ new Edge type
* chore(ui): organize generated files
* chore(ui): update .eslintignore, .prettierignore
* chore(ui): update openapi.json
* feat(backend): fixes for nodes/generator
* feat(ui): generate object args for api client
* feat(ui): more nodes api prototyping
* feat(ui): nodes cancel
* chore(ui): regenerate api client
* fix(ui): disable OG web server socket connection
* fix(ui): fix scrollbar styles typing and prop
just noticed the typo, and made the types stronger.
* feat(ui): add socketio types
* feat(ui): wip nodes
- extract api client method arg types instead of manually declaring them
- update example to display images
- general tidy up
* start building out node translations from frontend state and add notes about missing features
* use reference to sampler_name
* use reference to sampler_name
* add optional apiUrl prop
* feat(ui): start hooking up dynamic txt2img node generation, create middleware for session invocation
* feat(ui): write separate nodes socket layer, txt2img generating and rendering w single node
* feat(ui): img2img implementation
* feat(ui): get intermediate images working but types are stubbed out
* chore(ui): add support for package mode
* feat(ui): add nodes mode script
* feat(ui): handle random seeds
* fix(ui): fix middleware types
* feat(ui): add rtk action type guard
* feat(ui): disable NodeAPITest
This was polluting the network/socket logs.
* feat(ui): fix parameters panel border color
This commit should be elsewhere but I don't want to break my flow
* feat(ui): make thunk types more consistent
* feat(ui): add type guards for outputs
* feat(ui): load images on socket connect
Rudimentary
* chore(ui): bump redux-toolkit
* docs(ui): update readme
* chore(ui): regenerate api client
* chore(ui): add typescript as dev dependency
I am having trouble with TS versions after vscode updated and now uses TS 5. `madge` has installed 3.9.10 and for whatever reason my vscode wants to use that. Manually specifying 4.9.5 and then setting vscode to use that as the workspace TS fixes the issue.
* feat(ui): begin migrating gallery to nodes
Along the way, migrate to use RTK `createEntityAdapter` for gallery images, and separate `results` and `uploads` into separate slices. Much cleaner this way.
* feat(ui): clean up & comment results slice
* fix(ui): separate thunk for initial gallery load so it properly gets index 0
* feat(ui): POST upload working
* fix(ui): restore removed type
* feat(ui): patch api generation for headers access
* chore(ui): regenerate api
* feat(ui): wip gallery migration
* feat(ui): wip gallery migration
* chore(ui): regenerate api
* feat(ui): wip refactor socket events
* feat(ui): disable panels based on app props
* feat(ui): invert logic to be disabled
* disable panels when app mounts
* feat(ui): add support to disableTabs
* docs(ui): organise and update docs
* lang(ui): add toast strings
* feat(ui): wip events, comments, and general refactoring
* feat(ui): add optional token for auth
* feat(ui): export StatusIndicator and ModelSelect for header use
* feat(ui) working on making socket URL dynamic
* feat(ui): dynamic middleware loading
* feat(ui): prep for socket jwt
* feat(ui): migrate cancelation
also updated action names to be event-like instead of declaration-like
sorry, i was scattered and this commit has a lot of unrelated stuff in it.
* fix(ui): fix img2img type
* chore(ui): regenerate api client
* feat(ui): improve InvocationCompleteEvent types
* feat(ui): increase StatusIndicator font size
* fix(ui): fix middleware order for multi-node graphs
* feat(ui): add exampleGraphs object w/ iterations example
* feat(ui): generate iterations graph
* feat(ui): update ModelSelect for nodes API
* feat(ui): add hi-res functionality for txt2img generations
* feat(ui): "subscribe" to particular nodes
feels like a dirty hack but oh well it works
* feat(ui): first steps to node editor ui
* fix(ui): disable event subscription
it is not fully baked just yet
* feat(ui): wip node editor
* feat(ui): remove extraneous field types
* feat(ui): nodes before deleting stuff
* feat(ui): cleanup nodes ui stuff
* feat(ui): hook up nodes to redux
* fix(ui): fix handle
* fix(ui): add basic node edges & connection validation
* feat(ui): add connection validation styling
* feat(ui): increase edge width
* feat(ui): it blends
* feat(ui): wip model handling and graph topology validation
* feat(ui): validation connections w/ graphlib
* docs(ui): update nodes doc
* feat(ui): wip node editor
* chore(ui): rebuild api, update types
* add redux-dynamic-middlewares as a dependency
* feat(ui): add url host transformation
* feat(ui): handle already-connected fields
* feat(ui): rewrite SqliteItemStore in sqlalchemy
* fix(ui): fix sqlalchemy dynamic model instantiation
* feat(ui, nodes): metadata wip
* feat(ui, nodes): models
* feat(ui, nodes): more metadata wip
* feat(ui): wip range/iterate
* fix(nodes): fix sqlite typing
* feat(ui): export new type for invoke component
* tests(nodes): fix test instantiation of ImageField
* feat(nodes): fix LoadImageInvocation
* feat(nodes): add `title` ui hint
* feat(nodes): make ImageField attrs optional
* feat(ui): wip nodes etc
* feat(nodes): roll back sqlalchemy
* fix(nodes): partially address feedback
* fix(backend): roll back changes to pngwriter
* feat(nodes): wip address metadata feedback
* feat(nodes): add seeded rng to RandomRange
* feat(nodes): address feedback
* feat(nodes): move GET images error handling to DiskImageStorage
* feat(nodes): move GET images error handling to DiskImageStorage
* fix(nodes): fix image output schema customization
* feat(ui): img2img/txt2img -> linear
- remove txt2img and img2img tabs
- add linear tab
- add initial image selection to linear parameters accordion
* feat(ui): tidy graph builders
* feat(ui): tidy misc
* feat(ui): improve invocation union types
* feat(ui): wip metadata viewer recall
* feat(ui): move fonts to normal deps
* feat(nodes): fix broken upload
* feat(nodes): add metadata module + tests, thumbnails
- `MetadataModule` is stateless and needed in places where the `InvocationContext` is not available, so have not made it a `service`
- Handles loading/parsing/building metadata, and creating png info objects
- added tests for MetadataModule
- Lifted thumbnail stuff to util
* fix(nodes): revert change to RandomRangeInvocation
* feat(nodes): address feedback
- make metadata a service
- rip out pydantic validation, implement metadata parsing as simple functions
- update tests
- address other minor feedback items
* fix(nodes): fix other tests
* fix(nodes): add metadata service to cli
* fix(nodes): fix latents/image field parsing
* feat(nodes): customise LatentsField schema
* feat(nodes): move metadata parsing to frontend
* fix(nodes): fix metadata test
---------
Co-authored-by: maryhipp <maryhipp@gmail.com>
Co-authored-by: Mary Hipp <maryhipp@Marys-MacBook-Air.local>
2023-04-22 03:10:20 +00:00
"linear" : "Linear" ,
2023-09-13 07:31:34 +00:00
"load" : "Load" ,
2023-09-27 14:54:20 +00:00
"loading" : "Loading" ,
2024-01-18 13:30:56 +00:00
"localSystem" : "Local System" ,
2024-03-22 17:53:32 +00:00
"loglevel" : "Log Level" ,
feat(ui): refactor informational popover
- Change translations to use arrays of paragraphs instead of a single paragraph.
- Change component to accept a `feature` prop to identify the feature which the popover describes.
- Add optional `wrapperProps`: passed to the wrapper element, allowing more flexibility when using the popover
- Add optional `popoverProps`: passed to the `<Popover />` component, allowing for overriding individual instances of the popover's props
- Move definitions of features and popover settings to `invokeai/frontend/web/src/common/components/IAIInformationalPopover/constants.ts`
- Add some type safety to the `feature` prop
- Edit `POPOVER_DATA` to provide `image`, `href`, `buttonLabel`, and any popover props. The popover props are applied to all instances of the popover for the given feature. Note that the component prop `popoverProps` will override settings here.
- Remove the popover's arrow. Because the popover is wrapping groups of components, sometimes the error ends up pointing to nothing, which looks kinda janky. I've just removed the arrow entirely, but feel free to add it back if you think it looks better.
- Use a `link` variant button with external link icon to better communicate that clicking the button will open a new tab.
- Default the link button label to "Learn More" (if a label is provided, that will be used instead)
- Make default position `top`, but set manually set some to `right` - namely, anything with a dropdown. This prevents the popovers from obscuring or being obscured by the dropdowns.
- Do a bit more restructuring of the Popover component itself, and how it is integrated with other components
- More ref forwarding
- Make the open delay 1s
- Set the popovers to use lazy mounting (eg do not mount until the user opens the thing)
- Update the verbiage for many popover items and add missing dynamic prompts stuff
2023-09-22 11:04:35 +00:00
"learnMore" : "Learn More" ,
2023-07-08 00:19:51 +00:00
"modelManager" : "Model Manager" ,
2023-09-13 07:31:34 +00:00
"nodeEditor" : "Node Editor" ,
2024-03-14 00:42:32 +00:00
"nodes" : "Workflows" ,
2023-11-25 03:46:19 +00:00
"notInstalled" : "Not $t(common.installed)" ,
2023-09-13 07:31:34 +00:00
"openInNewTab" : "Open in New Tab" ,
feat: workflow library (#5148)
* chore: bump pydantic to 2.5.2
This release fixes pydantic/pydantic#8175 and allows us to use `JsonValue`
* fix(ui): exclude public/en.json from prettier config
* fix(workflow_records): fix SQLite workflow insertion to ignore duplicates
* feat(backend): update workflows handling
Update workflows handling for Workflow Library.
**Updated Workflow Storage**
"Embedded Workflows" are workflows associated with images, and are now only stored in the image files. "Library Workflows" are not associated with images, and are stored only in DB.
This works out nicely. We have always saved workflows to files, but recently began saving them to the DB in addition to in image files. When that happened, we stopped reading workflows from files, so all the workflows that only existed in images were inaccessible. With this change, access to those workflows is restored, and no workflows are lost.
**Updated Workflow Handling in Nodes**
Prior to this change, workflows were embedded in images by passing the whole workflow JSON to a special workflow field on a node. In the node's `invoke()` function, the node was able to access this workflow and save it with the image. This (inaccurately) models workflows as a property of an image and is rather awkward technically.
A workflow is now a property of a batch/session queue item. It is available in the InvocationContext and therefore available to all nodes during `invoke()`.
**Database Migrations**
Added a `SQLiteMigrator` class to handle database migrations. Migrations were needed to accomodate the DB-related changes in this PR. See the code for details.
The `images`, `workflows` and `session_queue` tables required migrations for this PR, and are using the new migrator. Other tables/services are still creating tables themselves. A followup PR will adapt them to use the migrator.
**Other/Support Changes**
- Add a `has_workflow` column to `images` table to indicate that the image has an embedded workflow.
- Add handling for retrieving the workflow from an image in python. The image file must be fetched, the workflow extracted, and then sent to client, avoiding needing the browser to parse the image file. With the `has_workflow` column, the UI knows if there is a workflow to be fetched, and only fetches when the user requests to load the workflow.
- Add route to get the workflow from an image
- Add CRUD service/routes for the library workflows
- `workflow_images` table and services removed (no longer needed now that embedded workflows are not in the DB)
* feat(ui): updated workflow handling (WIP)
Clientside updates for the backend workflow changes.
Includes roughed-out workflow library UI.
* feat: revert SQLiteMigrator class
Will pursue this in a separate PR.
* feat(nodes): do not overwrite custom node module names
Use a different, simpler method to detect if a node is custom.
* feat(nodes): restore WithWorkflow as no-op class
This class is deprecated and no longer needed. Set its workflow attr value to None (meaning it is now a no-op), and issue a warning when an invocation subclasses it.
* fix(nodes): fix get_workflow from queue item dict func
* feat(backend): add WorkflowRecordListItemDTO
This is the id, name, description, created at and updated at workflow columns/attrs. Used to display lists of workflowsl
* chore(ui): typegen
* feat(ui): add workflow loading, deleting to workflow library UI
* feat(ui): workflow library pagination button styles
* wip
* feat: workflow library WIP
- Save to library
- Duplicate
- Filter/sort
- UI/queries
* feat: workflow library - system graphs - wip
* feat(backend): sync system workflows to db
* fix: merge conflicts
* feat: simplify default workflows
- Rename "system" -> "default"
- Simplify syncing logic
- Update UI to match
* feat(workflows): update default workflows
- Update TextToImage_SD15
- Add TextToImage_SDXL
- Add README
* feat(ui): refine workflow list UI
* fix(workflow_records): typo
* fix(tests): fix tests
* feat(ui): clean up workflow library hooks
* fix(db): fix mis-ordered db cleanup step
It was happening before pruning queue items - should happen afterwards, else you have to restart the app again to free disk space made available by the pruning.
* feat(ui): tweak reset workflow editor translations
* feat(ui): split out workflow redux state
The `nodes` slice is a rather complicated slice. Removing `workflow` makes it a bit more reasonable.
Also helps to flatten state out a bit.
* docs: update default workflows README
* fix: tidy up unused files, unrelated changes
* fix(backend): revert unrelated service organisational changes
* feat(backend): workflow_records.get_many arg "filter_text" -> "query"
* feat(ui): use custom hook in current image buttons
Already in use elsewhere, forgot to use it here.
* fix(ui): remove commented out property
* fix(ui): fix workflow loading
- Different handling for loading from library vs external
- Fix bug where only nodes and edges loaded
* fix(ui): fix save/save-as workflow naming
* fix(ui): fix circular dependency
* fix(db): fix bug with releasing without lock in db.clean()
* fix(db): remove extraneous lock
* chore: bump ruff
* fix(workflow_records): default `category` to `WorkflowCategory.User`
This allows old workflows to validate when reading them from the db or image files.
* hide workflow library buttons if feature is disabled
---------
Co-authored-by: Mary Hipp <maryhipp@Marys-MacBook-Air.local>
2023-12-08 22:48:38 +00:00
"orderBy" : "Order By" ,
2023-11-20 06:24:03 +00:00
"outpaint" : "outpaint" ,
"outputs" : "Outputs" ,
2023-09-13 07:31:34 +00:00
"postprocessing" : "Post Processing" ,
"random" : "Random" ,
"reportBugLabel" : "Report Bug" ,
2023-11-20 06:24:03 +00:00
"safetensors" : "Safetensors" ,
feat: workflow library (#5148)
* chore: bump pydantic to 2.5.2
This release fixes pydantic/pydantic#8175 and allows us to use `JsonValue`
* fix(ui): exclude public/en.json from prettier config
* fix(workflow_records): fix SQLite workflow insertion to ignore duplicates
* feat(backend): update workflows handling
Update workflows handling for Workflow Library.
**Updated Workflow Storage**
"Embedded Workflows" are workflows associated with images, and are now only stored in the image files. "Library Workflows" are not associated with images, and are stored only in DB.
This works out nicely. We have always saved workflows to files, but recently began saving them to the DB in addition to in image files. When that happened, we stopped reading workflows from files, so all the workflows that only existed in images were inaccessible. With this change, access to those workflows is restored, and no workflows are lost.
**Updated Workflow Handling in Nodes**
Prior to this change, workflows were embedded in images by passing the whole workflow JSON to a special workflow field on a node. In the node's `invoke()` function, the node was able to access this workflow and save it with the image. This (inaccurately) models workflows as a property of an image and is rather awkward technically.
A workflow is now a property of a batch/session queue item. It is available in the InvocationContext and therefore available to all nodes during `invoke()`.
**Database Migrations**
Added a `SQLiteMigrator` class to handle database migrations. Migrations were needed to accomodate the DB-related changes in this PR. See the code for details.
The `images`, `workflows` and `session_queue` tables required migrations for this PR, and are using the new migrator. Other tables/services are still creating tables themselves. A followup PR will adapt them to use the migrator.
**Other/Support Changes**
- Add a `has_workflow` column to `images` table to indicate that the image has an embedded workflow.
- Add handling for retrieving the workflow from an image in python. The image file must be fetched, the workflow extracted, and then sent to client, avoiding needing the browser to parse the image file. With the `has_workflow` column, the UI knows if there is a workflow to be fetched, and only fetches when the user requests to load the workflow.
- Add route to get the workflow from an image
- Add CRUD service/routes for the library workflows
- `workflow_images` table and services removed (no longer needed now that embedded workflows are not in the DB)
* feat(ui): updated workflow handling (WIP)
Clientside updates for the backend workflow changes.
Includes roughed-out workflow library UI.
* feat: revert SQLiteMigrator class
Will pursue this in a separate PR.
* feat(nodes): do not overwrite custom node module names
Use a different, simpler method to detect if a node is custom.
* feat(nodes): restore WithWorkflow as no-op class
This class is deprecated and no longer needed. Set its workflow attr value to None (meaning it is now a no-op), and issue a warning when an invocation subclasses it.
* fix(nodes): fix get_workflow from queue item dict func
* feat(backend): add WorkflowRecordListItemDTO
This is the id, name, description, created at and updated at workflow columns/attrs. Used to display lists of workflowsl
* chore(ui): typegen
* feat(ui): add workflow loading, deleting to workflow library UI
* feat(ui): workflow library pagination button styles
* wip
* feat: workflow library WIP
- Save to library
- Duplicate
- Filter/sort
- UI/queries
* feat: workflow library - system graphs - wip
* feat(backend): sync system workflows to db
* fix: merge conflicts
* feat: simplify default workflows
- Rename "system" -> "default"
- Simplify syncing logic
- Update UI to match
* feat(workflows): update default workflows
- Update TextToImage_SD15
- Add TextToImage_SDXL
- Add README
* feat(ui): refine workflow list UI
* fix(workflow_records): typo
* fix(tests): fix tests
* feat(ui): clean up workflow library hooks
* fix(db): fix mis-ordered db cleanup step
It was happening before pruning queue items - should happen afterwards, else you have to restart the app again to free disk space made available by the pruning.
* feat(ui): tweak reset workflow editor translations
* feat(ui): split out workflow redux state
The `nodes` slice is a rather complicated slice. Removing `workflow` makes it a bit more reasonable.
Also helps to flatten state out a bit.
* docs: update default workflows README
* fix: tidy up unused files, unrelated changes
* fix(backend): revert unrelated service organisational changes
* feat(backend): workflow_records.get_many arg "filter_text" -> "query"
* feat(ui): use custom hook in current image buttons
Already in use elsewhere, forgot to use it here.
* fix(ui): remove commented out property
* fix(ui): fix workflow loading
- Different handling for loading from library vs external
- Fix bug where only nodes and edges loaded
* fix(ui): fix save/save-as workflow naming
* fix(ui): fix circular dependency
* fix(db): fix bug with releasing without lock in db.clean()
* fix(db): remove extraneous lock
* chore: bump ruff
* fix(workflow_records): default `category` to `WorkflowCategory.User`
This allows old workflows to validate when reading them from the db or image files.
* hide workflow library buttons if feature is disabled
---------
Co-authored-by: Mary Hipp <maryhipp@Marys-MacBook-Air.local>
2023-12-08 22:48:38 +00:00
"save" : "Save" ,
"saveAs" : "Save As" ,
2023-09-13 07:31:34 +00:00
"settingsLabel" : "Settings" ,
2023-11-20 06:24:03 +00:00
"simple" : "Simple" ,
2023-11-25 03:46:19 +00:00
"somethingWentWrong" : "Something went wrong" ,
2023-02-18 04:23:24 +00:00
"statusDisconnected" : "Disconnected" ,
2023-11-20 06:24:03 +00:00
"template" : "Template" ,
2024-02-14 14:02:07 +00:00
"toResolve" : "To resolve" ,
2023-09-13 07:31:34 +00:00
"txt2img" : "Text To Image" ,
"unifiedCanvas" : "Unified Canvas" ,
2023-11-27 01:51:19 +00:00
"unknown" : "Unknown" ,
feat: workflow library (#5148)
* chore: bump pydantic to 2.5.2
This release fixes pydantic/pydantic#8175 and allows us to use `JsonValue`
* fix(ui): exclude public/en.json from prettier config
* fix(workflow_records): fix SQLite workflow insertion to ignore duplicates
* feat(backend): update workflows handling
Update workflows handling for Workflow Library.
**Updated Workflow Storage**
"Embedded Workflows" are workflows associated with images, and are now only stored in the image files. "Library Workflows" are not associated with images, and are stored only in DB.
This works out nicely. We have always saved workflows to files, but recently began saving them to the DB in addition to in image files. When that happened, we stopped reading workflows from files, so all the workflows that only existed in images were inaccessible. With this change, access to those workflows is restored, and no workflows are lost.
**Updated Workflow Handling in Nodes**
Prior to this change, workflows were embedded in images by passing the whole workflow JSON to a special workflow field on a node. In the node's `invoke()` function, the node was able to access this workflow and save it with the image. This (inaccurately) models workflows as a property of an image and is rather awkward technically.
A workflow is now a property of a batch/session queue item. It is available in the InvocationContext and therefore available to all nodes during `invoke()`.
**Database Migrations**
Added a `SQLiteMigrator` class to handle database migrations. Migrations were needed to accomodate the DB-related changes in this PR. See the code for details.
The `images`, `workflows` and `session_queue` tables required migrations for this PR, and are using the new migrator. Other tables/services are still creating tables themselves. A followup PR will adapt them to use the migrator.
**Other/Support Changes**
- Add a `has_workflow` column to `images` table to indicate that the image has an embedded workflow.
- Add handling for retrieving the workflow from an image in python. The image file must be fetched, the workflow extracted, and then sent to client, avoiding needing the browser to parse the image file. With the `has_workflow` column, the UI knows if there is a workflow to be fetched, and only fetches when the user requests to load the workflow.
- Add route to get the workflow from an image
- Add CRUD service/routes for the library workflows
- `workflow_images` table and services removed (no longer needed now that embedded workflows are not in the DB)
* feat(ui): updated workflow handling (WIP)
Clientside updates for the backend workflow changes.
Includes roughed-out workflow library UI.
* feat: revert SQLiteMigrator class
Will pursue this in a separate PR.
* feat(nodes): do not overwrite custom node module names
Use a different, simpler method to detect if a node is custom.
* feat(nodes): restore WithWorkflow as no-op class
This class is deprecated and no longer needed. Set its workflow attr value to None (meaning it is now a no-op), and issue a warning when an invocation subclasses it.
* fix(nodes): fix get_workflow from queue item dict func
* feat(backend): add WorkflowRecordListItemDTO
This is the id, name, description, created at and updated at workflow columns/attrs. Used to display lists of workflowsl
* chore(ui): typegen
* feat(ui): add workflow loading, deleting to workflow library UI
* feat(ui): workflow library pagination button styles
* wip
* feat: workflow library WIP
- Save to library
- Duplicate
- Filter/sort
- UI/queries
* feat: workflow library - system graphs - wip
* feat(backend): sync system workflows to db
* fix: merge conflicts
* feat: simplify default workflows
- Rename "system" -> "default"
- Simplify syncing logic
- Update UI to match
* feat(workflows): update default workflows
- Update TextToImage_SD15
- Add TextToImage_SDXL
- Add README
* feat(ui): refine workflow list UI
* fix(workflow_records): typo
* fix(tests): fix tests
* feat(ui): clean up workflow library hooks
* fix(db): fix mis-ordered db cleanup step
It was happening before pruning queue items - should happen afterwards, else you have to restart the app again to free disk space made available by the pruning.
* feat(ui): tweak reset workflow editor translations
* feat(ui): split out workflow redux state
The `nodes` slice is a rather complicated slice. Removing `workflow` makes it a bit more reasonable.
Also helps to flatten state out a bit.
* docs: update default workflows README
* fix: tidy up unused files, unrelated changes
* fix(backend): revert unrelated service organisational changes
* feat(backend): workflow_records.get_many arg "filter_text" -> "query"
* feat(ui): use custom hook in current image buttons
Already in use elsewhere, forgot to use it here.
* fix(ui): remove commented out property
* fix(ui): fix workflow loading
- Different handling for loading from library vs external
- Fix bug where only nodes and edges loaded
* fix(ui): fix save/save-as workflow naming
* fix(ui): fix circular dependency
* fix(db): fix bug with releasing without lock in db.clean()
* fix(db): remove extraneous lock
* chore: bump ruff
* fix(workflow_records): default `category` to `WorkflowCategory.User`
This allows old workflows to validate when reading them from the db or image files.
* hide workflow library buttons if feature is disabled
---------
Co-authored-by: Mary Hipp <maryhipp@Marys-MacBook-Air.local>
2023-12-08 22:48:38 +00:00
"upload" : "Upload" ,
"updated" : "Updated" ,
"created" : "Created" ,
"prevPage" : "Previous Page" ,
"nextPage" : "Next Page" ,
2023-12-06 12:44:00 +00:00
"unknownError" : "Unknown Error" ,
2024-01-20 05:28:22 +00:00
"red" : "Red" ,
"green" : "Green" ,
"blue" : "Blue" ,
2024-04-12 08:46:31 +00:00
"alpha" : "Alpha" ,
2024-04-19 12:30:20 +00:00
"selected" : "Selected" ,
2024-05-06 08:29:13 +00:00
"tab" : "Tab" ,
2024-05-08 07:38:18 +00:00
"viewing" : "Viewing" ,
"viewingDesc" : "Review images in a large gallery view" ,
"editing" : "Editing" ,
2024-05-22 21:40:27 +00:00
"editingDesc" : "Edit on the Control Layers canvas" ,
2024-05-31 07:20:13 +00:00
"comparing" : "Comparing" ,
"comparingDesc" : "Comparing two images" ,
2024-05-22 21:40:27 +00:00
"enabled" : "Enabled" ,
"disabled" : "Disabled"
2023-09-13 07:31:34 +00:00
} ,
"controlnet" : {
2023-10-12 11:46:47 +00:00
"controlAdapter_one" : "Control Adapter" ,
"controlAdapter_other" : "Control Adapters" ,
2023-10-15 07:16:10 +00:00
"controlnet" : "$t(controlnet.controlAdapter_one) #{{number}} ($t(common.controlNet))" ,
"ip_adapter" : "$t(controlnet.controlAdapter_one) #{{number}} ($t(common.ipAdapter))" ,
"t2i_adapter" : "$t(controlnet.controlAdapter_one) #{{number}} ($t(common.t2iAdapter))" ,
2023-10-07 06:07:32 +00:00
"addControlNet" : "Add $t(common.controlNet)" ,
"addIPAdapter" : "Add $t(common.ipAdapter)" ,
"addT2IAdapter" : "Add $t(common.t2iAdapter)" ,
2023-09-13 07:31:34 +00:00
"amult" : "a_mult" ,
"autoConfigure" : "Auto configure processor" ,
"balanced" : "Balanced" ,
2024-01-22 22:28:24 +00:00
"base" : "Base" ,
2023-09-13 07:31:34 +00:00
"beginEndStepPercent" : "Begin / End Step Percentage" ,
2024-04-29 13:31:33 +00:00
"beginEndStepPercentShort" : "Begin/End %" ,
2023-09-13 07:31:34 +00:00
"bgth" : "bg_th" ,
"canny" : "Canny" ,
"cannyDescription" : "Canny edge detection" ,
2023-09-22 20:49:15 +00:00
"colorMap" : "Color" ,
"colorMapDescription" : "Generates a color map from the image" ,
2023-09-13 07:31:34 +00:00
"coarse" : "Coarse" ,
"contentShuffle" : "Content Shuffle" ,
"contentShuffleDescription" : "Shuffles the content in an image" ,
"control" : "Control" ,
"controlMode" : "Control Mode" ,
"crop" : "Crop" ,
"delete" : "Delete" ,
2024-01-22 22:28:24 +00:00
"depthAnything" : "Depth Anything" ,
"depthAnythingDescription" : "Depth map generation using the Depth Anything technique" ,
2024-07-31 17:59:43 +00:00
"depthAnythingSmallV2" : "Small V2" ,
2023-09-13 07:31:34 +00:00
"depthMidas" : "Depth (Midas)" ,
"depthMidasDescription" : "Depth map generation using Midas" ,
"depthZoe" : "Depth (Zoe)" ,
"depthZoeDescription" : "Depth map generation using Zoe" ,
"detectResolution" : "Detect Resolution" ,
"duplicate" : "Duplicate" ,
"f" : "F" ,
"fill" : "Fill" ,
"h" : "H" ,
2024-02-09 20:14:15 +00:00
"face" : "Face" ,
"body" : "Body" ,
"hands" : "Hands" ,
2023-09-13 07:31:34 +00:00
"hed" : "HED" ,
"hedDescription" : "Holistically-Nested Edge Detection" ,
"hideAdvanced" : "Hide Advanced" ,
"highThreshold" : "High Threshold" ,
"imageResolution" : "Image Resolution" ,
2023-09-22 20:49:15 +00:00
"colorMapTileSize" : "Tile Size" ,
2023-09-13 07:31:34 +00:00
"importImageFromCanvas" : "Import Image From Canvas" ,
"importMaskFromCanvas" : "Import Mask From Canvas" ,
2024-01-22 22:28:24 +00:00
"large" : "Large" ,
2023-09-13 07:31:34 +00:00
"lineart" : "Lineart" ,
"lineartAnime" : "Lineart Anime" ,
"lineartAnimeDescription" : "Anime-style lineart processing" ,
"lineartDescription" : "Converts image to lineart" ,
"lowThreshold" : "Low Threshold" ,
"maxFaces" : "Max Faces" ,
"mediapipeFace" : "Mediapipe Face" ,
"mediapipeFaceDescription" : "Face detection using Mediapipe" ,
"megaControl" : "Mega Control" ,
"minConfidence" : "Min Confidence" ,
"mlsd" : "M-LSD" ,
"mlsdDescription" : "Minimalist Line Segment Detector" ,
2024-01-22 22:28:24 +00:00
"modelSize" : "Model Size" ,
2023-09-13 07:31:34 +00:00
"none" : "None" ,
"noneDescription" : "No processing applied" ,
"normalBae" : "Normal BAE" ,
"normalBaeDescription" : "Normal BAE processing" ,
2024-02-11 08:00:51 +00:00
"dwOpenpose" : "DW Openpose" ,
"dwOpenposeDescription" : "Human pose estimation using DW Openpose" ,
2023-09-13 07:31:34 +00:00
"pidi" : "PIDI" ,
"pidiDescription" : "PIDI image processing" ,
"processor" : "Processor" ,
"prompt" : "Prompt" ,
"resetControlImage" : "Reset Control Image" ,
"resize" : "Resize" ,
2023-12-28 13:03:21 +00:00
"resizeSimple" : "Resize (Simple)" ,
2023-09-13 07:31:34 +00:00
"resizeMode" : "Resize Mode" ,
2024-04-13 06:36:59 +00:00
"ipAdapterMethod" : "Method" ,
"full" : "Full" ,
"style" : "Style Only" ,
"composition" : "Composition Only" ,
2023-09-13 07:31:34 +00:00
"safe" : "Safe" ,
"saveControlImage" : "Save Control Image" ,
2024-05-01 22:46:21 +00:00
"scribble" : "Scribble" ,
2023-09-13 07:31:34 +00:00
"selectModel" : "Select a model" ,
2024-03-27 15:02:41 +00:00
"selectCLIPVisionModel" : "Select a CLIP Vision model" ,
2024-04-30 02:02:24 +00:00
"setControlImageDimensions" : "Copy size to W/H (optimize for model)" ,
"setControlImageDimensionsForce" : "Copy size to W/H (ignore model)" ,
2023-09-13 07:31:34 +00:00
"showAdvanced" : "Show Advanced" ,
2024-01-22 22:28:24 +00:00
"small" : "Small" ,
2023-09-13 07:31:34 +00:00
"toggleControlNet" : "Toggle this ControlNet" ,
"w" : "W" ,
2024-03-13 11:37:47 +00:00
"weight" : "Weight"
2023-09-13 07:31:34 +00:00
} ,
2023-11-11 00:11:46 +00:00
"hrf" : {
"hrf" : "High Resolution Fix" ,
"enableHrf" : "Enable High Resolution Fix" ,
"upscaleMethod" : "Upscale Method" ,
"metadata" : {
"enabled" : "High Resolution Fix Enabled" ,
"strength" : "High Resolution Fix Strength" ,
"method" : "High Resolution Fix Method"
}
} ,
2024-02-28 16:45:34 +00:00
"prompt" : {
"addPromptTrigger" : "Add Prompt Trigger" ,
"compatibleEmbeddings" : "Compatible Embeddings" ,
"noMatchingTriggers" : "No matching triggers"
} ,
feat: queued generation (#4502)
* fix(config): fix typing issues in `config/`
`config/invokeai_config.py`:
- use `Optional` for things that are optional
- fix typing of `ram_cache_size()` and `vram_cache_size()`
- remove unused and incorrectly typed method `autoconvert_path`
- fix types and logic for `parse_args()`, in which `InvokeAIAppConfig.initconf` *must* be a `DictConfig`, but function would allow it to be set as a `ListConfig`, which presumably would cause issues elsewhere
`config/base.py`:
- use `cls` for first arg of class methods
- use `Optional` for things that are optional
- fix minor type issue related to setting of `env_prefix`
- remove unused `add_subparser()` method, which calls `add_parser()` on an `ArgumentParser` (method only available on the `_SubParsersAction` object, which is returned from ArgumentParser.add_subparsers()`)
* feat: queued generation and batches
Due to a very messy branch with broad addition of `isort` on `main` alongside it, some git surgery was needed to get an agreeable git history. This commit represents all of the work on queued generation. See PR for notes.
* chore: flake8, isort, black
* fix(nodes): fix incorrect service stop() method
* fix(nodes): improve names of a few variables
* fix(tests): fix up tests after changes to batches/queue
* feat(tests): add unit tests for session queue helper functions
* feat(ui): dynamic prompts is always enabled
* feat(queue): add queue_status_changed event
* feat(ui): wip queue graphs
* feat(nodes): move cleanup til after invoker startup
* feat(nodes): add cancel_by_batch_ids
* feat(ui): wip batch graphs & UI
* fix(nodes): remove `Batch.batch_id` from required
* fix(ui): cleanup and use fixedCacheKey for all mutations
* fix(ui): remove orphaned nodes from canvas graphs
* fix(nodes): fix cancel_by_batch_ids result count
* fix(ui): only show cancel batch tooltip when batches were canceled
* chore: isort
* fix(api): return `[""]` when dynamic prompts generates no prompts
Just a simple fallback so we always have a prompt.
* feat(ui): dynamicPrompts.combinatorial is always on
There seems to be little purpose in using the combinatorial generation for dynamic prompts. I've disabled it by hiding it from the UI and defaulting combinatorial to true. If we want to enable it again in the future it's straightforward to do so.
* feat: add queue_id & support logic
* feat(ui): fix upscale button
It prepends the upscale operation to queue
* feat(nodes): return queue item when enqueuing a single graph
This facilitates one-off graph async workflows in the client.
* feat(ui): move controlnet autoprocess to queue
* fix(ui): fix non-serializable DOMRect in redux state
* feat(ui): QueueTable performance tweaks
* feat(ui): update queue list
Queue items expand to show the full queue item. Just as JSON for now.
* wip threaded session_processor
* feat(nodes,ui): fully migrate queue to session_processor
* feat(nodes,ui): add processor events
* feat(ui): ui tweaks
* feat(nodes,ui): consolidate events, reduce network requests
* feat(ui): cleanup & abstract queue hooks
* feat(nodes): optimize batch permutation
Use a generator to do only as much work as is needed.
Previously, though we only ended up creating exactly as many queue items as was needed, there was still some intermediary work that calculated *all* permutations. When that number was very high, the system had a very hard time and used a lot of memory.
The logic has been refactored to use a generator. Additionally, the batch validators are optimized to return early and use less memory.
* feat(ui): add seed behaviour parameter
This dynamic prompts parameter allows the seed to be randomized per prompt or per iteration:
- Per iteration: Use the same seed for all prompts in a single dynamic prompt expansion
- Per prompt: Use a different seed for every single prompt
"Per iteration" is appropriate for exploring a the latents space with a stable starting noise, while "Per prompt" provides more variation.
* fix(ui): remove extraneous random seed nodes from linear graphs
* fix(ui): fix controlnet autoprocess not working when queue is running
* feat(queue): add timestamps to queue status updates
Also show execution time in queue list
* feat(queue): change all execution-related events to use the `queue_id` as the room, also include `queue_item_id` in InvocationQueueItem
This allows for much simpler handling of queue items.
* feat(api): deprecate sessions router
* chore(backend): tidy logging in `dependencies.py`
* fix(backend): respect `use_memory_db`
* feat(backend): add `config.log_sql` (enables sql trace logging)
* feat: add invocation cache
Supersedes #4574
The invocation cache provides simple node memoization functionality. Nodes that use the cache are memoized and not re-executed if their inputs haven't changed. Instead, the stored output is returned.
## Results
This feature provides anywhere some significant to massive performance improvement.
The improvement is most marked on large batches of generations where you only change a couple things (e.g. different seed or prompt for each iteration) and low-VRAM systems, where skipping an extraneous model load is a big deal.
## Overview
A new `invocation_cache` service is added to handle the caching. There's not much to it.
All nodes now inherit a boolean `use_cache` field from `BaseInvocation`. This is a node field and not a class attribute, because specific instances of nodes may want to opt in or out of caching.
The recently-added `invoke_internal()` method on `BaseInvocation` is used as an entrypoint for the cache logic.
To create a cache key, the invocation is first serialized using pydantic's provided `json()` method, skipping the unique `id` field. Then python's very fast builtin `hash()` is used to create an integer key. All implementations of `InvocationCacheBase` must provide a class method `create_key()` which accepts an invocation and outputs a string or integer key.
## In-Memory Implementation
An in-memory implementation is provided. In this implementation, the node outputs are stored in memory as python classes. The in-memory cache does not persist application restarts.
Max node cache size is added as `node_cache_size` under the `Generation` config category.
It defaults to 512 - this number is up for discussion, but given that these are relatively lightweight pydantic models, I think it's safe to up this even higher.
Note that the cache isn't storing the big stuff - tensors and images are store on disk, and outputs include only references to them.
## Node Definition
The default for all nodes is to use the cache. The `@invocation` decorator now accepts an optional `use_cache: bool` argument to override the default of `True`.
Non-deterministic nodes, however, should set this to `False`. Currently, all random-stuff nodes, including `dynamic_prompt`, are set to `False`.
The field name `use_cache` is now effectively a reserved field name and possibly a breaking change if any community nodes use this as a field name. In hindsight, all our reserved field names should have been prefixed with underscores or something.
## One Gotcha
Leaf nodes probably want to opt out of the cache, because if they are not cached, their outputs are not saved again.
If you run the same graph multiple times, you only end up with a single image output, because the image storage side-effects are in the `invoke()` method, which is bypassed if we have a cache hit.
## Linear UI
The linear graphs _almost_ just work, but due to the gotcha, we need to be careful about the final image-outputting node. To resolve this, a `SaveImageInvocation` node is added and used in the linear graphs.
This node is similar to `ImagePrimitive`, except it saves a copy of its input image, and has `use_cache` set to `False` by default.
This is now the leaf node in all linear graphs, and is the only node in those graphs with `use_cache == False` _and_ the only node with `is_intermedate == False`.
## Workflow Editor
All nodes now have a footer with a new `Use Cache [ ]` checkbox. It defaults to the value set by the invocation in its python definition, but can be changed by the user.
The workflow/node validation logic has been updated to migrate old workflows to use the new default values for `use_cache`. Users may still want to review the settings that have been chosen. In the event of catastrophic failure when running this migration, the default value of `True` is applied, as this is correct for most nodes.
Users should consider saving their workflows after loading them in and having them updated.
## Future Enhancements - Callback
A future enhancement would be to provide a callback to the `use_cache` flag that would be run as the node is executed to determine, based on its own internal state, if the cache should be used or not.
This would be useful for `DynamicPromptInvocation`, where the deterministic behaviour is determined by the `combinatorial: bool` field.
## Future Enhancements - Persisted Cache
Similar to how the latents storage is backed by disk, the invocation cache could be persisted to the database or disk. We'd need to be very careful about deserializing outputs, but it's perhaps worth exploring in the future.
* fix(ui): fix queue list item width
* feat(nodes): do not send the whole node on every generator progress
* feat(ui): strip out old logic related to sessions
Things like `isProcessing` are no longer relevant with queue. Removed them all & updated everything be appropriate for queue. May be a few little quirks I've missed...
* feat(ui): fix up param collapse labels
* feat(ui): click queue count to go to queue tab
* tidy(queue): update comment, query format
* feat(ui): fix progress bar when canceling
* fix(ui): fix circular dependency
* feat(nodes): bail on node caching logic if `node_cache_size == 0`
* feat(nodes): handle KeyError on node cache pop
* feat(nodes): bypass cache codepath if caches is disabled
more better no do thing
* fix(ui): reset api cache on connect/disconnect
* feat(ui): prevent enqueue when no prompts generated
* feat(ui): add queue controls to workflow editor
* feat(ui): update floating buttons & other incidental UI tweaks
* fix(ui): fix missing/incorrect translation keys
* fix(tests): add config service to mock invocation services
invoking needs access to `node_cache_size` to occur
* optionally remove pause/resume buttons from queue UI
* option to disable prepending
* chore(ui): remove unused file
* feat(queue): remove `order_id` entirely, `item_id` is now an autoinc pk
---------
Co-authored-by: Mary Hipp <maryhipp@Marys-MacBook-Air.local>
2023-09-20 05:09:24 +00:00
"queue" : {
"queue" : "Queue" ,
"queueFront" : "Add to Front of Queue" ,
"queueBack" : "Add to Queue" ,
"queueEmpty" : "Queue Empty" ,
"enqueueing" : "Queueing Batch" ,
"resume" : "Resume" ,
"resumeTooltip" : "Resume Processor" ,
"resumeSucceeded" : "Processor Resumed" ,
"resumeFailed" : "Problem Resuming Processor" ,
"pause" : "Pause" ,
"pauseTooltip" : "Pause Processor" ,
"pauseSucceeded" : "Processor Paused" ,
"pauseFailed" : "Problem Pausing Processor" ,
"cancel" : "Cancel" ,
"cancelTooltip" : "Cancel Current Item" ,
"cancelSucceeded" : "Item Canceled" ,
"cancelFailed" : "Problem Canceling Item" ,
"prune" : "Prune" ,
"pruneTooltip" : "Prune {{item_count}} Completed Items" ,
"pruneSucceeded" : "Pruned {{item_count}} Completed Items from Queue" ,
"pruneFailed" : "Problem Pruning Queue" ,
"clear" : "Clear" ,
"clearTooltip" : "Cancel and Clear All Items" ,
"clearSucceeded" : "Queue Cleared" ,
"clearFailed" : "Problem Clearing Queue" ,
"cancelBatch" : "Cancel Batch" ,
"cancelItem" : "Cancel Item" ,
"cancelBatchSucceeded" : "Batch Canceled" ,
"cancelBatchFailed" : "Problem Canceling Batch" ,
2023-09-20 12:39:54 +00:00
"clearQueueAlertDialog" : "Clearing the queue immediately cancels any processing items and clears the queue entirely." ,
"clearQueueAlertDialog2" : "Are you sure you want to clear the queue?" ,
feat: queued generation (#4502)
* fix(config): fix typing issues in `config/`
`config/invokeai_config.py`:
- use `Optional` for things that are optional
- fix typing of `ram_cache_size()` and `vram_cache_size()`
- remove unused and incorrectly typed method `autoconvert_path`
- fix types and logic for `parse_args()`, in which `InvokeAIAppConfig.initconf` *must* be a `DictConfig`, but function would allow it to be set as a `ListConfig`, which presumably would cause issues elsewhere
`config/base.py`:
- use `cls` for first arg of class methods
- use `Optional` for things that are optional
- fix minor type issue related to setting of `env_prefix`
- remove unused `add_subparser()` method, which calls `add_parser()` on an `ArgumentParser` (method only available on the `_SubParsersAction` object, which is returned from ArgumentParser.add_subparsers()`)
* feat: queued generation and batches
Due to a very messy branch with broad addition of `isort` on `main` alongside it, some git surgery was needed to get an agreeable git history. This commit represents all of the work on queued generation. See PR for notes.
* chore: flake8, isort, black
* fix(nodes): fix incorrect service stop() method
* fix(nodes): improve names of a few variables
* fix(tests): fix up tests after changes to batches/queue
* feat(tests): add unit tests for session queue helper functions
* feat(ui): dynamic prompts is always enabled
* feat(queue): add queue_status_changed event
* feat(ui): wip queue graphs
* feat(nodes): move cleanup til after invoker startup
* feat(nodes): add cancel_by_batch_ids
* feat(ui): wip batch graphs & UI
* fix(nodes): remove `Batch.batch_id` from required
* fix(ui): cleanup and use fixedCacheKey for all mutations
* fix(ui): remove orphaned nodes from canvas graphs
* fix(nodes): fix cancel_by_batch_ids result count
* fix(ui): only show cancel batch tooltip when batches were canceled
* chore: isort
* fix(api): return `[""]` when dynamic prompts generates no prompts
Just a simple fallback so we always have a prompt.
* feat(ui): dynamicPrompts.combinatorial is always on
There seems to be little purpose in using the combinatorial generation for dynamic prompts. I've disabled it by hiding it from the UI and defaulting combinatorial to true. If we want to enable it again in the future it's straightforward to do so.
* feat: add queue_id & support logic
* feat(ui): fix upscale button
It prepends the upscale operation to queue
* feat(nodes): return queue item when enqueuing a single graph
This facilitates one-off graph async workflows in the client.
* feat(ui): move controlnet autoprocess to queue
* fix(ui): fix non-serializable DOMRect in redux state
* feat(ui): QueueTable performance tweaks
* feat(ui): update queue list
Queue items expand to show the full queue item. Just as JSON for now.
* wip threaded session_processor
* feat(nodes,ui): fully migrate queue to session_processor
* feat(nodes,ui): add processor events
* feat(ui): ui tweaks
* feat(nodes,ui): consolidate events, reduce network requests
* feat(ui): cleanup & abstract queue hooks
* feat(nodes): optimize batch permutation
Use a generator to do only as much work as is needed.
Previously, though we only ended up creating exactly as many queue items as was needed, there was still some intermediary work that calculated *all* permutations. When that number was very high, the system had a very hard time and used a lot of memory.
The logic has been refactored to use a generator. Additionally, the batch validators are optimized to return early and use less memory.
* feat(ui): add seed behaviour parameter
This dynamic prompts parameter allows the seed to be randomized per prompt or per iteration:
- Per iteration: Use the same seed for all prompts in a single dynamic prompt expansion
- Per prompt: Use a different seed for every single prompt
"Per iteration" is appropriate for exploring a the latents space with a stable starting noise, while "Per prompt" provides more variation.
* fix(ui): remove extraneous random seed nodes from linear graphs
* fix(ui): fix controlnet autoprocess not working when queue is running
* feat(queue): add timestamps to queue status updates
Also show execution time in queue list
* feat(queue): change all execution-related events to use the `queue_id` as the room, also include `queue_item_id` in InvocationQueueItem
This allows for much simpler handling of queue items.
* feat(api): deprecate sessions router
* chore(backend): tidy logging in `dependencies.py`
* fix(backend): respect `use_memory_db`
* feat(backend): add `config.log_sql` (enables sql trace logging)
* feat: add invocation cache
Supersedes #4574
The invocation cache provides simple node memoization functionality. Nodes that use the cache are memoized and not re-executed if their inputs haven't changed. Instead, the stored output is returned.
## Results
This feature provides anywhere some significant to massive performance improvement.
The improvement is most marked on large batches of generations where you only change a couple things (e.g. different seed or prompt for each iteration) and low-VRAM systems, where skipping an extraneous model load is a big deal.
## Overview
A new `invocation_cache` service is added to handle the caching. There's not much to it.
All nodes now inherit a boolean `use_cache` field from `BaseInvocation`. This is a node field and not a class attribute, because specific instances of nodes may want to opt in or out of caching.
The recently-added `invoke_internal()` method on `BaseInvocation` is used as an entrypoint for the cache logic.
To create a cache key, the invocation is first serialized using pydantic's provided `json()` method, skipping the unique `id` field. Then python's very fast builtin `hash()` is used to create an integer key. All implementations of `InvocationCacheBase` must provide a class method `create_key()` which accepts an invocation and outputs a string or integer key.
## In-Memory Implementation
An in-memory implementation is provided. In this implementation, the node outputs are stored in memory as python classes. The in-memory cache does not persist application restarts.
Max node cache size is added as `node_cache_size` under the `Generation` config category.
It defaults to 512 - this number is up for discussion, but given that these are relatively lightweight pydantic models, I think it's safe to up this even higher.
Note that the cache isn't storing the big stuff - tensors and images are store on disk, and outputs include only references to them.
## Node Definition
The default for all nodes is to use the cache. The `@invocation` decorator now accepts an optional `use_cache: bool` argument to override the default of `True`.
Non-deterministic nodes, however, should set this to `False`. Currently, all random-stuff nodes, including `dynamic_prompt`, are set to `False`.
The field name `use_cache` is now effectively a reserved field name and possibly a breaking change if any community nodes use this as a field name. In hindsight, all our reserved field names should have been prefixed with underscores or something.
## One Gotcha
Leaf nodes probably want to opt out of the cache, because if they are not cached, their outputs are not saved again.
If you run the same graph multiple times, you only end up with a single image output, because the image storage side-effects are in the `invoke()` method, which is bypassed if we have a cache hit.
## Linear UI
The linear graphs _almost_ just work, but due to the gotcha, we need to be careful about the final image-outputting node. To resolve this, a `SaveImageInvocation` node is added and used in the linear graphs.
This node is similar to `ImagePrimitive`, except it saves a copy of its input image, and has `use_cache` set to `False` by default.
This is now the leaf node in all linear graphs, and is the only node in those graphs with `use_cache == False` _and_ the only node with `is_intermedate == False`.
## Workflow Editor
All nodes now have a footer with a new `Use Cache [ ]` checkbox. It defaults to the value set by the invocation in its python definition, but can be changed by the user.
The workflow/node validation logic has been updated to migrate old workflows to use the new default values for `use_cache`. Users may still want to review the settings that have been chosen. In the event of catastrophic failure when running this migration, the default value of `True` is applied, as this is correct for most nodes.
Users should consider saving their workflows after loading them in and having them updated.
## Future Enhancements - Callback
A future enhancement would be to provide a callback to the `use_cache` flag that would be run as the node is executed to determine, based on its own internal state, if the cache should be used or not.
This would be useful for `DynamicPromptInvocation`, where the deterministic behaviour is determined by the `combinatorial: bool` field.
## Future Enhancements - Persisted Cache
Similar to how the latents storage is backed by disk, the invocation cache could be persisted to the database or disk. We'd need to be very careful about deserializing outputs, but it's perhaps worth exploring in the future.
* fix(ui): fix queue list item width
* feat(nodes): do not send the whole node on every generator progress
* feat(ui): strip out old logic related to sessions
Things like `isProcessing` are no longer relevant with queue. Removed them all & updated everything be appropriate for queue. May be a few little quirks I've missed...
* feat(ui): fix up param collapse labels
* feat(ui): click queue count to go to queue tab
* tidy(queue): update comment, query format
* feat(ui): fix progress bar when canceling
* fix(ui): fix circular dependency
* feat(nodes): bail on node caching logic if `node_cache_size == 0`
* feat(nodes): handle KeyError on node cache pop
* feat(nodes): bypass cache codepath if caches is disabled
more better no do thing
* fix(ui): reset api cache on connect/disconnect
* feat(ui): prevent enqueue when no prompts generated
* feat(ui): add queue controls to workflow editor
* feat(ui): update floating buttons & other incidental UI tweaks
* fix(ui): fix missing/incorrect translation keys
* fix(tests): add config service to mock invocation services
invoking needs access to `node_cache_size` to occur
* optionally remove pause/resume buttons from queue UI
* option to disable prepending
* chore(ui): remove unused file
* feat(queue): remove `order_id` entirely, `item_id` is now an autoinc pk
---------
Co-authored-by: Mary Hipp <maryhipp@Marys-MacBook-Air.local>
2023-09-20 05:09:24 +00:00
"current" : "Current" ,
"next" : "Next" ,
"status" : "Status" ,
"total" : "Total" ,
2023-11-20 06:24:03 +00:00
"time" : "Time" ,
feat: queued generation (#4502)
* fix(config): fix typing issues in `config/`
`config/invokeai_config.py`:
- use `Optional` for things that are optional
- fix typing of `ram_cache_size()` and `vram_cache_size()`
- remove unused and incorrectly typed method `autoconvert_path`
- fix types and logic for `parse_args()`, in which `InvokeAIAppConfig.initconf` *must* be a `DictConfig`, but function would allow it to be set as a `ListConfig`, which presumably would cause issues elsewhere
`config/base.py`:
- use `cls` for first arg of class methods
- use `Optional` for things that are optional
- fix minor type issue related to setting of `env_prefix`
- remove unused `add_subparser()` method, which calls `add_parser()` on an `ArgumentParser` (method only available on the `_SubParsersAction` object, which is returned from ArgumentParser.add_subparsers()`)
* feat: queued generation and batches
Due to a very messy branch with broad addition of `isort` on `main` alongside it, some git surgery was needed to get an agreeable git history. This commit represents all of the work on queued generation. See PR for notes.
* chore: flake8, isort, black
* fix(nodes): fix incorrect service stop() method
* fix(nodes): improve names of a few variables
* fix(tests): fix up tests after changes to batches/queue
* feat(tests): add unit tests for session queue helper functions
* feat(ui): dynamic prompts is always enabled
* feat(queue): add queue_status_changed event
* feat(ui): wip queue graphs
* feat(nodes): move cleanup til after invoker startup
* feat(nodes): add cancel_by_batch_ids
* feat(ui): wip batch graphs & UI
* fix(nodes): remove `Batch.batch_id` from required
* fix(ui): cleanup and use fixedCacheKey for all mutations
* fix(ui): remove orphaned nodes from canvas graphs
* fix(nodes): fix cancel_by_batch_ids result count
* fix(ui): only show cancel batch tooltip when batches were canceled
* chore: isort
* fix(api): return `[""]` when dynamic prompts generates no prompts
Just a simple fallback so we always have a prompt.
* feat(ui): dynamicPrompts.combinatorial is always on
There seems to be little purpose in using the combinatorial generation for dynamic prompts. I've disabled it by hiding it from the UI and defaulting combinatorial to true. If we want to enable it again in the future it's straightforward to do so.
* feat: add queue_id & support logic
* feat(ui): fix upscale button
It prepends the upscale operation to queue
* feat(nodes): return queue item when enqueuing a single graph
This facilitates one-off graph async workflows in the client.
* feat(ui): move controlnet autoprocess to queue
* fix(ui): fix non-serializable DOMRect in redux state
* feat(ui): QueueTable performance tweaks
* feat(ui): update queue list
Queue items expand to show the full queue item. Just as JSON for now.
* wip threaded session_processor
* feat(nodes,ui): fully migrate queue to session_processor
* feat(nodes,ui): add processor events
* feat(ui): ui tweaks
* feat(nodes,ui): consolidate events, reduce network requests
* feat(ui): cleanup & abstract queue hooks
* feat(nodes): optimize batch permutation
Use a generator to do only as much work as is needed.
Previously, though we only ended up creating exactly as many queue items as was needed, there was still some intermediary work that calculated *all* permutations. When that number was very high, the system had a very hard time and used a lot of memory.
The logic has been refactored to use a generator. Additionally, the batch validators are optimized to return early and use less memory.
* feat(ui): add seed behaviour parameter
This dynamic prompts parameter allows the seed to be randomized per prompt or per iteration:
- Per iteration: Use the same seed for all prompts in a single dynamic prompt expansion
- Per prompt: Use a different seed for every single prompt
"Per iteration" is appropriate for exploring a the latents space with a stable starting noise, while "Per prompt" provides more variation.
* fix(ui): remove extraneous random seed nodes from linear graphs
* fix(ui): fix controlnet autoprocess not working when queue is running
* feat(queue): add timestamps to queue status updates
Also show execution time in queue list
* feat(queue): change all execution-related events to use the `queue_id` as the room, also include `queue_item_id` in InvocationQueueItem
This allows for much simpler handling of queue items.
* feat(api): deprecate sessions router
* chore(backend): tidy logging in `dependencies.py`
* fix(backend): respect `use_memory_db`
* feat(backend): add `config.log_sql` (enables sql trace logging)
* feat: add invocation cache
Supersedes #4574
The invocation cache provides simple node memoization functionality. Nodes that use the cache are memoized and not re-executed if their inputs haven't changed. Instead, the stored output is returned.
## Results
This feature provides anywhere some significant to massive performance improvement.
The improvement is most marked on large batches of generations where you only change a couple things (e.g. different seed or prompt for each iteration) and low-VRAM systems, where skipping an extraneous model load is a big deal.
## Overview
A new `invocation_cache` service is added to handle the caching. There's not much to it.
All nodes now inherit a boolean `use_cache` field from `BaseInvocation`. This is a node field and not a class attribute, because specific instances of nodes may want to opt in or out of caching.
The recently-added `invoke_internal()` method on `BaseInvocation` is used as an entrypoint for the cache logic.
To create a cache key, the invocation is first serialized using pydantic's provided `json()` method, skipping the unique `id` field. Then python's very fast builtin `hash()` is used to create an integer key. All implementations of `InvocationCacheBase` must provide a class method `create_key()` which accepts an invocation and outputs a string or integer key.
## In-Memory Implementation
An in-memory implementation is provided. In this implementation, the node outputs are stored in memory as python classes. The in-memory cache does not persist application restarts.
Max node cache size is added as `node_cache_size` under the `Generation` config category.
It defaults to 512 - this number is up for discussion, but given that these are relatively lightweight pydantic models, I think it's safe to up this even higher.
Note that the cache isn't storing the big stuff - tensors and images are store on disk, and outputs include only references to them.
## Node Definition
The default for all nodes is to use the cache. The `@invocation` decorator now accepts an optional `use_cache: bool` argument to override the default of `True`.
Non-deterministic nodes, however, should set this to `False`. Currently, all random-stuff nodes, including `dynamic_prompt`, are set to `False`.
The field name `use_cache` is now effectively a reserved field name and possibly a breaking change if any community nodes use this as a field name. In hindsight, all our reserved field names should have been prefixed with underscores or something.
## One Gotcha
Leaf nodes probably want to opt out of the cache, because if they are not cached, their outputs are not saved again.
If you run the same graph multiple times, you only end up with a single image output, because the image storage side-effects are in the `invoke()` method, which is bypassed if we have a cache hit.
## Linear UI
The linear graphs _almost_ just work, but due to the gotcha, we need to be careful about the final image-outputting node. To resolve this, a `SaveImageInvocation` node is added and used in the linear graphs.
This node is similar to `ImagePrimitive`, except it saves a copy of its input image, and has `use_cache` set to `False` by default.
This is now the leaf node in all linear graphs, and is the only node in those graphs with `use_cache == False` _and_ the only node with `is_intermedate == False`.
## Workflow Editor
All nodes now have a footer with a new `Use Cache [ ]` checkbox. It defaults to the value set by the invocation in its python definition, but can be changed by the user.
The workflow/node validation logic has been updated to migrate old workflows to use the new default values for `use_cache`. Users may still want to review the settings that have been chosen. In the event of catastrophic failure when running this migration, the default value of `True` is applied, as this is correct for most nodes.
Users should consider saving their workflows after loading them in and having them updated.
## Future Enhancements - Callback
A future enhancement would be to provide a callback to the `use_cache` flag that would be run as the node is executed to determine, based on its own internal state, if the cache should be used or not.
This would be useful for `DynamicPromptInvocation`, where the deterministic behaviour is determined by the `combinatorial: bool` field.
## Future Enhancements - Persisted Cache
Similar to how the latents storage is backed by disk, the invocation cache could be persisted to the database or disk. We'd need to be very careful about deserializing outputs, but it's perhaps worth exploring in the future.
* fix(ui): fix queue list item width
* feat(nodes): do not send the whole node on every generator progress
* feat(ui): strip out old logic related to sessions
Things like `isProcessing` are no longer relevant with queue. Removed them all & updated everything be appropriate for queue. May be a few little quirks I've missed...
* feat(ui): fix up param collapse labels
* feat(ui): click queue count to go to queue tab
* tidy(queue): update comment, query format
* feat(ui): fix progress bar when canceling
* fix(ui): fix circular dependency
* feat(nodes): bail on node caching logic if `node_cache_size == 0`
* feat(nodes): handle KeyError on node cache pop
* feat(nodes): bypass cache codepath if caches is disabled
more better no do thing
* fix(ui): reset api cache on connect/disconnect
* feat(ui): prevent enqueue when no prompts generated
* feat(ui): add queue controls to workflow editor
* feat(ui): update floating buttons & other incidental UI tweaks
* fix(ui): fix missing/incorrect translation keys
* fix(tests): add config service to mock invocation services
invoking needs access to `node_cache_size` to occur
* optionally remove pause/resume buttons from queue UI
* option to disable prepending
* chore(ui): remove unused file
* feat(queue): remove `order_id` entirely, `item_id` is now an autoinc pk
---------
Co-authored-by: Mary Hipp <maryhipp@Marys-MacBook-Air.local>
2023-09-20 05:09:24 +00:00
"pending" : "Pending" ,
"in_progress" : "In Progress" ,
"completed" : "Completed" ,
"failed" : "Failed" ,
"canceled" : "Canceled" ,
"completedIn" : "Completed in" ,
"batch" : "Batch" ,
2023-11-20 06:24:03 +00:00
"batchFieldValues" : "Batch Field Values" ,
feat: queued generation (#4502)
* fix(config): fix typing issues in `config/`
`config/invokeai_config.py`:
- use `Optional` for things that are optional
- fix typing of `ram_cache_size()` and `vram_cache_size()`
- remove unused and incorrectly typed method `autoconvert_path`
- fix types and logic for `parse_args()`, in which `InvokeAIAppConfig.initconf` *must* be a `DictConfig`, but function would allow it to be set as a `ListConfig`, which presumably would cause issues elsewhere
`config/base.py`:
- use `cls` for first arg of class methods
- use `Optional` for things that are optional
- fix minor type issue related to setting of `env_prefix`
- remove unused `add_subparser()` method, which calls `add_parser()` on an `ArgumentParser` (method only available on the `_SubParsersAction` object, which is returned from ArgumentParser.add_subparsers()`)
* feat: queued generation and batches
Due to a very messy branch with broad addition of `isort` on `main` alongside it, some git surgery was needed to get an agreeable git history. This commit represents all of the work on queued generation. See PR for notes.
* chore: flake8, isort, black
* fix(nodes): fix incorrect service stop() method
* fix(nodes): improve names of a few variables
* fix(tests): fix up tests after changes to batches/queue
* feat(tests): add unit tests for session queue helper functions
* feat(ui): dynamic prompts is always enabled
* feat(queue): add queue_status_changed event
* feat(ui): wip queue graphs
* feat(nodes): move cleanup til after invoker startup
* feat(nodes): add cancel_by_batch_ids
* feat(ui): wip batch graphs & UI
* fix(nodes): remove `Batch.batch_id` from required
* fix(ui): cleanup and use fixedCacheKey for all mutations
* fix(ui): remove orphaned nodes from canvas graphs
* fix(nodes): fix cancel_by_batch_ids result count
* fix(ui): only show cancel batch tooltip when batches were canceled
* chore: isort
* fix(api): return `[""]` when dynamic prompts generates no prompts
Just a simple fallback so we always have a prompt.
* feat(ui): dynamicPrompts.combinatorial is always on
There seems to be little purpose in using the combinatorial generation for dynamic prompts. I've disabled it by hiding it from the UI and defaulting combinatorial to true. If we want to enable it again in the future it's straightforward to do so.
* feat: add queue_id & support logic
* feat(ui): fix upscale button
It prepends the upscale operation to queue
* feat(nodes): return queue item when enqueuing a single graph
This facilitates one-off graph async workflows in the client.
* feat(ui): move controlnet autoprocess to queue
* fix(ui): fix non-serializable DOMRect in redux state
* feat(ui): QueueTable performance tweaks
* feat(ui): update queue list
Queue items expand to show the full queue item. Just as JSON for now.
* wip threaded session_processor
* feat(nodes,ui): fully migrate queue to session_processor
* feat(nodes,ui): add processor events
* feat(ui): ui tweaks
* feat(nodes,ui): consolidate events, reduce network requests
* feat(ui): cleanup & abstract queue hooks
* feat(nodes): optimize batch permutation
Use a generator to do only as much work as is needed.
Previously, though we only ended up creating exactly as many queue items as was needed, there was still some intermediary work that calculated *all* permutations. When that number was very high, the system had a very hard time and used a lot of memory.
The logic has been refactored to use a generator. Additionally, the batch validators are optimized to return early and use less memory.
* feat(ui): add seed behaviour parameter
This dynamic prompts parameter allows the seed to be randomized per prompt or per iteration:
- Per iteration: Use the same seed for all prompts in a single dynamic prompt expansion
- Per prompt: Use a different seed for every single prompt
"Per iteration" is appropriate for exploring a the latents space with a stable starting noise, while "Per prompt" provides more variation.
* fix(ui): remove extraneous random seed nodes from linear graphs
* fix(ui): fix controlnet autoprocess not working when queue is running
* feat(queue): add timestamps to queue status updates
Also show execution time in queue list
* feat(queue): change all execution-related events to use the `queue_id` as the room, also include `queue_item_id` in InvocationQueueItem
This allows for much simpler handling of queue items.
* feat(api): deprecate sessions router
* chore(backend): tidy logging in `dependencies.py`
* fix(backend): respect `use_memory_db`
* feat(backend): add `config.log_sql` (enables sql trace logging)
* feat: add invocation cache
Supersedes #4574
The invocation cache provides simple node memoization functionality. Nodes that use the cache are memoized and not re-executed if their inputs haven't changed. Instead, the stored output is returned.
## Results
This feature provides anywhere some significant to massive performance improvement.
The improvement is most marked on large batches of generations where you only change a couple things (e.g. different seed or prompt for each iteration) and low-VRAM systems, where skipping an extraneous model load is a big deal.
## Overview
A new `invocation_cache` service is added to handle the caching. There's not much to it.
All nodes now inherit a boolean `use_cache` field from `BaseInvocation`. This is a node field and not a class attribute, because specific instances of nodes may want to opt in or out of caching.
The recently-added `invoke_internal()` method on `BaseInvocation` is used as an entrypoint for the cache logic.
To create a cache key, the invocation is first serialized using pydantic's provided `json()` method, skipping the unique `id` field. Then python's very fast builtin `hash()` is used to create an integer key. All implementations of `InvocationCacheBase` must provide a class method `create_key()` which accepts an invocation and outputs a string or integer key.
## In-Memory Implementation
An in-memory implementation is provided. In this implementation, the node outputs are stored in memory as python classes. The in-memory cache does not persist application restarts.
Max node cache size is added as `node_cache_size` under the `Generation` config category.
It defaults to 512 - this number is up for discussion, but given that these are relatively lightweight pydantic models, I think it's safe to up this even higher.
Note that the cache isn't storing the big stuff - tensors and images are store on disk, and outputs include only references to them.
## Node Definition
The default for all nodes is to use the cache. The `@invocation` decorator now accepts an optional `use_cache: bool` argument to override the default of `True`.
Non-deterministic nodes, however, should set this to `False`. Currently, all random-stuff nodes, including `dynamic_prompt`, are set to `False`.
The field name `use_cache` is now effectively a reserved field name and possibly a breaking change if any community nodes use this as a field name. In hindsight, all our reserved field names should have been prefixed with underscores or something.
## One Gotcha
Leaf nodes probably want to opt out of the cache, because if they are not cached, their outputs are not saved again.
If you run the same graph multiple times, you only end up with a single image output, because the image storage side-effects are in the `invoke()` method, which is bypassed if we have a cache hit.
## Linear UI
The linear graphs _almost_ just work, but due to the gotcha, we need to be careful about the final image-outputting node. To resolve this, a `SaveImageInvocation` node is added and used in the linear graphs.
This node is similar to `ImagePrimitive`, except it saves a copy of its input image, and has `use_cache` set to `False` by default.
This is now the leaf node in all linear graphs, and is the only node in those graphs with `use_cache == False` _and_ the only node with `is_intermedate == False`.
## Workflow Editor
All nodes now have a footer with a new `Use Cache [ ]` checkbox. It defaults to the value set by the invocation in its python definition, but can be changed by the user.
The workflow/node validation logic has been updated to migrate old workflows to use the new default values for `use_cache`. Users may still want to review the settings that have been chosen. In the event of catastrophic failure when running this migration, the default value of `True` is applied, as this is correct for most nodes.
Users should consider saving their workflows after loading them in and having them updated.
## Future Enhancements - Callback
A future enhancement would be to provide a callback to the `use_cache` flag that would be run as the node is executed to determine, based on its own internal state, if the cache should be used or not.
This would be useful for `DynamicPromptInvocation`, where the deterministic behaviour is determined by the `combinatorial: bool` field.
## Future Enhancements - Persisted Cache
Similar to how the latents storage is backed by disk, the invocation cache could be persisted to the database or disk. We'd need to be very careful about deserializing outputs, but it's perhaps worth exploring in the future.
* fix(ui): fix queue list item width
* feat(nodes): do not send the whole node on every generator progress
* feat(ui): strip out old logic related to sessions
Things like `isProcessing` are no longer relevant with queue. Removed them all & updated everything be appropriate for queue. May be a few little quirks I've missed...
* feat(ui): fix up param collapse labels
* feat(ui): click queue count to go to queue tab
* tidy(queue): update comment, query format
* feat(ui): fix progress bar when canceling
* fix(ui): fix circular dependency
* feat(nodes): bail on node caching logic if `node_cache_size == 0`
* feat(nodes): handle KeyError on node cache pop
* feat(nodes): bypass cache codepath if caches is disabled
more better no do thing
* fix(ui): reset api cache on connect/disconnect
* feat(ui): prevent enqueue when no prompts generated
* feat(ui): add queue controls to workflow editor
* feat(ui): update floating buttons & other incidental UI tweaks
* fix(ui): fix missing/incorrect translation keys
* fix(tests): add config service to mock invocation services
invoking needs access to `node_cache_size` to occur
* optionally remove pause/resume buttons from queue UI
* option to disable prepending
* chore(ui): remove unused file
* feat(queue): remove `order_id` entirely, `item_id` is now an autoinc pk
---------
Co-authored-by: Mary Hipp <maryhipp@Marys-MacBook-Air.local>
2023-09-20 05:09:24 +00:00
"item" : "Item" ,
"session" : "Session" ,
"notReady" : "Unable to Queue" ,
"batchQueued" : "Batch Queued" ,
2023-10-12 10:34:24 +00:00
"batchQueuedDesc_one" : "Added {{count}} sessions to {{direction}} of queue" ,
"batchQueuedDesc_other" : "Added {{count}} sessions to {{direction}} of queue" ,
feat: queued generation (#4502)
* fix(config): fix typing issues in `config/`
`config/invokeai_config.py`:
- use `Optional` for things that are optional
- fix typing of `ram_cache_size()` and `vram_cache_size()`
- remove unused and incorrectly typed method `autoconvert_path`
- fix types and logic for `parse_args()`, in which `InvokeAIAppConfig.initconf` *must* be a `DictConfig`, but function would allow it to be set as a `ListConfig`, which presumably would cause issues elsewhere
`config/base.py`:
- use `cls` for first arg of class methods
- use `Optional` for things that are optional
- fix minor type issue related to setting of `env_prefix`
- remove unused `add_subparser()` method, which calls `add_parser()` on an `ArgumentParser` (method only available on the `_SubParsersAction` object, which is returned from ArgumentParser.add_subparsers()`)
* feat: queued generation and batches
Due to a very messy branch with broad addition of `isort` on `main` alongside it, some git surgery was needed to get an agreeable git history. This commit represents all of the work on queued generation. See PR for notes.
* chore: flake8, isort, black
* fix(nodes): fix incorrect service stop() method
* fix(nodes): improve names of a few variables
* fix(tests): fix up tests after changes to batches/queue
* feat(tests): add unit tests for session queue helper functions
* feat(ui): dynamic prompts is always enabled
* feat(queue): add queue_status_changed event
* feat(ui): wip queue graphs
* feat(nodes): move cleanup til after invoker startup
* feat(nodes): add cancel_by_batch_ids
* feat(ui): wip batch graphs & UI
* fix(nodes): remove `Batch.batch_id` from required
* fix(ui): cleanup and use fixedCacheKey for all mutations
* fix(ui): remove orphaned nodes from canvas graphs
* fix(nodes): fix cancel_by_batch_ids result count
* fix(ui): only show cancel batch tooltip when batches were canceled
* chore: isort
* fix(api): return `[""]` when dynamic prompts generates no prompts
Just a simple fallback so we always have a prompt.
* feat(ui): dynamicPrompts.combinatorial is always on
There seems to be little purpose in using the combinatorial generation for dynamic prompts. I've disabled it by hiding it from the UI and defaulting combinatorial to true. If we want to enable it again in the future it's straightforward to do so.
* feat: add queue_id & support logic
* feat(ui): fix upscale button
It prepends the upscale operation to queue
* feat(nodes): return queue item when enqueuing a single graph
This facilitates one-off graph async workflows in the client.
* feat(ui): move controlnet autoprocess to queue
* fix(ui): fix non-serializable DOMRect in redux state
* feat(ui): QueueTable performance tweaks
* feat(ui): update queue list
Queue items expand to show the full queue item. Just as JSON for now.
* wip threaded session_processor
* feat(nodes,ui): fully migrate queue to session_processor
* feat(nodes,ui): add processor events
* feat(ui): ui tweaks
* feat(nodes,ui): consolidate events, reduce network requests
* feat(ui): cleanup & abstract queue hooks
* feat(nodes): optimize batch permutation
Use a generator to do only as much work as is needed.
Previously, though we only ended up creating exactly as many queue items as was needed, there was still some intermediary work that calculated *all* permutations. When that number was very high, the system had a very hard time and used a lot of memory.
The logic has been refactored to use a generator. Additionally, the batch validators are optimized to return early and use less memory.
* feat(ui): add seed behaviour parameter
This dynamic prompts parameter allows the seed to be randomized per prompt or per iteration:
- Per iteration: Use the same seed for all prompts in a single dynamic prompt expansion
- Per prompt: Use a different seed for every single prompt
"Per iteration" is appropriate for exploring a the latents space with a stable starting noise, while "Per prompt" provides more variation.
* fix(ui): remove extraneous random seed nodes from linear graphs
* fix(ui): fix controlnet autoprocess not working when queue is running
* feat(queue): add timestamps to queue status updates
Also show execution time in queue list
* feat(queue): change all execution-related events to use the `queue_id` as the room, also include `queue_item_id` in InvocationQueueItem
This allows for much simpler handling of queue items.
* feat(api): deprecate sessions router
* chore(backend): tidy logging in `dependencies.py`
* fix(backend): respect `use_memory_db`
* feat(backend): add `config.log_sql` (enables sql trace logging)
* feat: add invocation cache
Supersedes #4574
The invocation cache provides simple node memoization functionality. Nodes that use the cache are memoized and not re-executed if their inputs haven't changed. Instead, the stored output is returned.
## Results
This feature provides anywhere some significant to massive performance improvement.
The improvement is most marked on large batches of generations where you only change a couple things (e.g. different seed or prompt for each iteration) and low-VRAM systems, where skipping an extraneous model load is a big deal.
## Overview
A new `invocation_cache` service is added to handle the caching. There's not much to it.
All nodes now inherit a boolean `use_cache` field from `BaseInvocation`. This is a node field and not a class attribute, because specific instances of nodes may want to opt in or out of caching.
The recently-added `invoke_internal()` method on `BaseInvocation` is used as an entrypoint for the cache logic.
To create a cache key, the invocation is first serialized using pydantic's provided `json()` method, skipping the unique `id` field. Then python's very fast builtin `hash()` is used to create an integer key. All implementations of `InvocationCacheBase` must provide a class method `create_key()` which accepts an invocation and outputs a string or integer key.
## In-Memory Implementation
An in-memory implementation is provided. In this implementation, the node outputs are stored in memory as python classes. The in-memory cache does not persist application restarts.
Max node cache size is added as `node_cache_size` under the `Generation` config category.
It defaults to 512 - this number is up for discussion, but given that these are relatively lightweight pydantic models, I think it's safe to up this even higher.
Note that the cache isn't storing the big stuff - tensors and images are store on disk, and outputs include only references to them.
## Node Definition
The default for all nodes is to use the cache. The `@invocation` decorator now accepts an optional `use_cache: bool` argument to override the default of `True`.
Non-deterministic nodes, however, should set this to `False`. Currently, all random-stuff nodes, including `dynamic_prompt`, are set to `False`.
The field name `use_cache` is now effectively a reserved field name and possibly a breaking change if any community nodes use this as a field name. In hindsight, all our reserved field names should have been prefixed with underscores or something.
## One Gotcha
Leaf nodes probably want to opt out of the cache, because if they are not cached, their outputs are not saved again.
If you run the same graph multiple times, you only end up with a single image output, because the image storage side-effects are in the `invoke()` method, which is bypassed if we have a cache hit.
## Linear UI
The linear graphs _almost_ just work, but due to the gotcha, we need to be careful about the final image-outputting node. To resolve this, a `SaveImageInvocation` node is added and used in the linear graphs.
This node is similar to `ImagePrimitive`, except it saves a copy of its input image, and has `use_cache` set to `False` by default.
This is now the leaf node in all linear graphs, and is the only node in those graphs with `use_cache == False` _and_ the only node with `is_intermedate == False`.
## Workflow Editor
All nodes now have a footer with a new `Use Cache [ ]` checkbox. It defaults to the value set by the invocation in its python definition, but can be changed by the user.
The workflow/node validation logic has been updated to migrate old workflows to use the new default values for `use_cache`. Users may still want to review the settings that have been chosen. In the event of catastrophic failure when running this migration, the default value of `True` is applied, as this is correct for most nodes.
Users should consider saving their workflows after loading them in and having them updated.
## Future Enhancements - Callback
A future enhancement would be to provide a callback to the `use_cache` flag that would be run as the node is executed to determine, based on its own internal state, if the cache should be used or not.
This would be useful for `DynamicPromptInvocation`, where the deterministic behaviour is determined by the `combinatorial: bool` field.
## Future Enhancements - Persisted Cache
Similar to how the latents storage is backed by disk, the invocation cache could be persisted to the database or disk. We'd need to be very careful about deserializing outputs, but it's perhaps worth exploring in the future.
* fix(ui): fix queue list item width
* feat(nodes): do not send the whole node on every generator progress
* feat(ui): strip out old logic related to sessions
Things like `isProcessing` are no longer relevant with queue. Removed them all & updated everything be appropriate for queue. May be a few little quirks I've missed...
* feat(ui): fix up param collapse labels
* feat(ui): click queue count to go to queue tab
* tidy(queue): update comment, query format
* feat(ui): fix progress bar when canceling
* fix(ui): fix circular dependency
* feat(nodes): bail on node caching logic if `node_cache_size == 0`
* feat(nodes): handle KeyError on node cache pop
* feat(nodes): bypass cache codepath if caches is disabled
more better no do thing
* fix(ui): reset api cache on connect/disconnect
* feat(ui): prevent enqueue when no prompts generated
* feat(ui): add queue controls to workflow editor
* feat(ui): update floating buttons & other incidental UI tweaks
* fix(ui): fix missing/incorrect translation keys
* fix(tests): add config service to mock invocation services
invoking needs access to `node_cache_size` to occur
* optionally remove pause/resume buttons from queue UI
* option to disable prepending
* chore(ui): remove unused file
* feat(queue): remove `order_id` entirely, `item_id` is now an autoinc pk
---------
Co-authored-by: Mary Hipp <maryhipp@Marys-MacBook-Air.local>
2023-09-20 05:09:24 +00:00
"front" : "front" ,
"back" : "back" ,
"batchFailedToQueue" : "Failed to Queue Batch" ,
"graphQueued" : "Graph queued" ,
2024-01-02 08:02:39 +00:00
"graphFailedToQueue" : "Failed to queue graph" ,
2024-05-12 21:44:33 +00:00
"openQueue" : "Open Queue" ,
"prompts_one" : "Prompt" ,
"prompts_other" : "Prompts" ,
"iterations_one" : "Iteration" ,
"iterations_other" : "Iterations" ,
"generations_one" : "Generation" ,
"generations_other" : "Generations"
feat: queued generation (#4502)
* fix(config): fix typing issues in `config/`
`config/invokeai_config.py`:
- use `Optional` for things that are optional
- fix typing of `ram_cache_size()` and `vram_cache_size()`
- remove unused and incorrectly typed method `autoconvert_path`
- fix types and logic for `parse_args()`, in which `InvokeAIAppConfig.initconf` *must* be a `DictConfig`, but function would allow it to be set as a `ListConfig`, which presumably would cause issues elsewhere
`config/base.py`:
- use `cls` for first arg of class methods
- use `Optional` for things that are optional
- fix minor type issue related to setting of `env_prefix`
- remove unused `add_subparser()` method, which calls `add_parser()` on an `ArgumentParser` (method only available on the `_SubParsersAction` object, which is returned from ArgumentParser.add_subparsers()`)
* feat: queued generation and batches
Due to a very messy branch with broad addition of `isort` on `main` alongside it, some git surgery was needed to get an agreeable git history. This commit represents all of the work on queued generation. See PR for notes.
* chore: flake8, isort, black
* fix(nodes): fix incorrect service stop() method
* fix(nodes): improve names of a few variables
* fix(tests): fix up tests after changes to batches/queue
* feat(tests): add unit tests for session queue helper functions
* feat(ui): dynamic prompts is always enabled
* feat(queue): add queue_status_changed event
* feat(ui): wip queue graphs
* feat(nodes): move cleanup til after invoker startup
* feat(nodes): add cancel_by_batch_ids
* feat(ui): wip batch graphs & UI
* fix(nodes): remove `Batch.batch_id` from required
* fix(ui): cleanup and use fixedCacheKey for all mutations
* fix(ui): remove orphaned nodes from canvas graphs
* fix(nodes): fix cancel_by_batch_ids result count
* fix(ui): only show cancel batch tooltip when batches were canceled
* chore: isort
* fix(api): return `[""]` when dynamic prompts generates no prompts
Just a simple fallback so we always have a prompt.
* feat(ui): dynamicPrompts.combinatorial is always on
There seems to be little purpose in using the combinatorial generation for dynamic prompts. I've disabled it by hiding it from the UI and defaulting combinatorial to true. If we want to enable it again in the future it's straightforward to do so.
* feat: add queue_id & support logic
* feat(ui): fix upscale button
It prepends the upscale operation to queue
* feat(nodes): return queue item when enqueuing a single graph
This facilitates one-off graph async workflows in the client.
* feat(ui): move controlnet autoprocess to queue
* fix(ui): fix non-serializable DOMRect in redux state
* feat(ui): QueueTable performance tweaks
* feat(ui): update queue list
Queue items expand to show the full queue item. Just as JSON for now.
* wip threaded session_processor
* feat(nodes,ui): fully migrate queue to session_processor
* feat(nodes,ui): add processor events
* feat(ui): ui tweaks
* feat(nodes,ui): consolidate events, reduce network requests
* feat(ui): cleanup & abstract queue hooks
* feat(nodes): optimize batch permutation
Use a generator to do only as much work as is needed.
Previously, though we only ended up creating exactly as many queue items as was needed, there was still some intermediary work that calculated *all* permutations. When that number was very high, the system had a very hard time and used a lot of memory.
The logic has been refactored to use a generator. Additionally, the batch validators are optimized to return early and use less memory.
* feat(ui): add seed behaviour parameter
This dynamic prompts parameter allows the seed to be randomized per prompt or per iteration:
- Per iteration: Use the same seed for all prompts in a single dynamic prompt expansion
- Per prompt: Use a different seed for every single prompt
"Per iteration" is appropriate for exploring a the latents space with a stable starting noise, while "Per prompt" provides more variation.
* fix(ui): remove extraneous random seed nodes from linear graphs
* fix(ui): fix controlnet autoprocess not working when queue is running
* feat(queue): add timestamps to queue status updates
Also show execution time in queue list
* feat(queue): change all execution-related events to use the `queue_id` as the room, also include `queue_item_id` in InvocationQueueItem
This allows for much simpler handling of queue items.
* feat(api): deprecate sessions router
* chore(backend): tidy logging in `dependencies.py`
* fix(backend): respect `use_memory_db`
* feat(backend): add `config.log_sql` (enables sql trace logging)
* feat: add invocation cache
Supersedes #4574
The invocation cache provides simple node memoization functionality. Nodes that use the cache are memoized and not re-executed if their inputs haven't changed. Instead, the stored output is returned.
## Results
This feature provides anywhere some significant to massive performance improvement.
The improvement is most marked on large batches of generations where you only change a couple things (e.g. different seed or prompt for each iteration) and low-VRAM systems, where skipping an extraneous model load is a big deal.
## Overview
A new `invocation_cache` service is added to handle the caching. There's not much to it.
All nodes now inherit a boolean `use_cache` field from `BaseInvocation`. This is a node field and not a class attribute, because specific instances of nodes may want to opt in or out of caching.
The recently-added `invoke_internal()` method on `BaseInvocation` is used as an entrypoint for the cache logic.
To create a cache key, the invocation is first serialized using pydantic's provided `json()` method, skipping the unique `id` field. Then python's very fast builtin `hash()` is used to create an integer key. All implementations of `InvocationCacheBase` must provide a class method `create_key()` which accepts an invocation and outputs a string or integer key.
## In-Memory Implementation
An in-memory implementation is provided. In this implementation, the node outputs are stored in memory as python classes. The in-memory cache does not persist application restarts.
Max node cache size is added as `node_cache_size` under the `Generation` config category.
It defaults to 512 - this number is up for discussion, but given that these are relatively lightweight pydantic models, I think it's safe to up this even higher.
Note that the cache isn't storing the big stuff - tensors and images are store on disk, and outputs include only references to them.
## Node Definition
The default for all nodes is to use the cache. The `@invocation` decorator now accepts an optional `use_cache: bool` argument to override the default of `True`.
Non-deterministic nodes, however, should set this to `False`. Currently, all random-stuff nodes, including `dynamic_prompt`, are set to `False`.
The field name `use_cache` is now effectively a reserved field name and possibly a breaking change if any community nodes use this as a field name. In hindsight, all our reserved field names should have been prefixed with underscores or something.
## One Gotcha
Leaf nodes probably want to opt out of the cache, because if they are not cached, their outputs are not saved again.
If you run the same graph multiple times, you only end up with a single image output, because the image storage side-effects are in the `invoke()` method, which is bypassed if we have a cache hit.
## Linear UI
The linear graphs _almost_ just work, but due to the gotcha, we need to be careful about the final image-outputting node. To resolve this, a `SaveImageInvocation` node is added and used in the linear graphs.
This node is similar to `ImagePrimitive`, except it saves a copy of its input image, and has `use_cache` set to `False` by default.
This is now the leaf node in all linear graphs, and is the only node in those graphs with `use_cache == False` _and_ the only node with `is_intermedate == False`.
## Workflow Editor
All nodes now have a footer with a new `Use Cache [ ]` checkbox. It defaults to the value set by the invocation in its python definition, but can be changed by the user.
The workflow/node validation logic has been updated to migrate old workflows to use the new default values for `use_cache`. Users may still want to review the settings that have been chosen. In the event of catastrophic failure when running this migration, the default value of `True` is applied, as this is correct for most nodes.
Users should consider saving their workflows after loading them in and having them updated.
## Future Enhancements - Callback
A future enhancement would be to provide a callback to the `use_cache` flag that would be run as the node is executed to determine, based on its own internal state, if the cache should be used or not.
This would be useful for `DynamicPromptInvocation`, where the deterministic behaviour is determined by the `combinatorial: bool` field.
## Future Enhancements - Persisted Cache
Similar to how the latents storage is backed by disk, the invocation cache could be persisted to the database or disk. We'd need to be very careful about deserializing outputs, but it's perhaps worth exploring in the future.
* fix(ui): fix queue list item width
* feat(nodes): do not send the whole node on every generator progress
* feat(ui): strip out old logic related to sessions
Things like `isProcessing` are no longer relevant with queue. Removed them all & updated everything be appropriate for queue. May be a few little quirks I've missed...
* feat(ui): fix up param collapse labels
* feat(ui): click queue count to go to queue tab
* tidy(queue): update comment, query format
* feat(ui): fix progress bar when canceling
* fix(ui): fix circular dependency
* feat(nodes): bail on node caching logic if `node_cache_size == 0`
* feat(nodes): handle KeyError on node cache pop
* feat(nodes): bypass cache codepath if caches is disabled
more better no do thing
* fix(ui): reset api cache on connect/disconnect
* feat(ui): prevent enqueue when no prompts generated
* feat(ui): add queue controls to workflow editor
* feat(ui): update floating buttons & other incidental UI tweaks
* fix(ui): fix missing/incorrect translation keys
* fix(tests): add config service to mock invocation services
invoking needs access to `node_cache_size` to occur
* optionally remove pause/resume buttons from queue UI
* option to disable prepending
* chore(ui): remove unused file
* feat(queue): remove `order_id` entirely, `item_id` is now an autoinc pk
---------
Co-authored-by: Mary Hipp <maryhipp@Marys-MacBook-Air.local>
2023-09-20 05:09:24 +00:00
} ,
2023-09-21 08:57:52 +00:00
"invocationCache" : {
"invocationCache" : "Invocation Cache" ,
"cacheSize" : "Cache Size" ,
"maxCacheSize" : "Max Cache Size" ,
"hits" : "Cache Hits" ,
"misses" : "Cache Misses" ,
"clear" : "Clear" ,
"clearSucceeded" : "Invocation Cache Cleared" ,
"clearFailed" : "Problem Clearing Invocation Cache" ,
"enable" : "Enable" ,
"enableSucceeded" : "Invocation Cache Enabled" ,
"enableFailed" : "Problem Enabling Invocation Cache" ,
"disable" : "Disable" ,
"disableSucceeded" : "Invocation Cache Disabled" ,
2023-11-25 03:46:19 +00:00
"disableFailed" : "Problem Disabling Invocation Cache" ,
"useCache" : "Use Cache"
2023-09-21 08:57:52 +00:00
} ,
2023-02-18 04:23:24 +00:00
"gallery" : {
2024-03-12 07:49:55 +00:00
"alwaysShowImageSizeBadge" : "Always Show Image Size Badge" ,
2023-09-13 07:31:34 +00:00
"assets" : "Assets" ,
"autoAssignBoardOnClick" : "Auto-Assign Board on Click" ,
"autoSwitchNewImages" : "Auto-Switch to New Images" ,
"copy" : "Copy" ,
"currentlyInUse" : "This image is currently in use in the following features:" ,
2023-11-26 09:44:39 +00:00
"drop" : "Drop" ,
"dropOrUpload" : "$t(gallery.drop) or Upload" ,
"dropToUpload" : "$t(gallery.drop) to Upload" ,
2024-04-14 22:51:05 +00:00
"deleteImage_one" : "Delete Image" ,
"deleteImage_other" : "Delete {{count}} Images" ,
2023-05-29 09:54:04 +00:00
"deleteImagePermanent" : "Deleted images cannot be restored." ,
2024-07-10 01:41:03 +00:00
"displayBoardSearch" : "Display Board Search" ,
2024-07-09 18:16:02 +00:00
"displaySearch" : "Display Search" ,
2023-09-13 07:31:34 +00:00
"download" : "Download" ,
2024-07-23 17:18:13 +00:00
"exitBoardSearch" : "Exit Board Search" ,
"exitSearch" : "Exit Search" ,
2023-09-13 07:31:34 +00:00
"featuresWillReset" : "If you delete this image, those features will immediately be reset." ,
"galleryImageSize" : "Image Size" ,
"gallerySettings" : "Gallery Settings" ,
2024-07-24 20:24:23 +00:00
"go" : "Go" ,
2023-11-25 03:46:19 +00:00
"image" : "image" ,
2024-07-24 20:24:23 +00:00
"jump" : "Jump" ,
2023-09-13 07:31:34 +00:00
"loading" : "Loading" ,
"loadMore" : "Load More" ,
2024-06-27 16:20:23 +00:00
"newestFirst" : "Newest First" ,
2024-06-28 06:27:52 +00:00
"oldestFirst" : "Oldest First" ,
"sortDirection" : "Sort Direction" ,
"showStarredImagesFirst" : "Show Starred Images First" ,
2023-11-20 06:24:03 +00:00
"noImageSelected" : "No Image Selected" ,
2023-09-13 07:31:34 +00:00
"noImagesInGallery" : "No Images to Display" ,
"setCurrentImage" : "Set as Current Image" ,
2023-11-30 10:06:50 +00:00
"starImage" : "Star Image" ,
"unstarImage" : "Unstar Image" ,
2023-09-13 07:31:34 +00:00
"unableToLoad" : "Unable to load Gallery" ,
2023-11-25 03:46:19 +00:00
"deleteSelection" : "Delete Selection" ,
2023-10-11 23:27:22 +00:00
"downloadSelection" : "Download Selection" ,
2024-02-20 11:26:38 +00:00
"bulkDownloadRequested" : "Preparing Download" ,
"bulkDownloadRequestedDesc" : "Your download request is being prepared. This may take a few moments." ,
"bulkDownloadRequestFailed" : "Problem Preparing Download" ,
"bulkDownloadFailed" : "Download Failed" ,
2023-12-08 00:26:15 +00:00
"problemDeletingImages" : "Problem Deleting Images" ,
2024-05-31 05:18:31 +00:00
"problemDeletingImagesDesc" : "One or more images could not be deleted" ,
2024-06-01 00:40:56 +00:00
"viewerImage" : "Viewer Image" ,
"compareImage" : "Compare Image" ,
2024-06-02 04:36:24 +00:00
"openInViewer" : "Open in Viewer" ,
2024-07-01 09:53:47 +00:00
"searchImages" : "Search by Metadata" ,
2024-06-24 18:09:32 +00:00
"selectAllOnPage" : "Select All On Page" ,
"selectAllOnBoard" : "Select All On Board" ,
2024-06-26 18:04:10 +00:00
"showArchivedBoards" : "Show Archived Boards" ,
2024-05-31 09:20:30 +00:00
"selectForCompare" : "Select for Compare" ,
"selectAnImageToCompare" : "Select an Image to Compare" ,
"slider" : "Slider" ,
"sideBySide" : "Side-by-Side" ,
2024-06-01 22:52:32 +00:00
"hover" : "Hover" ,
2024-05-31 09:46:23 +00:00
"swapImages" : "Swap Images" ,
"compareOptions" : "Comparison Options" ,
2024-05-31 23:42:04 +00:00
"stretchToFit" : "Stretch to Fit" ,
2024-06-02 05:20:47 +00:00
"exitCompare" : "Exit Compare" ,
"compareHelp1" : "Hold <Kbd>Alt</Kbd> while clicking a gallery image or using the arrow keys to change the compare image." ,
"compareHelp2" : "Press <Kbd>M</Kbd> to cycle through comparison modes." ,
"compareHelp3" : "Press <Kbd>C</Kbd> to swap the compared images." ,
"compareHelp4" : "Press <Kbd>Z</Kbd> or <Kbd>Esc</Kbd> to exit."
2023-02-18 04:23:24 +00:00
} ,
"hotkeys" : {
2024-01-03 10:18:51 +00:00
"searchHotkeys" : "Search Hotkeys" ,
"clearSearch" : "Clear Search" ,
"noHotkeysFound" : "No Hotkeys Found" ,
2023-09-13 07:31:34 +00:00
"acceptStagingImage" : {
"desc" : "Accept Current Staging Area Image" ,
"title" : "Accept Staging Image"
2023-02-18 04:23:24 +00:00
} ,
2023-09-13 07:31:34 +00:00
"addNodes" : {
"desc" : "Opens the add node menu" ,
"title" : "Add Nodes"
2023-02-18 04:23:24 +00:00
} ,
2024-01-03 10:18:51 +00:00
"appHotkeys" : "App" ,
2023-09-13 07:31:34 +00:00
"cancel" : {
2024-01-03 01:58:43 +00:00
"desc" : "Cancel current queue item" ,
2023-09-13 07:31:34 +00:00
"title" : "Cancel"
2023-02-18 04:23:24 +00:00
} ,
2024-01-03 01:58:43 +00:00
"cancelAndClear" : {
"desc" : "Cancel current queue item and clear all pending items" ,
"title" : "Cancel and Clear"
} ,
2023-09-13 07:31:34 +00:00
"changeTabs" : {
"desc" : "Switch to another workspace" ,
"title" : "Change Tabs"
2023-02-18 04:23:24 +00:00
} ,
2023-09-13 07:31:34 +00:00
"clearMask" : {
"desc" : "Clear the entire mask" ,
"title" : "Clear Mask"
2023-02-18 04:23:24 +00:00
} ,
2023-09-13 07:31:34 +00:00
"closePanels" : {
"desc" : "Closes open panels" ,
"title" : "Close Panels"
2023-02-18 04:23:24 +00:00
} ,
2023-09-13 07:31:34 +00:00
"colorPicker" : {
"desc" : "Selects the canvas color picker" ,
"title" : "Select Color Picker"
2023-02-18 04:23:24 +00:00
} ,
"consoleToggle" : {
2023-09-13 07:31:34 +00:00
"desc" : "Open and close console" ,
"title" : "Console Toggle"
2023-02-18 04:23:24 +00:00
} ,
2023-09-13 07:31:34 +00:00
"copyToClipboard" : {
"desc" : "Copy current canvas to clipboard" ,
"title" : "Copy to Clipboard"
2023-02-18 04:23:24 +00:00
} ,
2023-09-13 07:31:34 +00:00
"decreaseBrushOpacity" : {
"desc" : "Decreases the opacity of the canvas brush" ,
"title" : "Decrease Brush Opacity"
2023-02-18 04:23:24 +00:00
} ,
2023-09-13 07:31:34 +00:00
"decreaseBrushSize" : {
"desc" : "Decreases the size of the canvas brush/eraser" ,
"title" : "Decrease Brush Size"
2023-02-18 04:23:24 +00:00
} ,
2023-09-13 07:31:34 +00:00
"decreaseGalleryThumbSize" : {
"desc" : "Decreases gallery thumbnails size" ,
"title" : "Decrease Gallery Image Size"
2023-02-18 04:23:24 +00:00
} ,
2023-09-13 07:31:34 +00:00
"deleteImage" : {
"desc" : "Delete the current image" ,
"title" : "Delete Image"
2023-02-18 04:23:24 +00:00
} ,
2023-09-13 07:31:34 +00:00
"downloadImage" : {
"desc" : "Download current canvas" ,
"title" : "Download Image"
2023-02-18 04:23:24 +00:00
} ,
2023-09-13 07:31:34 +00:00
"eraseBoundingBox" : {
"desc" : "Erases the bounding box area" ,
"title" : "Erase Bounding Box"
2023-02-18 04:23:24 +00:00
} ,
2023-09-13 07:31:34 +00:00
"fillBoundingBox" : {
"desc" : "Fills the bounding box with brush color" ,
"title" : "Fill Bounding Box"
2023-02-18 04:23:24 +00:00
} ,
2023-09-13 07:31:34 +00:00
"focusPrompt" : {
"desc" : "Focus the prompt input area" ,
"title" : "Focus Prompt"
2023-02-18 04:23:24 +00:00
} ,
2024-01-03 10:18:51 +00:00
"galleryHotkeys" : "Gallery" ,
"generalHotkeys" : "General" ,
2023-09-13 07:31:34 +00:00
"hideMask" : {
"desc" : "Hide and unhide mask" ,
"title" : "Hide Mask"
2023-02-18 04:23:24 +00:00
} ,
2023-09-13 07:31:34 +00:00
"increaseBrushOpacity" : {
"desc" : "Increases the opacity of the canvas brush" ,
"title" : "Increase Brush Opacity"
2023-02-18 04:23:24 +00:00
} ,
2023-09-13 07:31:34 +00:00
"increaseBrushSize" : {
"desc" : "Increases the size of the canvas brush/eraser" ,
"title" : "Increase Brush Size"
2023-02-18 04:23:24 +00:00
} ,
"increaseGalleryThumbSize" : {
2023-09-13 07:31:34 +00:00
"desc" : "Increases gallery thumbnails size" ,
"title" : "Increase Gallery Image Size"
2023-02-18 04:23:24 +00:00
} ,
2023-09-13 07:31:34 +00:00
"invoke" : {
"desc" : "Generate an image" ,
"title" : "Invoke"
2023-02-18 04:23:24 +00:00
} ,
2024-01-03 10:18:51 +00:00
"keyboardShortcuts" : "Hotkeys" ,
2023-09-13 07:31:34 +00:00
"maximizeWorkSpace" : {
"desc" : "Close panels and maximize work area" ,
"title" : "Maximize Workspace"
2023-02-18 04:23:24 +00:00
} ,
2023-09-13 07:31:34 +00:00
"mergeVisible" : {
"desc" : "Merge all visible layers of canvas" ,
"title" : "Merge Visible"
2023-02-18 04:23:24 +00:00
} ,
2023-09-13 07:31:34 +00:00
"moveTool" : {
"desc" : "Allows canvas navigation" ,
"title" : "Move Tool"
2023-02-18 04:23:24 +00:00
} ,
2023-09-13 07:31:34 +00:00
"nextImage" : {
"desc" : "Display the next image in gallery" ,
"title" : "Next Image"
2023-02-18 04:23:24 +00:00
} ,
2023-09-13 07:31:34 +00:00
"nextStagingImage" : {
"desc" : "Next Staging Area Image" ,
"title" : "Next Staging Image"
2023-02-18 04:23:24 +00:00
} ,
2024-01-03 10:18:51 +00:00
"nodesHotkeys" : "Nodes" ,
2023-09-13 07:31:34 +00:00
"pinOptions" : {
"desc" : "Pin the options panel" ,
"title" : "Pin Options"
2023-02-18 04:23:24 +00:00
} ,
2023-09-13 07:31:34 +00:00
"previousImage" : {
"desc" : "Display the previous image in gallery" ,
"title" : "Previous Image"
2023-02-18 04:23:24 +00:00
} ,
2023-09-13 07:31:34 +00:00
"previousStagingImage" : {
"desc" : "Previous Staging Area Image" ,
"title" : "Previous Staging Image"
2023-02-18 04:23:24 +00:00
} ,
2023-09-13 07:31:34 +00:00
"quickToggleMove" : {
"desc" : "Temporarily toggles Move mode" ,
"title" : "Quick Toggle Move"
2023-02-18 04:23:24 +00:00
} ,
2023-09-13 07:31:34 +00:00
"redoStroke" : {
"desc" : "Redo a brush stroke" ,
"title" : "Redo Stroke"
2023-02-18 04:23:24 +00:00
} ,
2023-09-13 07:31:34 +00:00
"resetView" : {
"desc" : "Reset Canvas View" ,
"title" : "Reset View"
2023-02-18 04:23:24 +00:00
} ,
2023-09-13 07:31:34 +00:00
"restoreFaces" : {
"desc" : "Restore the current image" ,
"title" : "Restore Faces"
2023-02-18 04:23:24 +00:00
} ,
2023-09-13 07:31:34 +00:00
"saveToGallery" : {
"desc" : "Save current canvas to gallery" ,
"title" : "Save To Gallery"
2023-02-18 04:23:24 +00:00
} ,
2023-09-13 07:31:34 +00:00
"selectBrush" : {
"desc" : "Selects the canvas brush" ,
"title" : "Select Brush"
2023-02-18 04:23:24 +00:00
} ,
2023-09-13 07:31:34 +00:00
"selectEraser" : {
"desc" : "Selects the canvas eraser" ,
"title" : "Select Eraser"
2023-02-18 04:23:24 +00:00
} ,
2023-09-13 07:31:34 +00:00
"sendToImageToImage" : {
"desc" : "Send current image to Image to Image" ,
"title" : "Send To Image To Image"
2023-02-18 04:23:24 +00:00
} ,
2024-01-24 00:45:30 +00:00
"remixImage" : {
"desc" : "Use all parameters except seed from the current image" ,
"title" : "Remix image"
} ,
2023-09-13 07:31:34 +00:00
"setParameters" : {
"desc" : "Use all parameters of the current image" ,
"title" : "Set Parameters"
2023-02-18 04:23:24 +00:00
} ,
2023-09-13 07:31:34 +00:00
"setPrompt" : {
"desc" : "Use the prompt of the current image" ,
"title" : "Set Prompt"
2023-02-18 04:23:24 +00:00
} ,
2023-09-13 07:31:34 +00:00
"setSeed" : {
"desc" : "Use the seed of the current image" ,
"title" : "Set Seed"
2023-02-18 04:23:24 +00:00
} ,
2023-09-13 07:31:34 +00:00
"showHideBoundingBox" : {
"desc" : "Toggle visibility of bounding box" ,
"title" : "Show/Hide Bounding Box"
2023-02-18 04:23:24 +00:00
} ,
2023-09-13 07:31:34 +00:00
"showInfo" : {
"desc" : "Show metadata info of the current image" ,
"title" : "Show Info"
2023-02-18 04:23:24 +00:00
} ,
2023-09-13 07:31:34 +00:00
"toggleGallery" : {
"desc" : "Open and close the gallery drawer" ,
"title" : "Toggle Gallery"
2023-02-18 04:23:24 +00:00
} ,
2024-01-03 10:18:51 +00:00
"toggleOptions" : {
"desc" : "Open and close the options panel" ,
"title" : "Toggle Options"
} ,
"toggleOptionsAndGallery" : {
"desc" : "Open and close the options and gallery panels" ,
"title" : "Toggle Options and Gallery"
} ,
"resetOptionsAndGallery" : {
"desc" : "Resets the options and gallery panels" ,
"title" : "Reset Options and Gallery"
2023-02-18 04:23:24 +00:00
} ,
2023-09-13 07:31:34 +00:00
"toggleLayer" : {
"desc" : "Toggles mask/base layer selection" ,
"title" : "Toggle Layer"
2023-02-18 04:23:24 +00:00
} ,
2023-09-13 07:31:34 +00:00
"toggleSnap" : {
"desc" : "Toggles Snap to Grid" ,
"title" : "Toggle Snap"
2023-08-21 15:31:29 +00:00
} ,
2023-09-13 07:31:34 +00:00
"undoStroke" : {
"desc" : "Undo a brush stroke" ,
"title" : "Undo Stroke"
} ,
2024-01-03 10:18:51 +00:00
"unifiedCanvasHotkeys" : "Unified Canvas" ,
2024-07-23 21:32:33 +00:00
"postProcess" : {
"desc" : "Process the current image using the selected post-processing model" ,
"title" : "Process Image"
2024-05-02 12:29:00 +00:00
} ,
2024-05-06 05:35:12 +00:00
"toggleViewer" : {
"desc" : "Switches between the Image Viewer and workspace for the current tab." ,
"title" : "Toggle Image Viewer"
2023-02-18 04:23:24 +00:00
}
} ,
2023-09-13 07:31:34 +00:00
"metadata" : {
2024-02-23 14:25:58 +00:00
"allPrompts" : "All Prompts" ,
2023-09-13 07:31:34 +00:00
"cfgScale" : "CFG scale" ,
2023-11-30 09:55:20 +00:00
"cfgRescaleMultiplier" : "$t(parameters.cfgRescaleMultiplier)" ,
2023-09-13 07:31:34 +00:00
"createdBy" : "Created By" ,
"fit" : "Image to image fit" ,
"generationMode" : "Generation Mode" ,
"height" : "Height" ,
"imageDetails" : "Image Details" ,
2024-02-23 14:25:58 +00:00
"imageDimensions" : "Image Dimensions" ,
2023-09-13 07:31:34 +00:00
"initImage" : "Initial image" ,
"metadata" : "Metadata" ,
2023-02-18 04:23:24 +00:00
"model" : "Model" ,
2023-09-13 07:31:34 +00:00
"negativePrompt" : "Negative Prompt" ,
"noImageDetails" : "No image details found" ,
"noMetaData" : "No metadata found" ,
2023-10-14 10:25:07 +00:00
"noRecallParameters" : "No parameters to recall found" ,
2024-02-23 14:25:58 +00:00
"parameterSet" : "Parameter {{parameter}} set" ,
2024-02-27 04:36:00 +00:00
"parsingFailed" : "Parsing Failed" ,
2023-09-13 07:31:34 +00:00
"positivePrompt" : "Positive Prompt" ,
2023-10-14 10:25:07 +00:00
"recallParameters" : "Recall Parameters" ,
2024-02-23 14:25:58 +00:00
"recallParameter" : "Recall {{label}}" ,
2023-09-13 07:31:34 +00:00
"scheduler" : "Scheduler" ,
"seamless" : "Seamless" ,
"seed" : "Seed" ,
"steps" : "Steps" ,
"strength" : "Image to image strength" ,
"Threshold" : "Noise Threshold" ,
2023-11-20 06:24:03 +00:00
"vae" : "VAE" ,
2023-09-13 07:31:34 +00:00
"width" : "Width" ,
"workflow" : "Workflow"
} ,
"modelManager" : {
"active" : "active" ,
2023-02-18 04:23:24 +00:00
"addModel" : "Add Model" ,
2024-02-26 19:40:38 +00:00
"addModels" : "Add Models" ,
2023-09-13 07:31:34 +00:00
"advanced" : "Advanced" ,
"allModels" : "All Models" ,
"alpha" : "Alpha" ,
2023-02-18 04:23:24 +00:00
"availableModels" : "Available Models" ,
2023-09-13 07:31:34 +00:00
"baseModel" : "Base Model" ,
2024-02-20 18:39:27 +00:00
"cancel" : "Cancel" ,
2024-08-27 14:15:02 +00:00
"clipEmbed" : "CLIP Embed" ,
"clipVision" : "CLIP Vision" ,
2023-09-13 07:31:34 +00:00
"config" : "Config" ,
2023-02-18 04:23:24 +00:00
"convert" : "Convert" ,
2023-09-13 07:31:34 +00:00
"convertingModelBegin" : "Converting Model. Please wait." ,
2023-02-18 04:23:24 +00:00
"convertToDiffusers" : "Convert To Diffusers" ,
2024-03-13 11:37:47 +00:00
"convertToDiffusersHelpText1" : "This model will be converted to the \ud83e\udde8 Diffusers format." ,
2023-02-18 04:23:24 +00:00
"convertToDiffusersHelpText2" : "This process will replace your Model Manager entry with the Diffusers version of the same model." ,
2023-07-15 06:48:18 +00:00
"convertToDiffusersHelpText3" : "Your checkpoint file on disk WILL be deleted if it is in InvokeAI root folder. If it is in a custom location, then it WILL NOT be deleted." ,
2023-02-18 04:23:24 +00:00
"convertToDiffusersHelpText4" : "This is a one time process only. It might take around 30s-60s depending on the specifications of your computer." ,
2023-07-12 08:40:58 +00:00
"convertToDiffusersHelpText5" : "Please make sure you have enough disk space. Models generally vary between 2GB-7GB in size." ,
2023-02-18 04:23:24 +00:00
"convertToDiffusersHelpText6" : "Do you wish to convert this model?" ,
2024-03-04 14:39:03 +00:00
"defaultSettings" : "Default Settings" ,
"defaultSettingsSaved" : "Default Settings Saved" ,
2023-09-13 07:31:34 +00:00
"delete" : "Delete" ,
"deleteConfig" : "Delete Config" ,
"deleteModel" : "Delete Model" ,
2024-03-06 18:15:33 +00:00
"deleteModelImage" : "Delete Model Image" ,
2023-09-13 07:31:34 +00:00
"deleteMsg1" : "Are you sure you want to delete this model from InvokeAI?" ,
"deleteMsg2" : "This WILL delete the model from disk if it is in the InvokeAI root folder. If you are using a custom location, then the model WILL NOT be deleted from disk." ,
"description" : "Description" ,
2024-02-26 19:40:38 +00:00
"edit" : "Edit" ,
2023-09-13 07:31:34 +00:00
"height" : "Height" ,
2024-03-13 08:58:19 +00:00
"huggingFace" : "HuggingFace" ,
2024-03-11 14:48:30 +00:00
"huggingFacePlaceholder" : "owner/model-name" ,
2024-03-13 08:58:19 +00:00
"huggingFaceRepoID" : "HuggingFace Repo ID" ,
"huggingFaceHelper" : "If multiple models are found in this repo, you will be prompted to select one to install." ,
2024-03-19 22:24:10 +00:00
"hfToken" : "HuggingFace Token" ,
"hfTokenHelperText" : "A HF token is required to use checkpoint models. Click here to create or get your token." ,
"hfTokenInvalid" : "Invalid or Missing HF Token" ,
"hfTokenInvalidErrorMessage" : "Invalid or missing HuggingFace token." ,
"hfTokenInvalidErrorMessage2" : "Update it in the " ,
"hfTokenUnableToVerify" : "Unable to Verify HF Token" ,
"hfTokenUnableToVerifyErrorMessage" : "Unable to verify HuggingFace token. This is likely due to a network error. Please try again later." ,
"hfTokenSaved" : "HF Token Saved" ,
2024-02-26 19:40:38 +00:00
"imageEncoderModelId" : "Image Encoder Model ID" ,
2024-03-13 09:50:23 +00:00
"installQueue" : "Install Queue" ,
2024-03-08 03:25:58 +00:00
"inplaceInstall" : "In-place install" ,
"inplaceInstallDesc" : "Install models without copying the files. When using the model, it will be loaded from its this location. If disabled, the model file(s) will be copied into the Invoke-managed models directory during installation." ,
2024-03-13 08:58:19 +00:00
"install" : "Install" ,
"installAll" : "Install All" ,
"installRepo" : "Install Repo" ,
2024-03-23 20:57:38 +00:00
"ipAdapters" : "IP Adapters" ,
2023-09-13 07:31:34 +00:00
"load" : "Load" ,
2024-03-07 20:13:04 +00:00
"localOnly" : "local only" ,
2023-09-13 07:31:34 +00:00
"manual" : "Manual" ,
2024-03-15 16:31:36 +00:00
"loraModels" : "LoRAs" ,
"main" : "Main" ,
2024-03-04 14:39:03 +00:00
"metadata" : "Metadata" ,
2023-09-13 07:31:34 +00:00
"model" : "Model" ,
"modelConversionFailed" : "Model Conversion Failed" ,
"modelConverted" : "Model Converted" ,
"modelDeleted" : "Model Deleted" ,
"modelDeleteFailed" : "Failed to delete model" ,
2024-03-06 18:15:33 +00:00
"modelImageDeleted" : "Model Image Deleted" ,
"modelImageDeleteFailed" : "Model Image Delete Failed" ,
"modelImageUpdated" : "Model Image Updated" ,
"modelImageUpdateFailed" : "Model Image Update Failed" ,
2023-09-13 07:31:34 +00:00
"modelManager" : "Model Manager" ,
2024-02-26 19:40:38 +00:00
"modelName" : "Model Name" ,
"modelSettings" : "Model Settings" ,
2023-09-13 07:31:34 +00:00
"modelsSynced" : "Models Synced" ,
"modelSyncFailed" : "Model Sync Failed" ,
"modelType" : "Model Type" ,
"modelUpdated" : "Model Updated" ,
"modelUpdateFailed" : "Model Update Failed" ,
"name" : "Name" ,
2024-03-19 05:01:04 +00:00
"noModelsInstalled" : "No Models Installed" ,
"noModelsInstalledDesc1" : "Install models with the" ,
2023-11-20 06:24:03 +00:00
"noModelSelected" : "No Model Selected" ,
2024-04-04 22:21:13 +00:00
"noMatchingModels" : "No matching Models" ,
2023-03-06 09:02:40 +00:00
"none" : "none" ,
2024-02-26 19:40:38 +00:00
"path" : "Path" ,
2024-03-06 19:08:19 +00:00
"pathToConfig" : "Path To Config" ,
2024-02-26 19:40:38 +00:00
"predictionType" : "Prediction Type" ,
2024-02-21 22:17:34 +00:00
"prune" : "Prune" ,
"pruneTooltip" : "Prune finished imports from queue" ,
2023-09-13 07:31:34 +00:00
"repo_id" : "Repo ID" ,
2024-02-26 19:40:38 +00:00
"repoVariant" : "Repo Variant" ,
2024-03-13 09:50:23 +00:00
"scanFolder" : "Scan Folder" ,
"scanFolderHelper" : "The folder will be recursively scanned for models. This can take a few moments for very large folders." ,
2024-03-11 14:48:30 +00:00
"scanPlaceholder" : "Path to a local folder" ,
2024-02-26 19:40:38 +00:00
"scanResults" : "Scan Results" ,
2023-09-13 07:31:34 +00:00
"search" : "Search" ,
"selected" : "Selected" ,
2023-07-15 13:36:00 +00:00
"selectModel" : "Select Model" ,
2023-07-20 06:50:43 +00:00
"settings" : "Settings" ,
2024-03-11 15:46:49 +00:00
"simpleModelPlaceholder" : "URL or path to a local file or diffusers folder" ,
2024-02-26 19:40:38 +00:00
"source" : "Source" ,
2024-08-27 13:46:34 +00:00
"spandrelImageToImage" : "Image to Image (Spandrel)" ,
2024-03-19 05:02:19 +00:00
"starterModels" : "Starter Models" ,
2024-08-21 18:35:39 +00:00
"starterModelsInModelManager" : "Starter Models can be found in Model Manager" ,
2023-07-20 06:50:43 +00:00
"syncModels" : "Sync Models" ,
2024-03-15 16:31:36 +00:00
"textualInversions" : "Textual Inversions" ,
2024-03-07 04:36:45 +00:00
"triggerPhrases" : "Trigger Phrases" ,
2024-03-07 04:24:28 +00:00
"loraTriggerPhrases" : "LoRA Trigger Phrases" ,
"mainModelTriggerPhrases" : "Main Model Trigger Phrases" ,
2024-03-04 14:39:03 +00:00
"typePhraseHere" : "Type phrase here" ,
2024-08-27 13:46:34 +00:00
"t5Encoder" : "T5 Encoder" ,
2024-02-26 19:40:38 +00:00
"upcastAttention" : "Upcast Attention" ,
2024-03-06 19:05:27 +00:00
"uploadImage" : "Upload Image" ,
2024-03-13 09:50:23 +00:00
"urlOrLocalPath" : "URL or Local Path" ,
"urlOrLocalPathHelper" : "URLs should point to a single file. Local paths can point to a single file or folder for a single diffusers model." ,
2024-03-04 14:39:03 +00:00
"useDefaultSettings" : "Use Default Settings" ,
2023-09-13 07:31:34 +00:00
"v2_768" : "v2 (768px)" ,
"v2_base" : "v2 (512px)" ,
"vae" : "VAE" ,
2023-10-12 11:46:47 +00:00
"vaePrecision" : "VAE Precision" ,
2023-09-13 07:31:34 +00:00
"variant" : "Variant" ,
2024-03-13 11:37:47 +00:00
"width" : "Width"
2023-09-13 07:31:34 +00:00
} ,
"models" : {
2023-11-20 06:24:03 +00:00
"addLora" : "Add LoRA" ,
2024-03-01 15:37:30 +00:00
"concepts" : "Concepts" ,
2023-11-20 06:24:03 +00:00
"esrganModel" : "ESRGAN Model" ,
2023-09-13 07:31:34 +00:00
"loading" : "loading" ,
"noMatchingLoRAs" : "No matching LoRAs" ,
"noMatchingModels" : "No matching Models" ,
2023-10-03 21:37:51 +00:00
"noModelsAvailable" : "No models available" ,
2023-12-28 13:03:21 +00:00
"lora" : "LoRA" ,
2023-11-02 01:27:28 +00:00
"selectModel" : "Select a Model" ,
"noLoRAsInstalled" : "No LoRAs installed" ,
2023-12-28 13:03:21 +00:00
"noRefinerModelsInstalled" : "No SDXL Refiner models installed" ,
"defaultVAE" : "Default VAE"
2023-09-13 07:31:34 +00:00
} ,
"nodes" : {
"addNode" : "Add Node" ,
"addNodeToolTip" : "Add Node (Shift+A, Space)" ,
2023-11-26 09:44:39 +00:00
"addLinearView" : "Add to Linear View" ,
2023-09-13 07:31:34 +00:00
"animatedEdges" : "Animated Edges" ,
"animatedEdgesHelp" : "Animate selected edges and edges connected to selected nodes" ,
2023-09-13 11:15:36 +00:00
"boolean" : "Booleans" ,
2023-09-13 07:31:34 +00:00
"cannotConnectInputToInput" : "Cannot connect input to input" ,
"cannotConnectOutputToOutput" : "Cannot connect output to output" ,
"cannotConnectToSelf" : "Cannot connect to self" ,
2023-09-26 15:54:35 +00:00
"cannotDuplicateConnection" : "Cannot create duplicate connections" ,
2024-05-16 13:42:54 +00:00
"cannotMixAndMatchCollectionItemTypes" : "Cannot mix and match collection item types" ,
2024-05-18 14:11:15 +00:00
"missingNode" : "Missing invocation node" ,
"missingInvocationTemplate" : "Missing invocation template" ,
"missingFieldTemplate" : "Missing field template" ,
2023-11-27 01:51:19 +00:00
"nodePack" : "Node pack" ,
2023-09-13 11:15:36 +00:00
"collection" : "Collection" ,
2024-05-19 13:24:16 +00:00
"singleFieldType" : "{{name}} (Single)" ,
"collectionFieldType" : "{{name}} (Collection)" ,
"collectionOrScalarFieldType" : "{{name}} (Single or Collection)" ,
2023-09-13 07:31:34 +00:00
"colorCodeEdges" : "Color-Code Edges" ,
"colorCodeEdgesHelp" : "Color-code edges according to their connected fields" ,
"connectionWouldCreateCycle" : "Connection would create a cycle" ,
"currentImage" : "Current Image" ,
"currentImageDescription" : "Displays the current image in the Node Editor" ,
"downloadWorkflow" : "Download Workflow JSON" ,
2023-09-13 11:15:36 +00:00
"edge" : "Edge" ,
2024-03-06 16:19:03 +00:00
"edit" : "Edit" ,
2024-02-14 14:02:07 +00:00
"editMode" : "Edit in Workflow Editor" ,
2023-09-13 11:15:36 +00:00
"enum" : "Enum" ,
"executionStateCompleted" : "Completed" ,
"executionStateError" : "Error" ,
"executionStateInProgress" : "In Progress" ,
2023-09-13 07:31:34 +00:00
"fieldTypesMustMatch" : "Field types must match" ,
"fitViewportNodes" : "Fit View" ,
2023-09-13 11:15:36 +00:00
"float" : "Float" ,
2023-09-13 07:31:34 +00:00
"fullyContainNodes" : "Fully Contain Nodes to Select" ,
"fullyContainNodesHelp" : "Nodes must be fully inside the selection box to be selected" ,
2024-04-15 09:26:29 +00:00
"showEdgeLabels" : "Show Edge Labels" ,
"showEdgeLabelsHelp" : "Show labels on edges, indicating the connected nodes" ,
2023-09-13 07:31:34 +00:00
"hideLegendNodes" : "Hide Field Type Legend" ,
"hideMinimapnodes" : "Hide MiniMap" ,
"inputMayOnlyHaveOneConnection" : "Input may only have one connection" ,
2023-09-13 11:15:36 +00:00
"integer" : "Integer" ,
2023-10-06 21:15:13 +00:00
"ipAdapter" : "IP-Adapter" ,
2023-09-13 07:31:34 +00:00
"loadingNodes" : "Loading Nodes..." ,
"loadWorkflow" : "Load Workflow" ,
feat: queued generation (#4502)
* fix(config): fix typing issues in `config/`
`config/invokeai_config.py`:
- use `Optional` for things that are optional
- fix typing of `ram_cache_size()` and `vram_cache_size()`
- remove unused and incorrectly typed method `autoconvert_path`
- fix types and logic for `parse_args()`, in which `InvokeAIAppConfig.initconf` *must* be a `DictConfig`, but function would allow it to be set as a `ListConfig`, which presumably would cause issues elsewhere
`config/base.py`:
- use `cls` for first arg of class methods
- use `Optional` for things that are optional
- fix minor type issue related to setting of `env_prefix`
- remove unused `add_subparser()` method, which calls `add_parser()` on an `ArgumentParser` (method only available on the `_SubParsersAction` object, which is returned from ArgumentParser.add_subparsers()`)
* feat: queued generation and batches
Due to a very messy branch with broad addition of `isort` on `main` alongside it, some git surgery was needed to get an agreeable git history. This commit represents all of the work on queued generation. See PR for notes.
* chore: flake8, isort, black
* fix(nodes): fix incorrect service stop() method
* fix(nodes): improve names of a few variables
* fix(tests): fix up tests after changes to batches/queue
* feat(tests): add unit tests for session queue helper functions
* feat(ui): dynamic prompts is always enabled
* feat(queue): add queue_status_changed event
* feat(ui): wip queue graphs
* feat(nodes): move cleanup til after invoker startup
* feat(nodes): add cancel_by_batch_ids
* feat(ui): wip batch graphs & UI
* fix(nodes): remove `Batch.batch_id` from required
* fix(ui): cleanup and use fixedCacheKey for all mutations
* fix(ui): remove orphaned nodes from canvas graphs
* fix(nodes): fix cancel_by_batch_ids result count
* fix(ui): only show cancel batch tooltip when batches were canceled
* chore: isort
* fix(api): return `[""]` when dynamic prompts generates no prompts
Just a simple fallback so we always have a prompt.
* feat(ui): dynamicPrompts.combinatorial is always on
There seems to be little purpose in using the combinatorial generation for dynamic prompts. I've disabled it by hiding it from the UI and defaulting combinatorial to true. If we want to enable it again in the future it's straightforward to do so.
* feat: add queue_id & support logic
* feat(ui): fix upscale button
It prepends the upscale operation to queue
* feat(nodes): return queue item when enqueuing a single graph
This facilitates one-off graph async workflows in the client.
* feat(ui): move controlnet autoprocess to queue
* fix(ui): fix non-serializable DOMRect in redux state
* feat(ui): QueueTable performance tweaks
* feat(ui): update queue list
Queue items expand to show the full queue item. Just as JSON for now.
* wip threaded session_processor
* feat(nodes,ui): fully migrate queue to session_processor
* feat(nodes,ui): add processor events
* feat(ui): ui tweaks
* feat(nodes,ui): consolidate events, reduce network requests
* feat(ui): cleanup & abstract queue hooks
* feat(nodes): optimize batch permutation
Use a generator to do only as much work as is needed.
Previously, though we only ended up creating exactly as many queue items as was needed, there was still some intermediary work that calculated *all* permutations. When that number was very high, the system had a very hard time and used a lot of memory.
The logic has been refactored to use a generator. Additionally, the batch validators are optimized to return early and use less memory.
* feat(ui): add seed behaviour parameter
This dynamic prompts parameter allows the seed to be randomized per prompt or per iteration:
- Per iteration: Use the same seed for all prompts in a single dynamic prompt expansion
- Per prompt: Use a different seed for every single prompt
"Per iteration" is appropriate for exploring a the latents space with a stable starting noise, while "Per prompt" provides more variation.
* fix(ui): remove extraneous random seed nodes from linear graphs
* fix(ui): fix controlnet autoprocess not working when queue is running
* feat(queue): add timestamps to queue status updates
Also show execution time in queue list
* feat(queue): change all execution-related events to use the `queue_id` as the room, also include `queue_item_id` in InvocationQueueItem
This allows for much simpler handling of queue items.
* feat(api): deprecate sessions router
* chore(backend): tidy logging in `dependencies.py`
* fix(backend): respect `use_memory_db`
* feat(backend): add `config.log_sql` (enables sql trace logging)
* feat: add invocation cache
Supersedes #4574
The invocation cache provides simple node memoization functionality. Nodes that use the cache are memoized and not re-executed if their inputs haven't changed. Instead, the stored output is returned.
## Results
This feature provides anywhere some significant to massive performance improvement.
The improvement is most marked on large batches of generations where you only change a couple things (e.g. different seed or prompt for each iteration) and low-VRAM systems, where skipping an extraneous model load is a big deal.
## Overview
A new `invocation_cache` service is added to handle the caching. There's not much to it.
All nodes now inherit a boolean `use_cache` field from `BaseInvocation`. This is a node field and not a class attribute, because specific instances of nodes may want to opt in or out of caching.
The recently-added `invoke_internal()` method on `BaseInvocation` is used as an entrypoint for the cache logic.
To create a cache key, the invocation is first serialized using pydantic's provided `json()` method, skipping the unique `id` field. Then python's very fast builtin `hash()` is used to create an integer key. All implementations of `InvocationCacheBase` must provide a class method `create_key()` which accepts an invocation and outputs a string or integer key.
## In-Memory Implementation
An in-memory implementation is provided. In this implementation, the node outputs are stored in memory as python classes. The in-memory cache does not persist application restarts.
Max node cache size is added as `node_cache_size` under the `Generation` config category.
It defaults to 512 - this number is up for discussion, but given that these are relatively lightweight pydantic models, I think it's safe to up this even higher.
Note that the cache isn't storing the big stuff - tensors and images are store on disk, and outputs include only references to them.
## Node Definition
The default for all nodes is to use the cache. The `@invocation` decorator now accepts an optional `use_cache: bool` argument to override the default of `True`.
Non-deterministic nodes, however, should set this to `False`. Currently, all random-stuff nodes, including `dynamic_prompt`, are set to `False`.
The field name `use_cache` is now effectively a reserved field name and possibly a breaking change if any community nodes use this as a field name. In hindsight, all our reserved field names should have been prefixed with underscores or something.
## One Gotcha
Leaf nodes probably want to opt out of the cache, because if they are not cached, their outputs are not saved again.
If you run the same graph multiple times, you only end up with a single image output, because the image storage side-effects are in the `invoke()` method, which is bypassed if we have a cache hit.
## Linear UI
The linear graphs _almost_ just work, but due to the gotcha, we need to be careful about the final image-outputting node. To resolve this, a `SaveImageInvocation` node is added and used in the linear graphs.
This node is similar to `ImagePrimitive`, except it saves a copy of its input image, and has `use_cache` set to `False` by default.
This is now the leaf node in all linear graphs, and is the only node in those graphs with `use_cache == False` _and_ the only node with `is_intermedate == False`.
## Workflow Editor
All nodes now have a footer with a new `Use Cache [ ]` checkbox. It defaults to the value set by the invocation in its python definition, but can be changed by the user.
The workflow/node validation logic has been updated to migrate old workflows to use the new default values for `use_cache`. Users may still want to review the settings that have been chosen. In the event of catastrophic failure when running this migration, the default value of `True` is applied, as this is correct for most nodes.
Users should consider saving their workflows after loading them in and having them updated.
## Future Enhancements - Callback
A future enhancement would be to provide a callback to the `use_cache` flag that would be run as the node is executed to determine, based on its own internal state, if the cache should be used or not.
This would be useful for `DynamicPromptInvocation`, where the deterministic behaviour is determined by the `combinatorial: bool` field.
## Future Enhancements - Persisted Cache
Similar to how the latents storage is backed by disk, the invocation cache could be persisted to the database or disk. We'd need to be very careful about deserializing outputs, but it's perhaps worth exploring in the future.
* fix(ui): fix queue list item width
* feat(nodes): do not send the whole node on every generator progress
* feat(ui): strip out old logic related to sessions
Things like `isProcessing` are no longer relevant with queue. Removed them all & updated everything be appropriate for queue. May be a few little quirks I've missed...
* feat(ui): fix up param collapse labels
* feat(ui): click queue count to go to queue tab
* tidy(queue): update comment, query format
* feat(ui): fix progress bar when canceling
* fix(ui): fix circular dependency
* feat(nodes): bail on node caching logic if `node_cache_size == 0`
* feat(nodes): handle KeyError on node cache pop
* feat(nodes): bypass cache codepath if caches is disabled
more better no do thing
* fix(ui): reset api cache on connect/disconnect
* feat(ui): prevent enqueue when no prompts generated
* feat(ui): add queue controls to workflow editor
* feat(ui): update floating buttons & other incidental UI tweaks
* fix(ui): fix missing/incorrect translation keys
* fix(tests): add config service to mock invocation services
invoking needs access to `node_cache_size` to occur
* optionally remove pause/resume buttons from queue UI
* option to disable prepending
* chore(ui): remove unused file
* feat(queue): remove `order_id` entirely, `item_id` is now an autoinc pk
---------
Co-authored-by: Mary Hipp <maryhipp@Marys-MacBook-Air.local>
2023-09-20 05:09:24 +00:00
"noWorkflow" : "No Workflow" ,
feat(ui): add support for custom field types
Node authors may now create their own arbitrary/custom field types. Any pydantic model is supported.
Two notes:
1. Your field type's class name must be unique.
Suggest prefixing fields with something related to the node pack as a kind of namespace.
2. Custom field types function as connection-only fields.
For example, if your custom field has string attributes, you will not get a text input for that attribute when you give a node a field with your custom type.
This is the same behaviour as other complex fields that don't have custom UIs in the workflow editor - like, say, a string collection.
feat(ui): fix tooltips for custom types
We need to hold onto the original type of the field so they don't all just show up as "Unknown".
fix(ui): fix ts error with custom fields
feat(ui): custom field types connection validation
In the initial commit, a custom field's original type was added to the *field templates* only as `originalType`. Custom fields' `type` property was `"Custom"`*. This allowed for type safety throughout the UI logic.
*Actually, it was `"Unknown"`, but I changed it to custom for clarity.
Connection validation logic, however, uses the *field instance* of the node/field. Like the templates, *field instances* with custom types have their `type` set to `"Custom"`, but they didn't have an `originalType` property. As a result, all custom fields could be connected to all other custom fields.
To resolve this, we need to add `originalType` to the *field instances*, then switch the validation logic to use this instead of `type`.
This ended up needing a bit of fanagling:
- If we make `originalType` a required property on field instances, existing workflows will break during connection validation, because they won't have this property. We'd need a new layer of logic to migrate the workflows, adding the new `originalType` property.
While this layer is probably needed anyways, typing `originalType` as optional is much simpler. Workflow migration logic can come layer.
(Technically, we could remove all references to field types from the workflow files, and let the templates hold all this information. This feels like a significant change and I'm reluctant to do it now.)
- Because `originalType` is optional, anywhere we care about the type of a field, we need to use it over `type`. So there are a number of `field.originalType ?? field.type` expressions. This is a bit of a gotcha, we'll need to remember this in the future.
- We use `Array.prototype.includes()` often in the workflow editor, e.g. `COLLECTION_TYPES.includes(type)`. In these cases, the const array is of type `FieldType[]`, and `type` is is `FieldType`.
Because we now support custom types, the arg `type` is now widened from `FieldType` to `string`.
This causes a TS error. This behaviour is somewhat controversial (see https://github.com/microsoft/TypeScript/issues/14520). These expressions are now rewritten as `COLLECTION_TYPES.some((t) => t === type)` to satisfy TS. It's logically equivalent.
fix(ui): typo
feat(ui): add CustomCollection and CustomPolymorphic field types
feat(ui): add validation for CustomCollection & CustomPolymorphic types
- Update connection validation for custom types
- Use simple string parsing to determine if a field is a collection or polymorphic type.
- No longer need to keep a list of collection and polymorphic types.
- Added runtime checks in `baseinvocation.py` to ensure no fields are named in such a way that it could mess up the new parsing
chore(ui): remove errant console.log
fix(ui): rename 'nodes.currentConnectionFieldType' -> 'nodes.connectionStartFieldType'
This was confusingly named and kept tripping me up. Renamed to be consistent with the `reactflow` `ConnectionStartParams` type.
fix(ui): fix ts error
feat(nodes): add runtime check for custom field names
"Custom", "CustomCollection" and "CustomPolymorphic" are reserved field names.
chore(ui): add TODO for revising field type names
wip refactor fieldtype structured
wip refactor field types
wip refactor types
wip refactor types
fix node layout
refactor field types
chore: mypy
organisation
organisation
organisation
fix(nodes): fix field orig_required, field_kind and input statuses
feat(nodes): remove broken implementation of default_factory on InputField
Use of this could break connection validation due to the difference in node schemas required fields and invoke() required args.
Removed entirely for now. It wasn't ever actually used by the system, because all graphs always had values provided for fields where default_factory was used.
Also, pydantic is smart enough to not reuse the same object when specifying a default value - it clones the object first. So, the common pattern of `default_factory=list` is extraneous. It can just be `default=[]`.
fix(nodes): fix InputField name validation
workflow validation
validation
chore: ruff
feat(nodes): fix up baseinvocation comments
fix(ui): improve typing & logic of buildFieldInputTemplate
improved error handling in parseFieldType
fix: back compat for deprecated default_factory and UIType
feat(nodes): do not show node packs loaded log if none loaded
chore(ui): typegen
2023-11-17 00:32:35 +00:00
"mismatchedVersion" : "Invalid node: node {{node}} of type {{type}} has mismatched version (try updating?)" ,
"missingTemplate" : "Invalid node: node {{node}} of type {{type}} missing template (not installed?)" ,
"sourceNodeDoesNotExist" : "Invalid edge: source/output node {{node}} does not exist" ,
"targetNodeDoesNotExist" : "Invalid edge: target/input node {{node}} does not exist" ,
"sourceNodeFieldDoesNotExist" : "Invalid edge: source/output field {{node}}.{{field}} does not exist" ,
"targetNodeFieldDoesNotExist" : "Invalid edge: target/input field {{node}}.{{field}} does not exist" ,
"deletedInvalidEdge" : "Deleted invalid edge {{source}} -> {{target}}" ,
2023-09-13 07:31:34 +00:00
"noConnectionData" : "No connection data" ,
"noConnectionInProgress" : "No connection in progress" ,
2023-09-13 11:15:36 +00:00
"node" : "Node" ,
2023-09-13 07:31:34 +00:00
"nodeOutputs" : "Node Outputs" ,
"nodeSearch" : "Search for nodes" ,
"nodeTemplate" : "Node Template" ,
2023-09-13 11:15:36 +00:00
"nodeType" : "Node Type" ,
2023-09-13 07:31:34 +00:00
"noFieldsLinearview" : "No fields added to Linear View" ,
2024-03-06 16:19:03 +00:00
"noFieldsViewMode" : "This workflow has no selected fields to display. View the full workflow to configure values." ,
2023-09-13 07:31:34 +00:00
"noFieldType" : "No field type" ,
"noMatchingNodes" : "No matching nodes" ,
"noNodeSelected" : "No node selected" ,
feat: queued generation (#4502)
* fix(config): fix typing issues in `config/`
`config/invokeai_config.py`:
- use `Optional` for things that are optional
- fix typing of `ram_cache_size()` and `vram_cache_size()`
- remove unused and incorrectly typed method `autoconvert_path`
- fix types and logic for `parse_args()`, in which `InvokeAIAppConfig.initconf` *must* be a `DictConfig`, but function would allow it to be set as a `ListConfig`, which presumably would cause issues elsewhere
`config/base.py`:
- use `cls` for first arg of class methods
- use `Optional` for things that are optional
- fix minor type issue related to setting of `env_prefix`
- remove unused `add_subparser()` method, which calls `add_parser()` on an `ArgumentParser` (method only available on the `_SubParsersAction` object, which is returned from ArgumentParser.add_subparsers()`)
* feat: queued generation and batches
Due to a very messy branch with broad addition of `isort` on `main` alongside it, some git surgery was needed to get an agreeable git history. This commit represents all of the work on queued generation. See PR for notes.
* chore: flake8, isort, black
* fix(nodes): fix incorrect service stop() method
* fix(nodes): improve names of a few variables
* fix(tests): fix up tests after changes to batches/queue
* feat(tests): add unit tests for session queue helper functions
* feat(ui): dynamic prompts is always enabled
* feat(queue): add queue_status_changed event
* feat(ui): wip queue graphs
* feat(nodes): move cleanup til after invoker startup
* feat(nodes): add cancel_by_batch_ids
* feat(ui): wip batch graphs & UI
* fix(nodes): remove `Batch.batch_id` from required
* fix(ui): cleanup and use fixedCacheKey for all mutations
* fix(ui): remove orphaned nodes from canvas graphs
* fix(nodes): fix cancel_by_batch_ids result count
* fix(ui): only show cancel batch tooltip when batches were canceled
* chore: isort
* fix(api): return `[""]` when dynamic prompts generates no prompts
Just a simple fallback so we always have a prompt.
* feat(ui): dynamicPrompts.combinatorial is always on
There seems to be little purpose in using the combinatorial generation for dynamic prompts. I've disabled it by hiding it from the UI and defaulting combinatorial to true. If we want to enable it again in the future it's straightforward to do so.
* feat: add queue_id & support logic
* feat(ui): fix upscale button
It prepends the upscale operation to queue
* feat(nodes): return queue item when enqueuing a single graph
This facilitates one-off graph async workflows in the client.
* feat(ui): move controlnet autoprocess to queue
* fix(ui): fix non-serializable DOMRect in redux state
* feat(ui): QueueTable performance tweaks
* feat(ui): update queue list
Queue items expand to show the full queue item. Just as JSON for now.
* wip threaded session_processor
* feat(nodes,ui): fully migrate queue to session_processor
* feat(nodes,ui): add processor events
* feat(ui): ui tweaks
* feat(nodes,ui): consolidate events, reduce network requests
* feat(ui): cleanup & abstract queue hooks
* feat(nodes): optimize batch permutation
Use a generator to do only as much work as is needed.
Previously, though we only ended up creating exactly as many queue items as was needed, there was still some intermediary work that calculated *all* permutations. When that number was very high, the system had a very hard time and used a lot of memory.
The logic has been refactored to use a generator. Additionally, the batch validators are optimized to return early and use less memory.
* feat(ui): add seed behaviour parameter
This dynamic prompts parameter allows the seed to be randomized per prompt or per iteration:
- Per iteration: Use the same seed for all prompts in a single dynamic prompt expansion
- Per prompt: Use a different seed for every single prompt
"Per iteration" is appropriate for exploring a the latents space with a stable starting noise, while "Per prompt" provides more variation.
* fix(ui): remove extraneous random seed nodes from linear graphs
* fix(ui): fix controlnet autoprocess not working when queue is running
* feat(queue): add timestamps to queue status updates
Also show execution time in queue list
* feat(queue): change all execution-related events to use the `queue_id` as the room, also include `queue_item_id` in InvocationQueueItem
This allows for much simpler handling of queue items.
* feat(api): deprecate sessions router
* chore(backend): tidy logging in `dependencies.py`
* fix(backend): respect `use_memory_db`
* feat(backend): add `config.log_sql` (enables sql trace logging)
* feat: add invocation cache
Supersedes #4574
The invocation cache provides simple node memoization functionality. Nodes that use the cache are memoized and not re-executed if their inputs haven't changed. Instead, the stored output is returned.
## Results
This feature provides anywhere some significant to massive performance improvement.
The improvement is most marked on large batches of generations where you only change a couple things (e.g. different seed or prompt for each iteration) and low-VRAM systems, where skipping an extraneous model load is a big deal.
## Overview
A new `invocation_cache` service is added to handle the caching. There's not much to it.
All nodes now inherit a boolean `use_cache` field from `BaseInvocation`. This is a node field and not a class attribute, because specific instances of nodes may want to opt in or out of caching.
The recently-added `invoke_internal()` method on `BaseInvocation` is used as an entrypoint for the cache logic.
To create a cache key, the invocation is first serialized using pydantic's provided `json()` method, skipping the unique `id` field. Then python's very fast builtin `hash()` is used to create an integer key. All implementations of `InvocationCacheBase` must provide a class method `create_key()` which accepts an invocation and outputs a string or integer key.
## In-Memory Implementation
An in-memory implementation is provided. In this implementation, the node outputs are stored in memory as python classes. The in-memory cache does not persist application restarts.
Max node cache size is added as `node_cache_size` under the `Generation` config category.
It defaults to 512 - this number is up for discussion, but given that these are relatively lightweight pydantic models, I think it's safe to up this even higher.
Note that the cache isn't storing the big stuff - tensors and images are store on disk, and outputs include only references to them.
## Node Definition
The default for all nodes is to use the cache. The `@invocation` decorator now accepts an optional `use_cache: bool` argument to override the default of `True`.
Non-deterministic nodes, however, should set this to `False`. Currently, all random-stuff nodes, including `dynamic_prompt`, are set to `False`.
The field name `use_cache` is now effectively a reserved field name and possibly a breaking change if any community nodes use this as a field name. In hindsight, all our reserved field names should have been prefixed with underscores or something.
## One Gotcha
Leaf nodes probably want to opt out of the cache, because if they are not cached, their outputs are not saved again.
If you run the same graph multiple times, you only end up with a single image output, because the image storage side-effects are in the `invoke()` method, which is bypassed if we have a cache hit.
## Linear UI
The linear graphs _almost_ just work, but due to the gotcha, we need to be careful about the final image-outputting node. To resolve this, a `SaveImageInvocation` node is added and used in the linear graphs.
This node is similar to `ImagePrimitive`, except it saves a copy of its input image, and has `use_cache` set to `False` by default.
This is now the leaf node in all linear graphs, and is the only node in those graphs with `use_cache == False` _and_ the only node with `is_intermedate == False`.
## Workflow Editor
All nodes now have a footer with a new `Use Cache [ ]` checkbox. It defaults to the value set by the invocation in its python definition, but can be changed by the user.
The workflow/node validation logic has been updated to migrate old workflows to use the new default values for `use_cache`. Users may still want to review the settings that have been chosen. In the event of catastrophic failure when running this migration, the default value of `True` is applied, as this is correct for most nodes.
Users should consider saving their workflows after loading them in and having them updated.
## Future Enhancements - Callback
A future enhancement would be to provide a callback to the `use_cache` flag that would be run as the node is executed to determine, based on its own internal state, if the cache should be used or not.
This would be useful for `DynamicPromptInvocation`, where the deterministic behaviour is determined by the `combinatorial: bool` field.
## Future Enhancements - Persisted Cache
Similar to how the latents storage is backed by disk, the invocation cache could be persisted to the database or disk. We'd need to be very careful about deserializing outputs, but it's perhaps worth exploring in the future.
* fix(ui): fix queue list item width
* feat(nodes): do not send the whole node on every generator progress
* feat(ui): strip out old logic related to sessions
Things like `isProcessing` are no longer relevant with queue. Removed them all & updated everything be appropriate for queue. May be a few little quirks I've missed...
* feat(ui): fix up param collapse labels
* feat(ui): click queue count to go to queue tab
* tidy(queue): update comment, query format
* feat(ui): fix progress bar when canceling
* fix(ui): fix circular dependency
* feat(nodes): bail on node caching logic if `node_cache_size == 0`
* feat(nodes): handle KeyError on node cache pop
* feat(nodes): bypass cache codepath if caches is disabled
more better no do thing
* fix(ui): reset api cache on connect/disconnect
* feat(ui): prevent enqueue when no prompts generated
* feat(ui): add queue controls to workflow editor
* feat(ui): update floating buttons & other incidental UI tweaks
* fix(ui): fix missing/incorrect translation keys
* fix(tests): add config service to mock invocation services
invoking needs access to `node_cache_size` to occur
* optionally remove pause/resume buttons from queue UI
* option to disable prepending
* chore(ui): remove unused file
* feat(queue): remove `order_id` entirely, `item_id` is now an autoinc pk
---------
Co-authored-by: Mary Hipp <maryhipp@Marys-MacBook-Air.local>
2023-09-20 05:09:24 +00:00
"nodeOpacity" : "Node Opacity" ,
2023-11-25 03:46:19 +00:00
"nodeVersion" : "Node Version" ,
2023-09-13 07:31:34 +00:00
"noOutputRecorded" : "No outputs recorded" ,
"notes" : "Notes" ,
"notesDescription" : "Add notes about your workflow" ,
"problemSettingTitle" : "Problem Setting Title" ,
2024-02-14 14:02:07 +00:00
"resetToDefaultValue" : "Reset to default value" ,
2023-09-13 07:31:34 +00:00
"reloadNodeTemplates" : "Reload Node Templates" ,
"removeLinearView" : "Remove from Linear View" ,
2024-02-14 22:27:21 +00:00
"reorderLinearView" : "Reorder Linear View" ,
2023-12-14 08:55:47 +00:00
"newWorkflow" : "New Workflow" ,
"newWorkflowDesc" : "Create a new workflow?" ,
"newWorkflowDesc2" : "Your current workflow has unsaved changes." ,
2024-01-31 13:32:31 +00:00
"clearWorkflow" : "Clear Workflow" ,
"clearWorkflowDesc" : "Clear this workflow and start a new one?" ,
"clearWorkflowDesc2" : "Your current workflow has unsaved changes." ,
2023-09-13 11:15:36 +00:00
"scheduler" : "Scheduler" ,
2023-09-13 07:31:34 +00:00
"showLegendNodes" : "Show Field Type Legend" ,
"showMinimapnodes" : "Show MiniMap" ,
"snapToGrid" : "Snap to Grid" ,
"snapToGridHelp" : "Snap nodes to grid when moved" ,
2023-09-13 11:15:36 +00:00
"string" : "String" ,
feat(ui): add support for custom field types
Node authors may now create their own arbitrary/custom field types. Any pydantic model is supported.
Two notes:
1. Your field type's class name must be unique.
Suggest prefixing fields with something related to the node pack as a kind of namespace.
2. Custom field types function as connection-only fields.
For example, if your custom field has string attributes, you will not get a text input for that attribute when you give a node a field with your custom type.
This is the same behaviour as other complex fields that don't have custom UIs in the workflow editor - like, say, a string collection.
feat(ui): fix tooltips for custom types
We need to hold onto the original type of the field so they don't all just show up as "Unknown".
fix(ui): fix ts error with custom fields
feat(ui): custom field types connection validation
In the initial commit, a custom field's original type was added to the *field templates* only as `originalType`. Custom fields' `type` property was `"Custom"`*. This allowed for type safety throughout the UI logic.
*Actually, it was `"Unknown"`, but I changed it to custom for clarity.
Connection validation logic, however, uses the *field instance* of the node/field. Like the templates, *field instances* with custom types have their `type` set to `"Custom"`, but they didn't have an `originalType` property. As a result, all custom fields could be connected to all other custom fields.
To resolve this, we need to add `originalType` to the *field instances*, then switch the validation logic to use this instead of `type`.
This ended up needing a bit of fanagling:
- If we make `originalType` a required property on field instances, existing workflows will break during connection validation, because they won't have this property. We'd need a new layer of logic to migrate the workflows, adding the new `originalType` property.
While this layer is probably needed anyways, typing `originalType` as optional is much simpler. Workflow migration logic can come layer.
(Technically, we could remove all references to field types from the workflow files, and let the templates hold all this information. This feels like a significant change and I'm reluctant to do it now.)
- Because `originalType` is optional, anywhere we care about the type of a field, we need to use it over `type`. So there are a number of `field.originalType ?? field.type` expressions. This is a bit of a gotcha, we'll need to remember this in the future.
- We use `Array.prototype.includes()` often in the workflow editor, e.g. `COLLECTION_TYPES.includes(type)`. In these cases, the const array is of type `FieldType[]`, and `type` is is `FieldType`.
Because we now support custom types, the arg `type` is now widened from `FieldType` to `string`.
This causes a TS error. This behaviour is somewhat controversial (see https://github.com/microsoft/TypeScript/issues/14520). These expressions are now rewritten as `COLLECTION_TYPES.some((t) => t === type)` to satisfy TS. It's logically equivalent.
fix(ui): typo
feat(ui): add CustomCollection and CustomPolymorphic field types
feat(ui): add validation for CustomCollection & CustomPolymorphic types
- Update connection validation for custom types
- Use simple string parsing to determine if a field is a collection or polymorphic type.
- No longer need to keep a list of collection and polymorphic types.
- Added runtime checks in `baseinvocation.py` to ensure no fields are named in such a way that it could mess up the new parsing
chore(ui): remove errant console.log
fix(ui): rename 'nodes.currentConnectionFieldType' -> 'nodes.connectionStartFieldType'
This was confusingly named and kept tripping me up. Renamed to be consistent with the `reactflow` `ConnectionStartParams` type.
fix(ui): fix ts error
feat(nodes): add runtime check for custom field names
"Custom", "CustomCollection" and "CustomPolymorphic" are reserved field names.
chore(ui): add TODO for revising field type names
wip refactor fieldtype structured
wip refactor field types
wip refactor types
wip refactor types
fix node layout
refactor field types
chore: mypy
organisation
organisation
organisation
fix(nodes): fix field orig_required, field_kind and input statuses
feat(nodes): remove broken implementation of default_factory on InputField
Use of this could break connection validation due to the difference in node schemas required fields and invoke() required args.
Removed entirely for now. It wasn't ever actually used by the system, because all graphs always had values provided for fields where default_factory was used.
Also, pydantic is smart enough to not reuse the same object when specifying a default value - it clones the object first. So, the common pattern of `default_factory=list` is extraneous. It can just be `default=[]`.
fix(nodes): fix InputField name validation
workflow validation
validation
chore: ruff
feat(nodes): fix up baseinvocation comments
fix(ui): improve typing & logic of buildFieldInputTemplate
improved error handling in parseFieldType
fix: back compat for deprecated default_factory and UIType
feat(nodes): do not show node packs loaded log if none loaded
chore(ui): typegen
2023-11-17 00:32:35 +00:00
"unableToLoadWorkflow" : "Unable to Load Workflow" ,
2023-09-13 07:31:34 +00:00
"unableToValidateWorkflow" : "Unable to Validate Workflow" ,
feat(ui): add support for custom field types
Node authors may now create their own arbitrary/custom field types. Any pydantic model is supported.
Two notes:
1. Your field type's class name must be unique.
Suggest prefixing fields with something related to the node pack as a kind of namespace.
2. Custom field types function as connection-only fields.
For example, if your custom field has string attributes, you will not get a text input for that attribute when you give a node a field with your custom type.
This is the same behaviour as other complex fields that don't have custom UIs in the workflow editor - like, say, a string collection.
feat(ui): fix tooltips for custom types
We need to hold onto the original type of the field so they don't all just show up as "Unknown".
fix(ui): fix ts error with custom fields
feat(ui): custom field types connection validation
In the initial commit, a custom field's original type was added to the *field templates* only as `originalType`. Custom fields' `type` property was `"Custom"`*. This allowed for type safety throughout the UI logic.
*Actually, it was `"Unknown"`, but I changed it to custom for clarity.
Connection validation logic, however, uses the *field instance* of the node/field. Like the templates, *field instances* with custom types have their `type` set to `"Custom"`, but they didn't have an `originalType` property. As a result, all custom fields could be connected to all other custom fields.
To resolve this, we need to add `originalType` to the *field instances*, then switch the validation logic to use this instead of `type`.
This ended up needing a bit of fanagling:
- If we make `originalType` a required property on field instances, existing workflows will break during connection validation, because they won't have this property. We'd need a new layer of logic to migrate the workflows, adding the new `originalType` property.
While this layer is probably needed anyways, typing `originalType` as optional is much simpler. Workflow migration logic can come layer.
(Technically, we could remove all references to field types from the workflow files, and let the templates hold all this information. This feels like a significant change and I'm reluctant to do it now.)
- Because `originalType` is optional, anywhere we care about the type of a field, we need to use it over `type`. So there are a number of `field.originalType ?? field.type` expressions. This is a bit of a gotcha, we'll need to remember this in the future.
- We use `Array.prototype.includes()` often in the workflow editor, e.g. `COLLECTION_TYPES.includes(type)`. In these cases, the const array is of type `FieldType[]`, and `type` is is `FieldType`.
Because we now support custom types, the arg `type` is now widened from `FieldType` to `string`.
This causes a TS error. This behaviour is somewhat controversial (see https://github.com/microsoft/TypeScript/issues/14520). These expressions are now rewritten as `COLLECTION_TYPES.some((t) => t === type)` to satisfy TS. It's logically equivalent.
fix(ui): typo
feat(ui): add CustomCollection and CustomPolymorphic field types
feat(ui): add validation for CustomCollection & CustomPolymorphic types
- Update connection validation for custom types
- Use simple string parsing to determine if a field is a collection or polymorphic type.
- No longer need to keep a list of collection and polymorphic types.
- Added runtime checks in `baseinvocation.py` to ensure no fields are named in such a way that it could mess up the new parsing
chore(ui): remove errant console.log
fix(ui): rename 'nodes.currentConnectionFieldType' -> 'nodes.connectionStartFieldType'
This was confusingly named and kept tripping me up. Renamed to be consistent with the `reactflow` `ConnectionStartParams` type.
fix(ui): fix ts error
feat(nodes): add runtime check for custom field names
"Custom", "CustomCollection" and "CustomPolymorphic" are reserved field names.
chore(ui): add TODO for revising field type names
wip refactor fieldtype structured
wip refactor field types
wip refactor types
wip refactor types
fix node layout
refactor field types
chore: mypy
organisation
organisation
organisation
fix(nodes): fix field orig_required, field_kind and input statuses
feat(nodes): remove broken implementation of default_factory on InputField
Use of this could break connection validation due to the difference in node schemas required fields and invoke() required args.
Removed entirely for now. It wasn't ever actually used by the system, because all graphs always had values provided for fields where default_factory was used.
Also, pydantic is smart enough to not reuse the same object when specifying a default value - it clones the object first. So, the common pattern of `default_factory=list` is extraneous. It can just be `default=[]`.
fix(nodes): fix InputField name validation
workflow validation
validation
chore: ruff
feat(nodes): fix up baseinvocation comments
fix(ui): improve typing & logic of buildFieldInputTemplate
improved error handling in parseFieldType
fix: back compat for deprecated default_factory and UIType
feat(nodes): do not show node packs loaded log if none loaded
chore(ui): typegen
2023-11-17 00:32:35 +00:00
"unknownErrorValidatingWorkflow" : "Unknown error validating workflow" ,
"inputFieldTypeParseError" : "Unable to parse type of input field {{node}}.{{field}} ({{message}})" ,
"outputFieldTypeParseError" : "Unable to parse type of output field {{node}}.{{field}} ({{message}})" ,
"unableToExtractSchemaNameFromRef" : "unable to extract schema name from ref" ,
"unsupportedArrayItemType" : "unsupported array item type \"{{type}}\"" ,
2023-11-26 09:44:39 +00:00
"unsupportedAnyOfLength" : "too many union members ({{count}})" ,
2023-11-28 23:19:51 +00:00
"unsupportedMismatchedUnion" : "mismatched CollectionOrScalar type with base types {{firstType}} and {{secondType}}" ,
feat(ui): add support for custom field types
Node authors may now create their own arbitrary/custom field types. Any pydantic model is supported.
Two notes:
1. Your field type's class name must be unique.
Suggest prefixing fields with something related to the node pack as a kind of namespace.
2. Custom field types function as connection-only fields.
For example, if your custom field has string attributes, you will not get a text input for that attribute when you give a node a field with your custom type.
This is the same behaviour as other complex fields that don't have custom UIs in the workflow editor - like, say, a string collection.
feat(ui): fix tooltips for custom types
We need to hold onto the original type of the field so they don't all just show up as "Unknown".
fix(ui): fix ts error with custom fields
feat(ui): custom field types connection validation
In the initial commit, a custom field's original type was added to the *field templates* only as `originalType`. Custom fields' `type` property was `"Custom"`*. This allowed for type safety throughout the UI logic.
*Actually, it was `"Unknown"`, but I changed it to custom for clarity.
Connection validation logic, however, uses the *field instance* of the node/field. Like the templates, *field instances* with custom types have their `type` set to `"Custom"`, but they didn't have an `originalType` property. As a result, all custom fields could be connected to all other custom fields.
To resolve this, we need to add `originalType` to the *field instances*, then switch the validation logic to use this instead of `type`.
This ended up needing a bit of fanagling:
- If we make `originalType` a required property on field instances, existing workflows will break during connection validation, because they won't have this property. We'd need a new layer of logic to migrate the workflows, adding the new `originalType` property.
While this layer is probably needed anyways, typing `originalType` as optional is much simpler. Workflow migration logic can come layer.
(Technically, we could remove all references to field types from the workflow files, and let the templates hold all this information. This feels like a significant change and I'm reluctant to do it now.)
- Because `originalType` is optional, anywhere we care about the type of a field, we need to use it over `type`. So there are a number of `field.originalType ?? field.type` expressions. This is a bit of a gotcha, we'll need to remember this in the future.
- We use `Array.prototype.includes()` often in the workflow editor, e.g. `COLLECTION_TYPES.includes(type)`. In these cases, the const array is of type `FieldType[]`, and `type` is is `FieldType`.
Because we now support custom types, the arg `type` is now widened from `FieldType` to `string`.
This causes a TS error. This behaviour is somewhat controversial (see https://github.com/microsoft/TypeScript/issues/14520). These expressions are now rewritten as `COLLECTION_TYPES.some((t) => t === type)` to satisfy TS. It's logically equivalent.
fix(ui): typo
feat(ui): add CustomCollection and CustomPolymorphic field types
feat(ui): add validation for CustomCollection & CustomPolymorphic types
- Update connection validation for custom types
- Use simple string parsing to determine if a field is a collection or polymorphic type.
- No longer need to keep a list of collection and polymorphic types.
- Added runtime checks in `baseinvocation.py` to ensure no fields are named in such a way that it could mess up the new parsing
chore(ui): remove errant console.log
fix(ui): rename 'nodes.currentConnectionFieldType' -> 'nodes.connectionStartFieldType'
This was confusingly named and kept tripping me up. Renamed to be consistent with the `reactflow` `ConnectionStartParams` type.
fix(ui): fix ts error
feat(nodes): add runtime check for custom field names
"Custom", "CustomCollection" and "CustomPolymorphic" are reserved field names.
chore(ui): add TODO for revising field type names
wip refactor fieldtype structured
wip refactor field types
wip refactor types
wip refactor types
fix node layout
refactor field types
chore: mypy
organisation
organisation
organisation
fix(nodes): fix field orig_required, field_kind and input statuses
feat(nodes): remove broken implementation of default_factory on InputField
Use of this could break connection validation due to the difference in node schemas required fields and invoke() required args.
Removed entirely for now. It wasn't ever actually used by the system, because all graphs always had values provided for fields where default_factory was used.
Also, pydantic is smart enough to not reuse the same object when specifying a default value - it clones the object first. So, the common pattern of `default_factory=list` is extraneous. It can just be `default=[]`.
fix(nodes): fix InputField name validation
workflow validation
validation
chore: ruff
feat(nodes): fix up baseinvocation comments
fix(ui): improve typing & logic of buildFieldInputTemplate
improved error handling in parseFieldType
fix: back compat for deprecated default_factory and UIType
feat(nodes): do not show node packs loaded log if none loaded
chore(ui): typegen
2023-11-17 00:32:35 +00:00
"unableToParseFieldType" : "unable to parse field type" ,
2023-11-30 11:01:33 +00:00
"unableToExtractEnumOptions" : "unable to extract enum options" ,
2023-11-25 03:46:19 +00:00
"unknownField" : "Unknown field" ,
2023-11-27 01:51:19 +00:00
"unknownFieldType" : "$t(nodes.unknownField) type: {{type}}" ,
2023-09-13 11:15:36 +00:00
"unknownNode" : "Unknown Node" ,
2023-11-27 01:51:19 +00:00
"unknownNodeType" : "Unknown node type" ,
2023-09-13 11:15:36 +00:00
"unknownTemplate" : "Unknown Template" ,
2023-11-25 10:10:22 +00:00
"unknownInput" : "Unknown input: {{name}}" ,
"unknownOutput" : "Unknown output: {{name}}" ,
feat: queued generation (#4502)
* fix(config): fix typing issues in `config/`
`config/invokeai_config.py`:
- use `Optional` for things that are optional
- fix typing of `ram_cache_size()` and `vram_cache_size()`
- remove unused and incorrectly typed method `autoconvert_path`
- fix types and logic for `parse_args()`, in which `InvokeAIAppConfig.initconf` *must* be a `DictConfig`, but function would allow it to be set as a `ListConfig`, which presumably would cause issues elsewhere
`config/base.py`:
- use `cls` for first arg of class methods
- use `Optional` for things that are optional
- fix minor type issue related to setting of `env_prefix`
- remove unused `add_subparser()` method, which calls `add_parser()` on an `ArgumentParser` (method only available on the `_SubParsersAction` object, which is returned from ArgumentParser.add_subparsers()`)
* feat: queued generation and batches
Due to a very messy branch with broad addition of `isort` on `main` alongside it, some git surgery was needed to get an agreeable git history. This commit represents all of the work on queued generation. See PR for notes.
* chore: flake8, isort, black
* fix(nodes): fix incorrect service stop() method
* fix(nodes): improve names of a few variables
* fix(tests): fix up tests after changes to batches/queue
* feat(tests): add unit tests for session queue helper functions
* feat(ui): dynamic prompts is always enabled
* feat(queue): add queue_status_changed event
* feat(ui): wip queue graphs
* feat(nodes): move cleanup til after invoker startup
* feat(nodes): add cancel_by_batch_ids
* feat(ui): wip batch graphs & UI
* fix(nodes): remove `Batch.batch_id` from required
* fix(ui): cleanup and use fixedCacheKey for all mutations
* fix(ui): remove orphaned nodes from canvas graphs
* fix(nodes): fix cancel_by_batch_ids result count
* fix(ui): only show cancel batch tooltip when batches were canceled
* chore: isort
* fix(api): return `[""]` when dynamic prompts generates no prompts
Just a simple fallback so we always have a prompt.
* feat(ui): dynamicPrompts.combinatorial is always on
There seems to be little purpose in using the combinatorial generation for dynamic prompts. I've disabled it by hiding it from the UI and defaulting combinatorial to true. If we want to enable it again in the future it's straightforward to do so.
* feat: add queue_id & support logic
* feat(ui): fix upscale button
It prepends the upscale operation to queue
* feat(nodes): return queue item when enqueuing a single graph
This facilitates one-off graph async workflows in the client.
* feat(ui): move controlnet autoprocess to queue
* fix(ui): fix non-serializable DOMRect in redux state
* feat(ui): QueueTable performance tweaks
* feat(ui): update queue list
Queue items expand to show the full queue item. Just as JSON for now.
* wip threaded session_processor
* feat(nodes,ui): fully migrate queue to session_processor
* feat(nodes,ui): add processor events
* feat(ui): ui tweaks
* feat(nodes,ui): consolidate events, reduce network requests
* feat(ui): cleanup & abstract queue hooks
* feat(nodes): optimize batch permutation
Use a generator to do only as much work as is needed.
Previously, though we only ended up creating exactly as many queue items as was needed, there was still some intermediary work that calculated *all* permutations. When that number was very high, the system had a very hard time and used a lot of memory.
The logic has been refactored to use a generator. Additionally, the batch validators are optimized to return early and use less memory.
* feat(ui): add seed behaviour parameter
This dynamic prompts parameter allows the seed to be randomized per prompt or per iteration:
- Per iteration: Use the same seed for all prompts in a single dynamic prompt expansion
- Per prompt: Use a different seed for every single prompt
"Per iteration" is appropriate for exploring a the latents space with a stable starting noise, while "Per prompt" provides more variation.
* fix(ui): remove extraneous random seed nodes from linear graphs
* fix(ui): fix controlnet autoprocess not working when queue is running
* feat(queue): add timestamps to queue status updates
Also show execution time in queue list
* feat(queue): change all execution-related events to use the `queue_id` as the room, also include `queue_item_id` in InvocationQueueItem
This allows for much simpler handling of queue items.
* feat(api): deprecate sessions router
* chore(backend): tidy logging in `dependencies.py`
* fix(backend): respect `use_memory_db`
* feat(backend): add `config.log_sql` (enables sql trace logging)
* feat: add invocation cache
Supersedes #4574
The invocation cache provides simple node memoization functionality. Nodes that use the cache are memoized and not re-executed if their inputs haven't changed. Instead, the stored output is returned.
## Results
This feature provides anywhere some significant to massive performance improvement.
The improvement is most marked on large batches of generations where you only change a couple things (e.g. different seed or prompt for each iteration) and low-VRAM systems, where skipping an extraneous model load is a big deal.
## Overview
A new `invocation_cache` service is added to handle the caching. There's not much to it.
All nodes now inherit a boolean `use_cache` field from `BaseInvocation`. This is a node field and not a class attribute, because specific instances of nodes may want to opt in or out of caching.
The recently-added `invoke_internal()` method on `BaseInvocation` is used as an entrypoint for the cache logic.
To create a cache key, the invocation is first serialized using pydantic's provided `json()` method, skipping the unique `id` field. Then python's very fast builtin `hash()` is used to create an integer key. All implementations of `InvocationCacheBase` must provide a class method `create_key()` which accepts an invocation and outputs a string or integer key.
## In-Memory Implementation
An in-memory implementation is provided. In this implementation, the node outputs are stored in memory as python classes. The in-memory cache does not persist application restarts.
Max node cache size is added as `node_cache_size` under the `Generation` config category.
It defaults to 512 - this number is up for discussion, but given that these are relatively lightweight pydantic models, I think it's safe to up this even higher.
Note that the cache isn't storing the big stuff - tensors and images are store on disk, and outputs include only references to them.
## Node Definition
The default for all nodes is to use the cache. The `@invocation` decorator now accepts an optional `use_cache: bool` argument to override the default of `True`.
Non-deterministic nodes, however, should set this to `False`. Currently, all random-stuff nodes, including `dynamic_prompt`, are set to `False`.
The field name `use_cache` is now effectively a reserved field name and possibly a breaking change if any community nodes use this as a field name. In hindsight, all our reserved field names should have been prefixed with underscores or something.
## One Gotcha
Leaf nodes probably want to opt out of the cache, because if they are not cached, their outputs are not saved again.
If you run the same graph multiple times, you only end up with a single image output, because the image storage side-effects are in the `invoke()` method, which is bypassed if we have a cache hit.
## Linear UI
The linear graphs _almost_ just work, but due to the gotcha, we need to be careful about the final image-outputting node. To resolve this, a `SaveImageInvocation` node is added and used in the linear graphs.
This node is similar to `ImagePrimitive`, except it saves a copy of its input image, and has `use_cache` set to `False` by default.
This is now the leaf node in all linear graphs, and is the only node in those graphs with `use_cache == False` _and_ the only node with `is_intermedate == False`.
## Workflow Editor
All nodes now have a footer with a new `Use Cache [ ]` checkbox. It defaults to the value set by the invocation in its python definition, but can be changed by the user.
The workflow/node validation logic has been updated to migrate old workflows to use the new default values for `use_cache`. Users may still want to review the settings that have been chosen. In the event of catastrophic failure when running this migration, the default value of `True` is applied, as this is correct for most nodes.
Users should consider saving their workflows after loading them in and having them updated.
## Future Enhancements - Callback
A future enhancement would be to provide a callback to the `use_cache` flag that would be run as the node is executed to determine, based on its own internal state, if the cache should be used or not.
This would be useful for `DynamicPromptInvocation`, where the deterministic behaviour is determined by the `combinatorial: bool` field.
## Future Enhancements - Persisted Cache
Similar to how the latents storage is backed by disk, the invocation cache could be persisted to the database or disk. We'd need to be very careful about deserializing outputs, but it's perhaps worth exploring in the future.
* fix(ui): fix queue list item width
* feat(nodes): do not send the whole node on every generator progress
* feat(ui): strip out old logic related to sessions
Things like `isProcessing` are no longer relevant with queue. Removed them all & updated everything be appropriate for queue. May be a few little quirks I've missed...
* feat(ui): fix up param collapse labels
* feat(ui): click queue count to go to queue tab
* tidy(queue): update comment, query format
* feat(ui): fix progress bar when canceling
* fix(ui): fix circular dependency
* feat(nodes): bail on node caching logic if `node_cache_size == 0`
* feat(nodes): handle KeyError on node cache pop
* feat(nodes): bypass cache codepath if caches is disabled
more better no do thing
* fix(ui): reset api cache on connect/disconnect
* feat(ui): prevent enqueue when no prompts generated
* feat(ui): add queue controls to workflow editor
* feat(ui): update floating buttons & other incidental UI tweaks
* fix(ui): fix missing/incorrect translation keys
* fix(tests): add config service to mock invocation services
invoking needs access to `node_cache_size` to occur
* optionally remove pause/resume buttons from queue UI
* option to disable prepending
* chore(ui): remove unused file
* feat(queue): remove `order_id` entirely, `item_id` is now an autoinc pk
---------
Co-authored-by: Mary Hipp <maryhipp@Marys-MacBook-Air.local>
2023-09-20 05:09:24 +00:00
"updateNode" : "Update Node" ,
2023-09-13 11:15:36 +00:00
"updateApp" : "Update App" ,
2023-11-25 10:10:22 +00:00
"updateAllNodes" : "Update Nodes" ,
feat(ui): add support for custom field types
Node authors may now create their own arbitrary/custom field types. Any pydantic model is supported.
Two notes:
1. Your field type's class name must be unique.
Suggest prefixing fields with something related to the node pack as a kind of namespace.
2. Custom field types function as connection-only fields.
For example, if your custom field has string attributes, you will not get a text input for that attribute when you give a node a field with your custom type.
This is the same behaviour as other complex fields that don't have custom UIs in the workflow editor - like, say, a string collection.
feat(ui): fix tooltips for custom types
We need to hold onto the original type of the field so they don't all just show up as "Unknown".
fix(ui): fix ts error with custom fields
feat(ui): custom field types connection validation
In the initial commit, a custom field's original type was added to the *field templates* only as `originalType`. Custom fields' `type` property was `"Custom"`*. This allowed for type safety throughout the UI logic.
*Actually, it was `"Unknown"`, but I changed it to custom for clarity.
Connection validation logic, however, uses the *field instance* of the node/field. Like the templates, *field instances* with custom types have their `type` set to `"Custom"`, but they didn't have an `originalType` property. As a result, all custom fields could be connected to all other custom fields.
To resolve this, we need to add `originalType` to the *field instances*, then switch the validation logic to use this instead of `type`.
This ended up needing a bit of fanagling:
- If we make `originalType` a required property on field instances, existing workflows will break during connection validation, because they won't have this property. We'd need a new layer of logic to migrate the workflows, adding the new `originalType` property.
While this layer is probably needed anyways, typing `originalType` as optional is much simpler. Workflow migration logic can come layer.
(Technically, we could remove all references to field types from the workflow files, and let the templates hold all this information. This feels like a significant change and I'm reluctant to do it now.)
- Because `originalType` is optional, anywhere we care about the type of a field, we need to use it over `type`. So there are a number of `field.originalType ?? field.type` expressions. This is a bit of a gotcha, we'll need to remember this in the future.
- We use `Array.prototype.includes()` often in the workflow editor, e.g. `COLLECTION_TYPES.includes(type)`. In these cases, the const array is of type `FieldType[]`, and `type` is is `FieldType`.
Because we now support custom types, the arg `type` is now widened from `FieldType` to `string`.
This causes a TS error. This behaviour is somewhat controversial (see https://github.com/microsoft/TypeScript/issues/14520). These expressions are now rewritten as `COLLECTION_TYPES.some((t) => t === type)` to satisfy TS. It's logically equivalent.
fix(ui): typo
feat(ui): add CustomCollection and CustomPolymorphic field types
feat(ui): add validation for CustomCollection & CustomPolymorphic types
- Update connection validation for custom types
- Use simple string parsing to determine if a field is a collection or polymorphic type.
- No longer need to keep a list of collection and polymorphic types.
- Added runtime checks in `baseinvocation.py` to ensure no fields are named in such a way that it could mess up the new parsing
chore(ui): remove errant console.log
fix(ui): rename 'nodes.currentConnectionFieldType' -> 'nodes.connectionStartFieldType'
This was confusingly named and kept tripping me up. Renamed to be consistent with the `reactflow` `ConnectionStartParams` type.
fix(ui): fix ts error
feat(nodes): add runtime check for custom field names
"Custom", "CustomCollection" and "CustomPolymorphic" are reserved field names.
chore(ui): add TODO for revising field type names
wip refactor fieldtype structured
wip refactor field types
wip refactor types
wip refactor types
fix node layout
refactor field types
chore: mypy
organisation
organisation
organisation
fix(nodes): fix field orig_required, field_kind and input statuses
feat(nodes): remove broken implementation of default_factory on InputField
Use of this could break connection validation due to the difference in node schemas required fields and invoke() required args.
Removed entirely for now. It wasn't ever actually used by the system, because all graphs always had values provided for fields where default_factory was used.
Also, pydantic is smart enough to not reuse the same object when specifying a default value - it clones the object first. So, the common pattern of `default_factory=list` is extraneous. It can just be `default=[]`.
fix(nodes): fix InputField name validation
workflow validation
validation
chore: ruff
feat(nodes): fix up baseinvocation comments
fix(ui): improve typing & logic of buildFieldInputTemplate
improved error handling in parseFieldType
fix: back compat for deprecated default_factory and UIType
feat(nodes): do not show node packs loaded log if none loaded
chore(ui): typegen
2023-11-17 00:32:35 +00:00
"allNodesUpdated" : "All Nodes Updated" ,
2023-11-16 01:42:25 +00:00
"unableToUpdateNodes_one" : "Unable to update {{count}} node" ,
"unableToUpdateNodes_other" : "Unable to update {{count}} nodes" ,
2023-09-13 07:31:34 +00:00
"validateConnections" : "Validate Connections and Graph" ,
"validateConnectionsHelp" : "Prevent invalid connections from being made, and invalid graphs from being invoked" ,
2024-02-14 14:02:07 +00:00
"viewMode" : "Use in Linear View" ,
feat(ui): add support for custom field types
Node authors may now create their own arbitrary/custom field types. Any pydantic model is supported.
Two notes:
1. Your field type's class name must be unique.
Suggest prefixing fields with something related to the node pack as a kind of namespace.
2. Custom field types function as connection-only fields.
For example, if your custom field has string attributes, you will not get a text input for that attribute when you give a node a field with your custom type.
This is the same behaviour as other complex fields that don't have custom UIs in the workflow editor - like, say, a string collection.
feat(ui): fix tooltips for custom types
We need to hold onto the original type of the field so they don't all just show up as "Unknown".
fix(ui): fix ts error with custom fields
feat(ui): custom field types connection validation
In the initial commit, a custom field's original type was added to the *field templates* only as `originalType`. Custom fields' `type` property was `"Custom"`*. This allowed for type safety throughout the UI logic.
*Actually, it was `"Unknown"`, but I changed it to custom for clarity.
Connection validation logic, however, uses the *field instance* of the node/field. Like the templates, *field instances* with custom types have their `type` set to `"Custom"`, but they didn't have an `originalType` property. As a result, all custom fields could be connected to all other custom fields.
To resolve this, we need to add `originalType` to the *field instances*, then switch the validation logic to use this instead of `type`.
This ended up needing a bit of fanagling:
- If we make `originalType` a required property on field instances, existing workflows will break during connection validation, because they won't have this property. We'd need a new layer of logic to migrate the workflows, adding the new `originalType` property.
While this layer is probably needed anyways, typing `originalType` as optional is much simpler. Workflow migration logic can come layer.
(Technically, we could remove all references to field types from the workflow files, and let the templates hold all this information. This feels like a significant change and I'm reluctant to do it now.)
- Because `originalType` is optional, anywhere we care about the type of a field, we need to use it over `type`. So there are a number of `field.originalType ?? field.type` expressions. This is a bit of a gotcha, we'll need to remember this in the future.
- We use `Array.prototype.includes()` often in the workflow editor, e.g. `COLLECTION_TYPES.includes(type)`. In these cases, the const array is of type `FieldType[]`, and `type` is is `FieldType`.
Because we now support custom types, the arg `type` is now widened from `FieldType` to `string`.
This causes a TS error. This behaviour is somewhat controversial (see https://github.com/microsoft/TypeScript/issues/14520). These expressions are now rewritten as `COLLECTION_TYPES.some((t) => t === type)` to satisfy TS. It's logically equivalent.
fix(ui): typo
feat(ui): add CustomCollection and CustomPolymorphic field types
feat(ui): add validation for CustomCollection & CustomPolymorphic types
- Update connection validation for custom types
- Use simple string parsing to determine if a field is a collection or polymorphic type.
- No longer need to keep a list of collection and polymorphic types.
- Added runtime checks in `baseinvocation.py` to ensure no fields are named in such a way that it could mess up the new parsing
chore(ui): remove errant console.log
fix(ui): rename 'nodes.currentConnectionFieldType' -> 'nodes.connectionStartFieldType'
This was confusingly named and kept tripping me up. Renamed to be consistent with the `reactflow` `ConnectionStartParams` type.
fix(ui): fix ts error
feat(nodes): add runtime check for custom field names
"Custom", "CustomCollection" and "CustomPolymorphic" are reserved field names.
chore(ui): add TODO for revising field type names
wip refactor fieldtype structured
wip refactor field types
wip refactor types
wip refactor types
fix node layout
refactor field types
chore: mypy
organisation
organisation
organisation
fix(nodes): fix field orig_required, field_kind and input statuses
feat(nodes): remove broken implementation of default_factory on InputField
Use of this could break connection validation due to the difference in node schemas required fields and invoke() required args.
Removed entirely for now. It wasn't ever actually used by the system, because all graphs always had values provided for fields where default_factory was used.
Also, pydantic is smart enough to not reuse the same object when specifying a default value - it clones the object first. So, the common pattern of `default_factory=list` is extraneous. It can just be `default=[]`.
fix(nodes): fix InputField name validation
workflow validation
validation
chore: ruff
feat(nodes): fix up baseinvocation comments
fix(ui): improve typing & logic of buildFieldInputTemplate
improved error handling in parseFieldType
fix: back compat for deprecated default_factory and UIType
feat(nodes): do not show node packs loaded log if none loaded
chore(ui): typegen
2023-11-17 00:32:35 +00:00
"unableToGetWorkflowVersion" : "Unable to get workflow schema version" ,
2023-09-13 11:15:36 +00:00
"version" : "Version" ,
"versionUnknown" : " Version Unknown" ,
2023-09-13 07:31:34 +00:00
"workflow" : "Workflow" ,
2024-04-09 00:17:03 +00:00
"graph" : "Graph" ,
feat(ui): store graph in image metadata
The previous super-minimal implementation had a major issue - the saved workflow didn't take into account batched field values. When generating with multiple iterations or dynamic prompts, the same workflow with the first prompt, seed, etc was stored in each image.
As a result, when the batch results in multiple queue items, only one of the images has the correct workflow - the others are mismatched.
To work around this, we can store the _graph_ in the image metadata (alongside the workflow, if generated via workflow editor). When loading a workflow from an image, we can choose to load the workflow or the graph, preferring the workflow.
Internally, we need to update images router image-saving services. The changes are minimal.
To avoid pydantic errors deserializing the graph, when we extract it from the image, we will leave it as stringified JSON and let the frontend's more sophisticated and flexible parsing handle it. The worklow is also changed to just return stringified JSON, so the API is consistent.
2024-05-17 08:10:04 +00:00
"noGraph" : "No Graph" ,
2023-09-13 07:31:34 +00:00
"workflowAuthor" : "Author" ,
"workflowContact" : "Contact" ,
"workflowDescription" : "Short Description" ,
"workflowName" : "Name" ,
"workflowNotes" : "Notes" ,
"workflowSettings" : "Workflow Editor Settings" ,
"workflowTags" : "Tags" ,
"workflowValidation" : "Workflow Validation Error" ,
"workflowVersion" : "Version" ,
"zoomInNodes" : "Zoom In" ,
2023-12-12 04:04:13 +00:00
"zoomOutNodes" : "Zoom Out" ,
"betaDesc" : "This invocation is in beta. Until it is stable, it may have breaking changes during app updates. We plan to support this invocation long-term." ,
2024-05-20 23:54:13 +00:00
"prototypeDesc" : "This invocation is a prototype. It may have breaking changes during app updates and may be removed at any time." ,
"imageAccessError" : "Unable to find image {{image_name}}, resetting to default" ,
"boardAccessError" : "Unable to find board {{board_id}}, resetting to default" ,
"modelAccessError" : "Unable to find model {{key}}, resetting to default"
2023-02-18 04:23:24 +00:00
} ,
"parameters" : {
2023-12-28 13:03:21 +00:00
"aspect" : "Aspect" ,
2024-01-03 12:56:28 +00:00
"lockAspectRatio" : "Lock Aspect Ratio" ,
"swapDimensions" : "Swap Dimensions" ,
"setToOptimalSize" : "Optimize size for model" ,
"setToOptimalSizeTooSmall" : "$t(parameters.setToOptimalSize) (may be too small)" ,
"setToOptimalSizeTooLarge" : "$t(parameters.setToOptimalSize) (may be too large)" ,
2023-02-19 00:25:01 +00:00
"cancel" : {
2024-03-13 11:37:47 +00:00
"cancel" : "Cancel"
2023-02-18 21:35:33 +00:00
} ,
2023-09-13 07:31:34 +00:00
"cfgScale" : "CFG Scale" ,
2023-11-30 09:55:20 +00:00
"cfgRescaleMultiplier" : "CFG Rescale Multiplier" ,
2023-09-13 07:31:34 +00:00
"clipSkip" : "CLIP Skip" ,
"coherenceMode" : "Mode" ,
2024-02-21 17:32:49 +00:00
"coherenceEdgeSize" : "Edge Size" ,
"coherenceMinDenoise" : "Min Denoise" ,
2023-09-13 07:31:34 +00:00
"controlNetControlMode" : "Control Mode" ,
2023-02-18 04:23:24 +00:00
"copyImage" : "Copy Image" ,
2023-09-13 07:31:34 +00:00
"denoisingStrength" : "Denoising Strength" ,
2023-02-18 04:23:24 +00:00
"downloadImage" : "Download Image" ,
2023-09-13 07:31:34 +00:00
"general" : "General" ,
2024-04-22 22:53:28 +00:00
"globalSettings" : "Global Settings" ,
2023-09-13 07:31:34 +00:00
"height" : "Height" ,
"imageFit" : "Fit Initial Image To Output Size" ,
"images" : "Images" ,
"infillMethod" : "Infill Method" ,
2024-03-20 08:06:55 +00:00
"infillMosaicTileWidth" : "Tile Width" ,
"infillMosaicTileHeight" : "Tile Height" ,
"infillMosaicMinColor" : "Min Color" ,
"infillMosaicMaxColor" : "Max Color" ,
2024-03-30 09:32:51 +00:00
"infillColorValue" : "Fill Color" ,
2023-02-18 04:23:24 +00:00
"info" : "Info" ,
2023-09-13 07:31:34 +00:00
"invoke" : {
"addingImagesTo" : "Adding images to" ,
"invoke" : "Invoke" ,
"missingFieldTemplate" : "Missing field template" ,
"missingInputForField" : "{{nodeLabel}} -> {{fieldLabel}} missing input" ,
"missingNodeTemplate" : "Missing node template" ,
2023-10-12 03:16:54 +00:00
"noControlImageForControlAdapter" : "Control Adapter #{{number}} has no control image" ,
2024-05-01 11:36:51 +00:00
"imageNotProcessedForControlAdapter" : "Control Adapter #{{number}}'s image is not processed" ,
2023-09-13 07:31:34 +00:00
"noInitialImageSelected" : "No initial image selected" ,
2023-10-12 03:16:54 +00:00
"noModelForControlAdapter" : "Control Adapter #{{number}} has no model selected." ,
2023-12-28 22:13:43 +00:00
"incompatibleBaseModelForControlAdapter" : "Control Adapter #{{number}} model is incompatible with main model." ,
2023-09-13 07:31:34 +00:00
"noModelSelected" : "No model selected" ,
feat: queued generation (#4502)
* fix(config): fix typing issues in `config/`
`config/invokeai_config.py`:
- use `Optional` for things that are optional
- fix typing of `ram_cache_size()` and `vram_cache_size()`
- remove unused and incorrectly typed method `autoconvert_path`
- fix types and logic for `parse_args()`, in which `InvokeAIAppConfig.initconf` *must* be a `DictConfig`, but function would allow it to be set as a `ListConfig`, which presumably would cause issues elsewhere
`config/base.py`:
- use `cls` for first arg of class methods
- use `Optional` for things that are optional
- fix minor type issue related to setting of `env_prefix`
- remove unused `add_subparser()` method, which calls `add_parser()` on an `ArgumentParser` (method only available on the `_SubParsersAction` object, which is returned from ArgumentParser.add_subparsers()`)
* feat: queued generation and batches
Due to a very messy branch with broad addition of `isort` on `main` alongside it, some git surgery was needed to get an agreeable git history. This commit represents all of the work on queued generation. See PR for notes.
* chore: flake8, isort, black
* fix(nodes): fix incorrect service stop() method
* fix(nodes): improve names of a few variables
* fix(tests): fix up tests after changes to batches/queue
* feat(tests): add unit tests for session queue helper functions
* feat(ui): dynamic prompts is always enabled
* feat(queue): add queue_status_changed event
* feat(ui): wip queue graphs
* feat(nodes): move cleanup til after invoker startup
* feat(nodes): add cancel_by_batch_ids
* feat(ui): wip batch graphs & UI
* fix(nodes): remove `Batch.batch_id` from required
* fix(ui): cleanup and use fixedCacheKey for all mutations
* fix(ui): remove orphaned nodes from canvas graphs
* fix(nodes): fix cancel_by_batch_ids result count
* fix(ui): only show cancel batch tooltip when batches were canceled
* chore: isort
* fix(api): return `[""]` when dynamic prompts generates no prompts
Just a simple fallback so we always have a prompt.
* feat(ui): dynamicPrompts.combinatorial is always on
There seems to be little purpose in using the combinatorial generation for dynamic prompts. I've disabled it by hiding it from the UI and defaulting combinatorial to true. If we want to enable it again in the future it's straightforward to do so.
* feat: add queue_id & support logic
* feat(ui): fix upscale button
It prepends the upscale operation to queue
* feat(nodes): return queue item when enqueuing a single graph
This facilitates one-off graph async workflows in the client.
* feat(ui): move controlnet autoprocess to queue
* fix(ui): fix non-serializable DOMRect in redux state
* feat(ui): QueueTable performance tweaks
* feat(ui): update queue list
Queue items expand to show the full queue item. Just as JSON for now.
* wip threaded session_processor
* feat(nodes,ui): fully migrate queue to session_processor
* feat(nodes,ui): add processor events
* feat(ui): ui tweaks
* feat(nodes,ui): consolidate events, reduce network requests
* feat(ui): cleanup & abstract queue hooks
* feat(nodes): optimize batch permutation
Use a generator to do only as much work as is needed.
Previously, though we only ended up creating exactly as many queue items as was needed, there was still some intermediary work that calculated *all* permutations. When that number was very high, the system had a very hard time and used a lot of memory.
The logic has been refactored to use a generator. Additionally, the batch validators are optimized to return early and use less memory.
* feat(ui): add seed behaviour parameter
This dynamic prompts parameter allows the seed to be randomized per prompt or per iteration:
- Per iteration: Use the same seed for all prompts in a single dynamic prompt expansion
- Per prompt: Use a different seed for every single prompt
"Per iteration" is appropriate for exploring a the latents space with a stable starting noise, while "Per prompt" provides more variation.
* fix(ui): remove extraneous random seed nodes from linear graphs
* fix(ui): fix controlnet autoprocess not working when queue is running
* feat(queue): add timestamps to queue status updates
Also show execution time in queue list
* feat(queue): change all execution-related events to use the `queue_id` as the room, also include `queue_item_id` in InvocationQueueItem
This allows for much simpler handling of queue items.
* feat(api): deprecate sessions router
* chore(backend): tidy logging in `dependencies.py`
* fix(backend): respect `use_memory_db`
* feat(backend): add `config.log_sql` (enables sql trace logging)
* feat: add invocation cache
Supersedes #4574
The invocation cache provides simple node memoization functionality. Nodes that use the cache are memoized and not re-executed if their inputs haven't changed. Instead, the stored output is returned.
## Results
This feature provides anywhere some significant to massive performance improvement.
The improvement is most marked on large batches of generations where you only change a couple things (e.g. different seed or prompt for each iteration) and low-VRAM systems, where skipping an extraneous model load is a big deal.
## Overview
A new `invocation_cache` service is added to handle the caching. There's not much to it.
All nodes now inherit a boolean `use_cache` field from `BaseInvocation`. This is a node field and not a class attribute, because specific instances of nodes may want to opt in or out of caching.
The recently-added `invoke_internal()` method on `BaseInvocation` is used as an entrypoint for the cache logic.
To create a cache key, the invocation is first serialized using pydantic's provided `json()` method, skipping the unique `id` field. Then python's very fast builtin `hash()` is used to create an integer key. All implementations of `InvocationCacheBase` must provide a class method `create_key()` which accepts an invocation and outputs a string or integer key.
## In-Memory Implementation
An in-memory implementation is provided. In this implementation, the node outputs are stored in memory as python classes. The in-memory cache does not persist application restarts.
Max node cache size is added as `node_cache_size` under the `Generation` config category.
It defaults to 512 - this number is up for discussion, but given that these are relatively lightweight pydantic models, I think it's safe to up this even higher.
Note that the cache isn't storing the big stuff - tensors and images are store on disk, and outputs include only references to them.
## Node Definition
The default for all nodes is to use the cache. The `@invocation` decorator now accepts an optional `use_cache: bool` argument to override the default of `True`.
Non-deterministic nodes, however, should set this to `False`. Currently, all random-stuff nodes, including `dynamic_prompt`, are set to `False`.
The field name `use_cache` is now effectively a reserved field name and possibly a breaking change if any community nodes use this as a field name. In hindsight, all our reserved field names should have been prefixed with underscores or something.
## One Gotcha
Leaf nodes probably want to opt out of the cache, because if they are not cached, their outputs are not saved again.
If you run the same graph multiple times, you only end up with a single image output, because the image storage side-effects are in the `invoke()` method, which is bypassed if we have a cache hit.
## Linear UI
The linear graphs _almost_ just work, but due to the gotcha, we need to be careful about the final image-outputting node. To resolve this, a `SaveImageInvocation` node is added and used in the linear graphs.
This node is similar to `ImagePrimitive`, except it saves a copy of its input image, and has `use_cache` set to `False` by default.
This is now the leaf node in all linear graphs, and is the only node in those graphs with `use_cache == False` _and_ the only node with `is_intermedate == False`.
## Workflow Editor
All nodes now have a footer with a new `Use Cache [ ]` checkbox. It defaults to the value set by the invocation in its python definition, but can be changed by the user.
The workflow/node validation logic has been updated to migrate old workflows to use the new default values for `use_cache`. Users may still want to review the settings that have been chosen. In the event of catastrophic failure when running this migration, the default value of `True` is applied, as this is correct for most nodes.
Users should consider saving their workflows after loading them in and having them updated.
## Future Enhancements - Callback
A future enhancement would be to provide a callback to the `use_cache` flag that would be run as the node is executed to determine, based on its own internal state, if the cache should be used or not.
This would be useful for `DynamicPromptInvocation`, where the deterministic behaviour is determined by the `combinatorial: bool` field.
## Future Enhancements - Persisted Cache
Similar to how the latents storage is backed by disk, the invocation cache could be persisted to the database or disk. We'd need to be very careful about deserializing outputs, but it's perhaps worth exploring in the future.
* fix(ui): fix queue list item width
* feat(nodes): do not send the whole node on every generator progress
* feat(ui): strip out old logic related to sessions
Things like `isProcessing` are no longer relevant with queue. Removed them all & updated everything be appropriate for queue. May be a few little quirks I've missed...
* feat(ui): fix up param collapse labels
* feat(ui): click queue count to go to queue tab
* tidy(queue): update comment, query format
* feat(ui): fix progress bar when canceling
* fix(ui): fix circular dependency
* feat(nodes): bail on node caching logic if `node_cache_size == 0`
* feat(nodes): handle KeyError on node cache pop
* feat(nodes): bypass cache codepath if caches is disabled
more better no do thing
* fix(ui): reset api cache on connect/disconnect
* feat(ui): prevent enqueue when no prompts generated
* feat(ui): add queue controls to workflow editor
* feat(ui): update floating buttons & other incidental UI tweaks
* fix(ui): fix missing/incorrect translation keys
* fix(tests): add config service to mock invocation services
invoking needs access to `node_cache_size` to occur
* optionally remove pause/resume buttons from queue UI
* option to disable prepending
* chore(ui): remove unused file
* feat(queue): remove `order_id` entirely, `item_id` is now an autoinc pk
---------
Co-authored-by: Mary Hipp <maryhipp@Marys-MacBook-Air.local>
2023-09-20 05:09:24 +00:00
"noPrompts" : "No prompts generated" ,
2023-09-13 07:31:34 +00:00
"noNodesInGraph" : "No nodes in graph" ,
2024-05-10 06:35:04 +00:00
"systemDisconnected" : "System disconnected" ,
2024-05-10 08:08:11 +00:00
"layer" : {
"initialImageNoImageSelected" : "no initial image selected" ,
"controlAdapterNoModelSelected" : "no Control Adapter model selected" ,
"controlAdapterIncompatibleBaseModel" : "incompatible Control Adapter base model" ,
"controlAdapterNoImageSelected" : "no Control Adapter image selected" ,
"controlAdapterImageNotProcessed" : "Control Adapter image not processed" ,
2024-05-17 23:12:10 +00:00
"t2iAdapterIncompatibleDimensions" : "T2I Adapter requires image dimension to be multiples of {{multiple}}" ,
2024-05-10 08:08:11 +00:00
"ipAdapterNoModelSelected" : "no IP adapter selected" ,
"ipAdapterIncompatibleBaseModel" : "incompatible IP Adapter base model" ,
"ipAdapterNoImageSelected" : "no IP Adapter image selected" ,
"rgNoPromptsOrIPAdapters" : "no text prompts or IP Adapters" ,
"rgNoRegion" : "no region selected"
}
2023-09-13 07:31:34 +00:00
} ,
2024-02-26 13:35:11 +00:00
"maskBlur" : "Mask Blur" ,
2023-09-13 07:31:34 +00:00
"negativePromptPlaceholder" : "Negative Prompt" ,
2024-05-09 02:03:31 +00:00
"globalNegativePromptPlaceholder" : "Global Negative Prompt" ,
2023-09-13 07:31:34 +00:00
"noiseThreshold" : "Noise Threshold" ,
"patchmatchDownScaleSize" : "Downscale" ,
"perlinNoise" : "Perlin Noise" ,
"positivePromptPlaceholder" : "Positive Prompt" ,
2024-08-12 17:33:39 +00:00
"globalPositivePromptPlaceholder" : "Global Positive Prompt" ,
feat: queued generation (#4502)
* fix(config): fix typing issues in `config/`
`config/invokeai_config.py`:
- use `Optional` for things that are optional
- fix typing of `ram_cache_size()` and `vram_cache_size()`
- remove unused and incorrectly typed method `autoconvert_path`
- fix types and logic for `parse_args()`, in which `InvokeAIAppConfig.initconf` *must* be a `DictConfig`, but function would allow it to be set as a `ListConfig`, which presumably would cause issues elsewhere
`config/base.py`:
- use `cls` for first arg of class methods
- use `Optional` for things that are optional
- fix minor type issue related to setting of `env_prefix`
- remove unused `add_subparser()` method, which calls `add_parser()` on an `ArgumentParser` (method only available on the `_SubParsersAction` object, which is returned from ArgumentParser.add_subparsers()`)
* feat: queued generation and batches
Due to a very messy branch with broad addition of `isort` on `main` alongside it, some git surgery was needed to get an agreeable git history. This commit represents all of the work on queued generation. See PR for notes.
* chore: flake8, isort, black
* fix(nodes): fix incorrect service stop() method
* fix(nodes): improve names of a few variables
* fix(tests): fix up tests after changes to batches/queue
* feat(tests): add unit tests for session queue helper functions
* feat(ui): dynamic prompts is always enabled
* feat(queue): add queue_status_changed event
* feat(ui): wip queue graphs
* feat(nodes): move cleanup til after invoker startup
* feat(nodes): add cancel_by_batch_ids
* feat(ui): wip batch graphs & UI
* fix(nodes): remove `Batch.batch_id` from required
* fix(ui): cleanup and use fixedCacheKey for all mutations
* fix(ui): remove orphaned nodes from canvas graphs
* fix(nodes): fix cancel_by_batch_ids result count
* fix(ui): only show cancel batch tooltip when batches were canceled
* chore: isort
* fix(api): return `[""]` when dynamic prompts generates no prompts
Just a simple fallback so we always have a prompt.
* feat(ui): dynamicPrompts.combinatorial is always on
There seems to be little purpose in using the combinatorial generation for dynamic prompts. I've disabled it by hiding it from the UI and defaulting combinatorial to true. If we want to enable it again in the future it's straightforward to do so.
* feat: add queue_id & support logic
* feat(ui): fix upscale button
It prepends the upscale operation to queue
* feat(nodes): return queue item when enqueuing a single graph
This facilitates one-off graph async workflows in the client.
* feat(ui): move controlnet autoprocess to queue
* fix(ui): fix non-serializable DOMRect in redux state
* feat(ui): QueueTable performance tweaks
* feat(ui): update queue list
Queue items expand to show the full queue item. Just as JSON for now.
* wip threaded session_processor
* feat(nodes,ui): fully migrate queue to session_processor
* feat(nodes,ui): add processor events
* feat(ui): ui tweaks
* feat(nodes,ui): consolidate events, reduce network requests
* feat(ui): cleanup & abstract queue hooks
* feat(nodes): optimize batch permutation
Use a generator to do only as much work as is needed.
Previously, though we only ended up creating exactly as many queue items as was needed, there was still some intermediary work that calculated *all* permutations. When that number was very high, the system had a very hard time and used a lot of memory.
The logic has been refactored to use a generator. Additionally, the batch validators are optimized to return early and use less memory.
* feat(ui): add seed behaviour parameter
This dynamic prompts parameter allows the seed to be randomized per prompt or per iteration:
- Per iteration: Use the same seed for all prompts in a single dynamic prompt expansion
- Per prompt: Use a different seed for every single prompt
"Per iteration" is appropriate for exploring a the latents space with a stable starting noise, while "Per prompt" provides more variation.
* fix(ui): remove extraneous random seed nodes from linear graphs
* fix(ui): fix controlnet autoprocess not working when queue is running
* feat(queue): add timestamps to queue status updates
Also show execution time in queue list
* feat(queue): change all execution-related events to use the `queue_id` as the room, also include `queue_item_id` in InvocationQueueItem
This allows for much simpler handling of queue items.
* feat(api): deprecate sessions router
* chore(backend): tidy logging in `dependencies.py`
* fix(backend): respect `use_memory_db`
* feat(backend): add `config.log_sql` (enables sql trace logging)
* feat: add invocation cache
Supersedes #4574
The invocation cache provides simple node memoization functionality. Nodes that use the cache are memoized and not re-executed if their inputs haven't changed. Instead, the stored output is returned.
## Results
This feature provides anywhere some significant to massive performance improvement.
The improvement is most marked on large batches of generations where you only change a couple things (e.g. different seed or prompt for each iteration) and low-VRAM systems, where skipping an extraneous model load is a big deal.
## Overview
A new `invocation_cache` service is added to handle the caching. There's not much to it.
All nodes now inherit a boolean `use_cache` field from `BaseInvocation`. This is a node field and not a class attribute, because specific instances of nodes may want to opt in or out of caching.
The recently-added `invoke_internal()` method on `BaseInvocation` is used as an entrypoint for the cache logic.
To create a cache key, the invocation is first serialized using pydantic's provided `json()` method, skipping the unique `id` field. Then python's very fast builtin `hash()` is used to create an integer key. All implementations of `InvocationCacheBase` must provide a class method `create_key()` which accepts an invocation and outputs a string or integer key.
## In-Memory Implementation
An in-memory implementation is provided. In this implementation, the node outputs are stored in memory as python classes. The in-memory cache does not persist application restarts.
Max node cache size is added as `node_cache_size` under the `Generation` config category.
It defaults to 512 - this number is up for discussion, but given that these are relatively lightweight pydantic models, I think it's safe to up this even higher.
Note that the cache isn't storing the big stuff - tensors and images are store on disk, and outputs include only references to them.
## Node Definition
The default for all nodes is to use the cache. The `@invocation` decorator now accepts an optional `use_cache: bool` argument to override the default of `True`.
Non-deterministic nodes, however, should set this to `False`. Currently, all random-stuff nodes, including `dynamic_prompt`, are set to `False`.
The field name `use_cache` is now effectively a reserved field name and possibly a breaking change if any community nodes use this as a field name. In hindsight, all our reserved field names should have been prefixed with underscores or something.
## One Gotcha
Leaf nodes probably want to opt out of the cache, because if they are not cached, their outputs are not saved again.
If you run the same graph multiple times, you only end up with a single image output, because the image storage side-effects are in the `invoke()` method, which is bypassed if we have a cache hit.
## Linear UI
The linear graphs _almost_ just work, but due to the gotcha, we need to be careful about the final image-outputting node. To resolve this, a `SaveImageInvocation` node is added and used in the linear graphs.
This node is similar to `ImagePrimitive`, except it saves a copy of its input image, and has `use_cache` set to `False` by default.
This is now the leaf node in all linear graphs, and is the only node in those graphs with `use_cache == False` _and_ the only node with `is_intermedate == False`.
## Workflow Editor
All nodes now have a footer with a new `Use Cache [ ]` checkbox. It defaults to the value set by the invocation in its python definition, but can be changed by the user.
The workflow/node validation logic has been updated to migrate old workflows to use the new default values for `use_cache`. Users may still want to review the settings that have been chosen. In the event of catastrophic failure when running this migration, the default value of `True` is applied, as this is correct for most nodes.
Users should consider saving their workflows after loading them in and having them updated.
## Future Enhancements - Callback
A future enhancement would be to provide a callback to the `use_cache` flag that would be run as the node is executed to determine, based on its own internal state, if the cache should be used or not.
This would be useful for `DynamicPromptInvocation`, where the deterministic behaviour is determined by the `combinatorial: bool` field.
## Future Enhancements - Persisted Cache
Similar to how the latents storage is backed by disk, the invocation cache could be persisted to the database or disk. We'd need to be very careful about deserializing outputs, but it's perhaps worth exploring in the future.
* fix(ui): fix queue list item width
* feat(nodes): do not send the whole node on every generator progress
* feat(ui): strip out old logic related to sessions
Things like `isProcessing` are no longer relevant with queue. Removed them all & updated everything be appropriate for queue. May be a few little quirks I've missed...
* feat(ui): fix up param collapse labels
* feat(ui): click queue count to go to queue tab
* tidy(queue): update comment, query format
* feat(ui): fix progress bar when canceling
* fix(ui): fix circular dependency
* feat(nodes): bail on node caching logic if `node_cache_size == 0`
* feat(nodes): handle KeyError on node cache pop
* feat(nodes): bypass cache codepath if caches is disabled
more better no do thing
* fix(ui): reset api cache on connect/disconnect
* feat(ui): prevent enqueue when no prompts generated
* feat(ui): add queue controls to workflow editor
* feat(ui): update floating buttons & other incidental UI tweaks
* fix(ui): fix missing/incorrect translation keys
* fix(tests): add config service to mock invocation services
invoking needs access to `node_cache_size` to occur
* optionally remove pause/resume buttons from queue UI
* option to disable prepending
* chore(ui): remove unused file
* feat(queue): remove `order_id` entirely, `item_id` is now an autoinc pk
---------
Co-authored-by: Mary Hipp <maryhipp@Marys-MacBook-Air.local>
2023-09-20 05:09:24 +00:00
"iterations" : "Iterations" ,
2023-09-13 07:31:34 +00:00
"scale" : "Scale" ,
"scaleBeforeProcessing" : "Scale Before Processing" ,
"scaledHeight" : "Scaled H" ,
"scaledWidth" : "Scaled W" ,
"scheduler" : "Scheduler" ,
2023-12-28 13:03:21 +00:00
"seamlessXAxis" : "Seamless Tiling X Axis" ,
"seamlessYAxis" : "Seamless Tiling Y Axis" ,
2023-09-13 07:31:34 +00:00
"seed" : "Seed" ,
feat: queued generation (#4502)
* fix(config): fix typing issues in `config/`
`config/invokeai_config.py`:
- use `Optional` for things that are optional
- fix typing of `ram_cache_size()` and `vram_cache_size()`
- remove unused and incorrectly typed method `autoconvert_path`
- fix types and logic for `parse_args()`, in which `InvokeAIAppConfig.initconf` *must* be a `DictConfig`, but function would allow it to be set as a `ListConfig`, which presumably would cause issues elsewhere
`config/base.py`:
- use `cls` for first arg of class methods
- use `Optional` for things that are optional
- fix minor type issue related to setting of `env_prefix`
- remove unused `add_subparser()` method, which calls `add_parser()` on an `ArgumentParser` (method only available on the `_SubParsersAction` object, which is returned from ArgumentParser.add_subparsers()`)
* feat: queued generation and batches
Due to a very messy branch with broad addition of `isort` on `main` alongside it, some git surgery was needed to get an agreeable git history. This commit represents all of the work on queued generation. See PR for notes.
* chore: flake8, isort, black
* fix(nodes): fix incorrect service stop() method
* fix(nodes): improve names of a few variables
* fix(tests): fix up tests after changes to batches/queue
* feat(tests): add unit tests for session queue helper functions
* feat(ui): dynamic prompts is always enabled
* feat(queue): add queue_status_changed event
* feat(ui): wip queue graphs
* feat(nodes): move cleanup til after invoker startup
* feat(nodes): add cancel_by_batch_ids
* feat(ui): wip batch graphs & UI
* fix(nodes): remove `Batch.batch_id` from required
* fix(ui): cleanup and use fixedCacheKey for all mutations
* fix(ui): remove orphaned nodes from canvas graphs
* fix(nodes): fix cancel_by_batch_ids result count
* fix(ui): only show cancel batch tooltip when batches were canceled
* chore: isort
* fix(api): return `[""]` when dynamic prompts generates no prompts
Just a simple fallback so we always have a prompt.
* feat(ui): dynamicPrompts.combinatorial is always on
There seems to be little purpose in using the combinatorial generation for dynamic prompts. I've disabled it by hiding it from the UI and defaulting combinatorial to true. If we want to enable it again in the future it's straightforward to do so.
* feat: add queue_id & support logic
* feat(ui): fix upscale button
It prepends the upscale operation to queue
* feat(nodes): return queue item when enqueuing a single graph
This facilitates one-off graph async workflows in the client.
* feat(ui): move controlnet autoprocess to queue
* fix(ui): fix non-serializable DOMRect in redux state
* feat(ui): QueueTable performance tweaks
* feat(ui): update queue list
Queue items expand to show the full queue item. Just as JSON for now.
* wip threaded session_processor
* feat(nodes,ui): fully migrate queue to session_processor
* feat(nodes,ui): add processor events
* feat(ui): ui tweaks
* feat(nodes,ui): consolidate events, reduce network requests
* feat(ui): cleanup & abstract queue hooks
* feat(nodes): optimize batch permutation
Use a generator to do only as much work as is needed.
Previously, though we only ended up creating exactly as many queue items as was needed, there was still some intermediary work that calculated *all* permutations. When that number was very high, the system had a very hard time and used a lot of memory.
The logic has been refactored to use a generator. Additionally, the batch validators are optimized to return early and use less memory.
* feat(ui): add seed behaviour parameter
This dynamic prompts parameter allows the seed to be randomized per prompt or per iteration:
- Per iteration: Use the same seed for all prompts in a single dynamic prompt expansion
- Per prompt: Use a different seed for every single prompt
"Per iteration" is appropriate for exploring a the latents space with a stable starting noise, while "Per prompt" provides more variation.
* fix(ui): remove extraneous random seed nodes from linear graphs
* fix(ui): fix controlnet autoprocess not working when queue is running
* feat(queue): add timestamps to queue status updates
Also show execution time in queue list
* feat(queue): change all execution-related events to use the `queue_id` as the room, also include `queue_item_id` in InvocationQueueItem
This allows for much simpler handling of queue items.
* feat(api): deprecate sessions router
* chore(backend): tidy logging in `dependencies.py`
* fix(backend): respect `use_memory_db`
* feat(backend): add `config.log_sql` (enables sql trace logging)
* feat: add invocation cache
Supersedes #4574
The invocation cache provides simple node memoization functionality. Nodes that use the cache are memoized and not re-executed if their inputs haven't changed. Instead, the stored output is returned.
## Results
This feature provides anywhere some significant to massive performance improvement.
The improvement is most marked on large batches of generations where you only change a couple things (e.g. different seed or prompt for each iteration) and low-VRAM systems, where skipping an extraneous model load is a big deal.
## Overview
A new `invocation_cache` service is added to handle the caching. There's not much to it.
All nodes now inherit a boolean `use_cache` field from `BaseInvocation`. This is a node field and not a class attribute, because specific instances of nodes may want to opt in or out of caching.
The recently-added `invoke_internal()` method on `BaseInvocation` is used as an entrypoint for the cache logic.
To create a cache key, the invocation is first serialized using pydantic's provided `json()` method, skipping the unique `id` field. Then python's very fast builtin `hash()` is used to create an integer key. All implementations of `InvocationCacheBase` must provide a class method `create_key()` which accepts an invocation and outputs a string or integer key.
## In-Memory Implementation
An in-memory implementation is provided. In this implementation, the node outputs are stored in memory as python classes. The in-memory cache does not persist application restarts.
Max node cache size is added as `node_cache_size` under the `Generation` config category.
It defaults to 512 - this number is up for discussion, but given that these are relatively lightweight pydantic models, I think it's safe to up this even higher.
Note that the cache isn't storing the big stuff - tensors and images are store on disk, and outputs include only references to them.
## Node Definition
The default for all nodes is to use the cache. The `@invocation` decorator now accepts an optional `use_cache: bool` argument to override the default of `True`.
Non-deterministic nodes, however, should set this to `False`. Currently, all random-stuff nodes, including `dynamic_prompt`, are set to `False`.
The field name `use_cache` is now effectively a reserved field name and possibly a breaking change if any community nodes use this as a field name. In hindsight, all our reserved field names should have been prefixed with underscores or something.
## One Gotcha
Leaf nodes probably want to opt out of the cache, because if they are not cached, their outputs are not saved again.
If you run the same graph multiple times, you only end up with a single image output, because the image storage side-effects are in the `invoke()` method, which is bypassed if we have a cache hit.
## Linear UI
The linear graphs _almost_ just work, but due to the gotcha, we need to be careful about the final image-outputting node. To resolve this, a `SaveImageInvocation` node is added and used in the linear graphs.
This node is similar to `ImagePrimitive`, except it saves a copy of its input image, and has `use_cache` set to `False` by default.
This is now the leaf node in all linear graphs, and is the only node in those graphs with `use_cache == False` _and_ the only node with `is_intermedate == False`.
## Workflow Editor
All nodes now have a footer with a new `Use Cache [ ]` checkbox. It defaults to the value set by the invocation in its python definition, but can be changed by the user.
The workflow/node validation logic has been updated to migrate old workflows to use the new default values for `use_cache`. Users may still want to review the settings that have been chosen. In the event of catastrophic failure when running this migration, the default value of `True` is applied, as this is correct for most nodes.
Users should consider saving their workflows after loading them in and having them updated.
## Future Enhancements - Callback
A future enhancement would be to provide a callback to the `use_cache` flag that would be run as the node is executed to determine, based on its own internal state, if the cache should be used or not.
This would be useful for `DynamicPromptInvocation`, where the deterministic behaviour is determined by the `combinatorial: bool` field.
## Future Enhancements - Persisted Cache
Similar to how the latents storage is backed by disk, the invocation cache could be persisted to the database or disk. We'd need to be very careful about deserializing outputs, but it's perhaps worth exploring in the future.
* fix(ui): fix queue list item width
* feat(nodes): do not send the whole node on every generator progress
* feat(ui): strip out old logic related to sessions
Things like `isProcessing` are no longer relevant with queue. Removed them all & updated everything be appropriate for queue. May be a few little quirks I've missed...
* feat(ui): fix up param collapse labels
* feat(ui): click queue count to go to queue tab
* tidy(queue): update comment, query format
* feat(ui): fix progress bar when canceling
* fix(ui): fix circular dependency
* feat(nodes): bail on node caching logic if `node_cache_size == 0`
* feat(nodes): handle KeyError on node cache pop
* feat(nodes): bypass cache codepath if caches is disabled
more better no do thing
* fix(ui): reset api cache on connect/disconnect
* feat(ui): prevent enqueue when no prompts generated
* feat(ui): add queue controls to workflow editor
* feat(ui): update floating buttons & other incidental UI tweaks
* fix(ui): fix missing/incorrect translation keys
* fix(tests): add config service to mock invocation services
invoking needs access to `node_cache_size` to occur
* optionally remove pause/resume buttons from queue UI
* option to disable prepending
* chore(ui): remove unused file
* feat(queue): remove `order_id` entirely, `item_id` is now an autoinc pk
---------
Co-authored-by: Mary Hipp <maryhipp@Marys-MacBook-Air.local>
2023-09-20 05:09:24 +00:00
"imageActions" : "Image Actions" ,
2023-09-13 07:31:34 +00:00
"sendToImg2Img" : "Send to Image to Image" ,
"sendToUnifiedCanvas" : "Send To Unified Canvas" ,
2024-07-19 00:26:17 +00:00
"sendToUpscale" : "Send To Upscale" ,
feat: queued generation (#4502)
* fix(config): fix typing issues in `config/`
`config/invokeai_config.py`:
- use `Optional` for things that are optional
- fix typing of `ram_cache_size()` and `vram_cache_size()`
- remove unused and incorrectly typed method `autoconvert_path`
- fix types and logic for `parse_args()`, in which `InvokeAIAppConfig.initconf` *must* be a `DictConfig`, but function would allow it to be set as a `ListConfig`, which presumably would cause issues elsewhere
`config/base.py`:
- use `cls` for first arg of class methods
- use `Optional` for things that are optional
- fix minor type issue related to setting of `env_prefix`
- remove unused `add_subparser()` method, which calls `add_parser()` on an `ArgumentParser` (method only available on the `_SubParsersAction` object, which is returned from ArgumentParser.add_subparsers()`)
* feat: queued generation and batches
Due to a very messy branch with broad addition of `isort` on `main` alongside it, some git surgery was needed to get an agreeable git history. This commit represents all of the work on queued generation. See PR for notes.
* chore: flake8, isort, black
* fix(nodes): fix incorrect service stop() method
* fix(nodes): improve names of a few variables
* fix(tests): fix up tests after changes to batches/queue
* feat(tests): add unit tests for session queue helper functions
* feat(ui): dynamic prompts is always enabled
* feat(queue): add queue_status_changed event
* feat(ui): wip queue graphs
* feat(nodes): move cleanup til after invoker startup
* feat(nodes): add cancel_by_batch_ids
* feat(ui): wip batch graphs & UI
* fix(nodes): remove `Batch.batch_id` from required
* fix(ui): cleanup and use fixedCacheKey for all mutations
* fix(ui): remove orphaned nodes from canvas graphs
* fix(nodes): fix cancel_by_batch_ids result count
* fix(ui): only show cancel batch tooltip when batches were canceled
* chore: isort
* fix(api): return `[""]` when dynamic prompts generates no prompts
Just a simple fallback so we always have a prompt.
* feat(ui): dynamicPrompts.combinatorial is always on
There seems to be little purpose in using the combinatorial generation for dynamic prompts. I've disabled it by hiding it from the UI and defaulting combinatorial to true. If we want to enable it again in the future it's straightforward to do so.
* feat: add queue_id & support logic
* feat(ui): fix upscale button
It prepends the upscale operation to queue
* feat(nodes): return queue item when enqueuing a single graph
This facilitates one-off graph async workflows in the client.
* feat(ui): move controlnet autoprocess to queue
* fix(ui): fix non-serializable DOMRect in redux state
* feat(ui): QueueTable performance tweaks
* feat(ui): update queue list
Queue items expand to show the full queue item. Just as JSON for now.
* wip threaded session_processor
* feat(nodes,ui): fully migrate queue to session_processor
* feat(nodes,ui): add processor events
* feat(ui): ui tweaks
* feat(nodes,ui): consolidate events, reduce network requests
* feat(ui): cleanup & abstract queue hooks
* feat(nodes): optimize batch permutation
Use a generator to do only as much work as is needed.
Previously, though we only ended up creating exactly as many queue items as was needed, there was still some intermediary work that calculated *all* permutations. When that number was very high, the system had a very hard time and used a lot of memory.
The logic has been refactored to use a generator. Additionally, the batch validators are optimized to return early and use less memory.
* feat(ui): add seed behaviour parameter
This dynamic prompts parameter allows the seed to be randomized per prompt or per iteration:
- Per iteration: Use the same seed for all prompts in a single dynamic prompt expansion
- Per prompt: Use a different seed for every single prompt
"Per iteration" is appropriate for exploring a the latents space with a stable starting noise, while "Per prompt" provides more variation.
* fix(ui): remove extraneous random seed nodes from linear graphs
* fix(ui): fix controlnet autoprocess not working when queue is running
* feat(queue): add timestamps to queue status updates
Also show execution time in queue list
* feat(queue): change all execution-related events to use the `queue_id` as the room, also include `queue_item_id` in InvocationQueueItem
This allows for much simpler handling of queue items.
* feat(api): deprecate sessions router
* chore(backend): tidy logging in `dependencies.py`
* fix(backend): respect `use_memory_db`
* feat(backend): add `config.log_sql` (enables sql trace logging)
* feat: add invocation cache
Supersedes #4574
The invocation cache provides simple node memoization functionality. Nodes that use the cache are memoized and not re-executed if their inputs haven't changed. Instead, the stored output is returned.
## Results
This feature provides anywhere some significant to massive performance improvement.
The improvement is most marked on large batches of generations where you only change a couple things (e.g. different seed or prompt for each iteration) and low-VRAM systems, where skipping an extraneous model load is a big deal.
## Overview
A new `invocation_cache` service is added to handle the caching. There's not much to it.
All nodes now inherit a boolean `use_cache` field from `BaseInvocation`. This is a node field and not a class attribute, because specific instances of nodes may want to opt in or out of caching.
The recently-added `invoke_internal()` method on `BaseInvocation` is used as an entrypoint for the cache logic.
To create a cache key, the invocation is first serialized using pydantic's provided `json()` method, skipping the unique `id` field. Then python's very fast builtin `hash()` is used to create an integer key. All implementations of `InvocationCacheBase` must provide a class method `create_key()` which accepts an invocation and outputs a string or integer key.
## In-Memory Implementation
An in-memory implementation is provided. In this implementation, the node outputs are stored in memory as python classes. The in-memory cache does not persist application restarts.
Max node cache size is added as `node_cache_size` under the `Generation` config category.
It defaults to 512 - this number is up for discussion, but given that these are relatively lightweight pydantic models, I think it's safe to up this even higher.
Note that the cache isn't storing the big stuff - tensors and images are store on disk, and outputs include only references to them.
## Node Definition
The default for all nodes is to use the cache. The `@invocation` decorator now accepts an optional `use_cache: bool` argument to override the default of `True`.
Non-deterministic nodes, however, should set this to `False`. Currently, all random-stuff nodes, including `dynamic_prompt`, are set to `False`.
The field name `use_cache` is now effectively a reserved field name and possibly a breaking change if any community nodes use this as a field name. In hindsight, all our reserved field names should have been prefixed with underscores or something.
## One Gotcha
Leaf nodes probably want to opt out of the cache, because if they are not cached, their outputs are not saved again.
If you run the same graph multiple times, you only end up with a single image output, because the image storage side-effects are in the `invoke()` method, which is bypassed if we have a cache hit.
## Linear UI
The linear graphs _almost_ just work, but due to the gotcha, we need to be careful about the final image-outputting node. To resolve this, a `SaveImageInvocation` node is added and used in the linear graphs.
This node is similar to `ImagePrimitive`, except it saves a copy of its input image, and has `use_cache` set to `False` by default.
This is now the leaf node in all linear graphs, and is the only node in those graphs with `use_cache == False` _and_ the only node with `is_intermedate == False`.
## Workflow Editor
All nodes now have a footer with a new `Use Cache [ ]` checkbox. It defaults to the value set by the invocation in its python definition, but can be changed by the user.
The workflow/node validation logic has been updated to migrate old workflows to use the new default values for `use_cache`. Users may still want to review the settings that have been chosen. In the event of catastrophic failure when running this migration, the default value of `True` is applied, as this is correct for most nodes.
Users should consider saving their workflows after loading them in and having them updated.
## Future Enhancements - Callback
A future enhancement would be to provide a callback to the `use_cache` flag that would be run as the node is executed to determine, based on its own internal state, if the cache should be used or not.
This would be useful for `DynamicPromptInvocation`, where the deterministic behaviour is determined by the `combinatorial: bool` field.
## Future Enhancements - Persisted Cache
Similar to how the latents storage is backed by disk, the invocation cache could be persisted to the database or disk. We'd need to be very careful about deserializing outputs, but it's perhaps worth exploring in the future.
* fix(ui): fix queue list item width
* feat(nodes): do not send the whole node on every generator progress
* feat(ui): strip out old logic related to sessions
Things like `isProcessing` are no longer relevant with queue. Removed them all & updated everything be appropriate for queue. May be a few little quirks I've missed...
* feat(ui): fix up param collapse labels
* feat(ui): click queue count to go to queue tab
* tidy(queue): update comment, query format
* feat(ui): fix progress bar when canceling
* fix(ui): fix circular dependency
* feat(nodes): bail on node caching logic if `node_cache_size == 0`
* feat(nodes): handle KeyError on node cache pop
* feat(nodes): bypass cache codepath if caches is disabled
more better no do thing
* fix(ui): reset api cache on connect/disconnect
* feat(ui): prevent enqueue when no prompts generated
* feat(ui): add queue controls to workflow editor
* feat(ui): update floating buttons & other incidental UI tweaks
* fix(ui): fix missing/incorrect translation keys
* fix(tests): add config service to mock invocation services
invoking needs access to `node_cache_size` to occur
* optionally remove pause/resume buttons from queue UI
* option to disable prepending
* chore(ui): remove unused file
* feat(queue): remove `order_id` entirely, `item_id` is now an autoinc pk
---------
Co-authored-by: Mary Hipp <maryhipp@Marys-MacBook-Air.local>
2023-09-20 05:09:24 +00:00
"showOptionsPanel" : "Show Side Panel (O or T)" ,
2023-09-13 07:31:34 +00:00
"shuffle" : "Shuffle Seed" ,
"steps" : "Steps" ,
"strength" : "Strength" ,
"symmetry" : "Symmetry" ,
"tileSize" : "Tile Size" ,
"type" : "Type" ,
2024-07-23 21:32:33 +00:00
"postProcessing" : "Post-Processing (Shift + U)" ,
"processImage" : "Process Image" ,
2023-09-13 07:31:34 +00:00
"upscaling" : "Upscaling" ,
"useAll" : "Use All" ,
2023-11-25 03:58:11 +00:00
"useSize" : "Use Size" ,
2023-09-13 07:31:34 +00:00
"useCpuNoise" : "Use CPU Noise" ,
2024-01-24 00:45:30 +00:00
"remixImage" : "Remix Image" ,
2023-09-13 07:31:34 +00:00
"usePrompt" : "Use Prompt" ,
"useSeed" : "Use Seed" ,
2024-08-05 15:58:17 +00:00
"width" : "Width"
2023-09-13 07:31:34 +00:00
} ,
feat: queued generation (#4502)
* fix(config): fix typing issues in `config/`
`config/invokeai_config.py`:
- use `Optional` for things that are optional
- fix typing of `ram_cache_size()` and `vram_cache_size()`
- remove unused and incorrectly typed method `autoconvert_path`
- fix types and logic for `parse_args()`, in which `InvokeAIAppConfig.initconf` *must* be a `DictConfig`, but function would allow it to be set as a `ListConfig`, which presumably would cause issues elsewhere
`config/base.py`:
- use `cls` for first arg of class methods
- use `Optional` for things that are optional
- fix minor type issue related to setting of `env_prefix`
- remove unused `add_subparser()` method, which calls `add_parser()` on an `ArgumentParser` (method only available on the `_SubParsersAction` object, which is returned from ArgumentParser.add_subparsers()`)
* feat: queued generation and batches
Due to a very messy branch with broad addition of `isort` on `main` alongside it, some git surgery was needed to get an agreeable git history. This commit represents all of the work on queued generation. See PR for notes.
* chore: flake8, isort, black
* fix(nodes): fix incorrect service stop() method
* fix(nodes): improve names of a few variables
* fix(tests): fix up tests after changes to batches/queue
* feat(tests): add unit tests for session queue helper functions
* feat(ui): dynamic prompts is always enabled
* feat(queue): add queue_status_changed event
* feat(ui): wip queue graphs
* feat(nodes): move cleanup til after invoker startup
* feat(nodes): add cancel_by_batch_ids
* feat(ui): wip batch graphs & UI
* fix(nodes): remove `Batch.batch_id` from required
* fix(ui): cleanup and use fixedCacheKey for all mutations
* fix(ui): remove orphaned nodes from canvas graphs
* fix(nodes): fix cancel_by_batch_ids result count
* fix(ui): only show cancel batch tooltip when batches were canceled
* chore: isort
* fix(api): return `[""]` when dynamic prompts generates no prompts
Just a simple fallback so we always have a prompt.
* feat(ui): dynamicPrompts.combinatorial is always on
There seems to be little purpose in using the combinatorial generation for dynamic prompts. I've disabled it by hiding it from the UI and defaulting combinatorial to true. If we want to enable it again in the future it's straightforward to do so.
* feat: add queue_id & support logic
* feat(ui): fix upscale button
It prepends the upscale operation to queue
* feat(nodes): return queue item when enqueuing a single graph
This facilitates one-off graph async workflows in the client.
* feat(ui): move controlnet autoprocess to queue
* fix(ui): fix non-serializable DOMRect in redux state
* feat(ui): QueueTable performance tweaks
* feat(ui): update queue list
Queue items expand to show the full queue item. Just as JSON for now.
* wip threaded session_processor
* feat(nodes,ui): fully migrate queue to session_processor
* feat(nodes,ui): add processor events
* feat(ui): ui tweaks
* feat(nodes,ui): consolidate events, reduce network requests
* feat(ui): cleanup & abstract queue hooks
* feat(nodes): optimize batch permutation
Use a generator to do only as much work as is needed.
Previously, though we only ended up creating exactly as many queue items as was needed, there was still some intermediary work that calculated *all* permutations. When that number was very high, the system had a very hard time and used a lot of memory.
The logic has been refactored to use a generator. Additionally, the batch validators are optimized to return early and use less memory.
* feat(ui): add seed behaviour parameter
This dynamic prompts parameter allows the seed to be randomized per prompt or per iteration:
- Per iteration: Use the same seed for all prompts in a single dynamic prompt expansion
- Per prompt: Use a different seed for every single prompt
"Per iteration" is appropriate for exploring a the latents space with a stable starting noise, while "Per prompt" provides more variation.
* fix(ui): remove extraneous random seed nodes from linear graphs
* fix(ui): fix controlnet autoprocess not working when queue is running
* feat(queue): add timestamps to queue status updates
Also show execution time in queue list
* feat(queue): change all execution-related events to use the `queue_id` as the room, also include `queue_item_id` in InvocationQueueItem
This allows for much simpler handling of queue items.
* feat(api): deprecate sessions router
* chore(backend): tidy logging in `dependencies.py`
* fix(backend): respect `use_memory_db`
* feat(backend): add `config.log_sql` (enables sql trace logging)
* feat: add invocation cache
Supersedes #4574
The invocation cache provides simple node memoization functionality. Nodes that use the cache are memoized and not re-executed if their inputs haven't changed. Instead, the stored output is returned.
## Results
This feature provides anywhere some significant to massive performance improvement.
The improvement is most marked on large batches of generations where you only change a couple things (e.g. different seed or prompt for each iteration) and low-VRAM systems, where skipping an extraneous model load is a big deal.
## Overview
A new `invocation_cache` service is added to handle the caching. There's not much to it.
All nodes now inherit a boolean `use_cache` field from `BaseInvocation`. This is a node field and not a class attribute, because specific instances of nodes may want to opt in or out of caching.
The recently-added `invoke_internal()` method on `BaseInvocation` is used as an entrypoint for the cache logic.
To create a cache key, the invocation is first serialized using pydantic's provided `json()` method, skipping the unique `id` field. Then python's very fast builtin `hash()` is used to create an integer key. All implementations of `InvocationCacheBase` must provide a class method `create_key()` which accepts an invocation and outputs a string or integer key.
## In-Memory Implementation
An in-memory implementation is provided. In this implementation, the node outputs are stored in memory as python classes. The in-memory cache does not persist application restarts.
Max node cache size is added as `node_cache_size` under the `Generation` config category.
It defaults to 512 - this number is up for discussion, but given that these are relatively lightweight pydantic models, I think it's safe to up this even higher.
Note that the cache isn't storing the big stuff - tensors and images are store on disk, and outputs include only references to them.
## Node Definition
The default for all nodes is to use the cache. The `@invocation` decorator now accepts an optional `use_cache: bool` argument to override the default of `True`.
Non-deterministic nodes, however, should set this to `False`. Currently, all random-stuff nodes, including `dynamic_prompt`, are set to `False`.
The field name `use_cache` is now effectively a reserved field name and possibly a breaking change if any community nodes use this as a field name. In hindsight, all our reserved field names should have been prefixed with underscores or something.
## One Gotcha
Leaf nodes probably want to opt out of the cache, because if they are not cached, their outputs are not saved again.
If you run the same graph multiple times, you only end up with a single image output, because the image storage side-effects are in the `invoke()` method, which is bypassed if we have a cache hit.
## Linear UI
The linear graphs _almost_ just work, but due to the gotcha, we need to be careful about the final image-outputting node. To resolve this, a `SaveImageInvocation` node is added and used in the linear graphs.
This node is similar to `ImagePrimitive`, except it saves a copy of its input image, and has `use_cache` set to `False` by default.
This is now the leaf node in all linear graphs, and is the only node in those graphs with `use_cache == False` _and_ the only node with `is_intermedate == False`.
## Workflow Editor
All nodes now have a footer with a new `Use Cache [ ]` checkbox. It defaults to the value set by the invocation in its python definition, but can be changed by the user.
The workflow/node validation logic has been updated to migrate old workflows to use the new default values for `use_cache`. Users may still want to review the settings that have been chosen. In the event of catastrophic failure when running this migration, the default value of `True` is applied, as this is correct for most nodes.
Users should consider saving their workflows after loading them in and having them updated.
## Future Enhancements - Callback
A future enhancement would be to provide a callback to the `use_cache` flag that would be run as the node is executed to determine, based on its own internal state, if the cache should be used or not.
This would be useful for `DynamicPromptInvocation`, where the deterministic behaviour is determined by the `combinatorial: bool` field.
## Future Enhancements - Persisted Cache
Similar to how the latents storage is backed by disk, the invocation cache could be persisted to the database or disk. We'd need to be very careful about deserializing outputs, but it's perhaps worth exploring in the future.
* fix(ui): fix queue list item width
* feat(nodes): do not send the whole node on every generator progress
* feat(ui): strip out old logic related to sessions
Things like `isProcessing` are no longer relevant with queue. Removed them all & updated everything be appropriate for queue. May be a few little quirks I've missed...
* feat(ui): fix up param collapse labels
* feat(ui): click queue count to go to queue tab
* tidy(queue): update comment, query format
* feat(ui): fix progress bar when canceling
* fix(ui): fix circular dependency
* feat(nodes): bail on node caching logic if `node_cache_size == 0`
* feat(nodes): handle KeyError on node cache pop
* feat(nodes): bypass cache codepath if caches is disabled
more better no do thing
* fix(ui): reset api cache on connect/disconnect
* feat(ui): prevent enqueue when no prompts generated
* feat(ui): add queue controls to workflow editor
* feat(ui): update floating buttons & other incidental UI tweaks
* fix(ui): fix missing/incorrect translation keys
* fix(tests): add config service to mock invocation services
invoking needs access to `node_cache_size` to occur
* optionally remove pause/resume buttons from queue UI
* option to disable prepending
* chore(ui): remove unused file
* feat(queue): remove `order_id` entirely, `item_id` is now an autoinc pk
---------
Co-authored-by: Mary Hipp <maryhipp@Marys-MacBook-Air.local>
2023-09-20 05:09:24 +00:00
"dynamicPrompts" : {
2023-12-28 13:03:21 +00:00
"showDynamicPrompts" : "Show Dynamic Prompts" ,
2023-09-13 07:31:34 +00:00
"dynamicPrompts" : "Dynamic Prompts" ,
feat: queued generation (#4502)
* fix(config): fix typing issues in `config/`
`config/invokeai_config.py`:
- use `Optional` for things that are optional
- fix typing of `ram_cache_size()` and `vram_cache_size()`
- remove unused and incorrectly typed method `autoconvert_path`
- fix types and logic for `parse_args()`, in which `InvokeAIAppConfig.initconf` *must* be a `DictConfig`, but function would allow it to be set as a `ListConfig`, which presumably would cause issues elsewhere
`config/base.py`:
- use `cls` for first arg of class methods
- use `Optional` for things that are optional
- fix minor type issue related to setting of `env_prefix`
- remove unused `add_subparser()` method, which calls `add_parser()` on an `ArgumentParser` (method only available on the `_SubParsersAction` object, which is returned from ArgumentParser.add_subparsers()`)
* feat: queued generation and batches
Due to a very messy branch with broad addition of `isort` on `main` alongside it, some git surgery was needed to get an agreeable git history. This commit represents all of the work on queued generation. See PR for notes.
* chore: flake8, isort, black
* fix(nodes): fix incorrect service stop() method
* fix(nodes): improve names of a few variables
* fix(tests): fix up tests after changes to batches/queue
* feat(tests): add unit tests for session queue helper functions
* feat(ui): dynamic prompts is always enabled
* feat(queue): add queue_status_changed event
* feat(ui): wip queue graphs
* feat(nodes): move cleanup til after invoker startup
* feat(nodes): add cancel_by_batch_ids
* feat(ui): wip batch graphs & UI
* fix(nodes): remove `Batch.batch_id` from required
* fix(ui): cleanup and use fixedCacheKey for all mutations
* fix(ui): remove orphaned nodes from canvas graphs
* fix(nodes): fix cancel_by_batch_ids result count
* fix(ui): only show cancel batch tooltip when batches were canceled
* chore: isort
* fix(api): return `[""]` when dynamic prompts generates no prompts
Just a simple fallback so we always have a prompt.
* feat(ui): dynamicPrompts.combinatorial is always on
There seems to be little purpose in using the combinatorial generation for dynamic prompts. I've disabled it by hiding it from the UI and defaulting combinatorial to true. If we want to enable it again in the future it's straightforward to do so.
* feat: add queue_id & support logic
* feat(ui): fix upscale button
It prepends the upscale operation to queue
* feat(nodes): return queue item when enqueuing a single graph
This facilitates one-off graph async workflows in the client.
* feat(ui): move controlnet autoprocess to queue
* fix(ui): fix non-serializable DOMRect in redux state
* feat(ui): QueueTable performance tweaks
* feat(ui): update queue list
Queue items expand to show the full queue item. Just as JSON for now.
* wip threaded session_processor
* feat(nodes,ui): fully migrate queue to session_processor
* feat(nodes,ui): add processor events
* feat(ui): ui tweaks
* feat(nodes,ui): consolidate events, reduce network requests
* feat(ui): cleanup & abstract queue hooks
* feat(nodes): optimize batch permutation
Use a generator to do only as much work as is needed.
Previously, though we only ended up creating exactly as many queue items as was needed, there was still some intermediary work that calculated *all* permutations. When that number was very high, the system had a very hard time and used a lot of memory.
The logic has been refactored to use a generator. Additionally, the batch validators are optimized to return early and use less memory.
* feat(ui): add seed behaviour parameter
This dynamic prompts parameter allows the seed to be randomized per prompt or per iteration:
- Per iteration: Use the same seed for all prompts in a single dynamic prompt expansion
- Per prompt: Use a different seed for every single prompt
"Per iteration" is appropriate for exploring a the latents space with a stable starting noise, while "Per prompt" provides more variation.
* fix(ui): remove extraneous random seed nodes from linear graphs
* fix(ui): fix controlnet autoprocess not working when queue is running
* feat(queue): add timestamps to queue status updates
Also show execution time in queue list
* feat(queue): change all execution-related events to use the `queue_id` as the room, also include `queue_item_id` in InvocationQueueItem
This allows for much simpler handling of queue items.
* feat(api): deprecate sessions router
* chore(backend): tidy logging in `dependencies.py`
* fix(backend): respect `use_memory_db`
* feat(backend): add `config.log_sql` (enables sql trace logging)
* feat: add invocation cache
Supersedes #4574
The invocation cache provides simple node memoization functionality. Nodes that use the cache are memoized and not re-executed if their inputs haven't changed. Instead, the stored output is returned.
## Results
This feature provides anywhere some significant to massive performance improvement.
The improvement is most marked on large batches of generations where you only change a couple things (e.g. different seed or prompt for each iteration) and low-VRAM systems, where skipping an extraneous model load is a big deal.
## Overview
A new `invocation_cache` service is added to handle the caching. There's not much to it.
All nodes now inherit a boolean `use_cache` field from `BaseInvocation`. This is a node field and not a class attribute, because specific instances of nodes may want to opt in or out of caching.
The recently-added `invoke_internal()` method on `BaseInvocation` is used as an entrypoint for the cache logic.
To create a cache key, the invocation is first serialized using pydantic's provided `json()` method, skipping the unique `id` field. Then python's very fast builtin `hash()` is used to create an integer key. All implementations of `InvocationCacheBase` must provide a class method `create_key()` which accepts an invocation and outputs a string or integer key.
## In-Memory Implementation
An in-memory implementation is provided. In this implementation, the node outputs are stored in memory as python classes. The in-memory cache does not persist application restarts.
Max node cache size is added as `node_cache_size` under the `Generation` config category.
It defaults to 512 - this number is up for discussion, but given that these are relatively lightweight pydantic models, I think it's safe to up this even higher.
Note that the cache isn't storing the big stuff - tensors and images are store on disk, and outputs include only references to them.
## Node Definition
The default for all nodes is to use the cache. The `@invocation` decorator now accepts an optional `use_cache: bool` argument to override the default of `True`.
Non-deterministic nodes, however, should set this to `False`. Currently, all random-stuff nodes, including `dynamic_prompt`, are set to `False`.
The field name `use_cache` is now effectively a reserved field name and possibly a breaking change if any community nodes use this as a field name. In hindsight, all our reserved field names should have been prefixed with underscores or something.
## One Gotcha
Leaf nodes probably want to opt out of the cache, because if they are not cached, their outputs are not saved again.
If you run the same graph multiple times, you only end up with a single image output, because the image storage side-effects are in the `invoke()` method, which is bypassed if we have a cache hit.
## Linear UI
The linear graphs _almost_ just work, but due to the gotcha, we need to be careful about the final image-outputting node. To resolve this, a `SaveImageInvocation` node is added and used in the linear graphs.
This node is similar to `ImagePrimitive`, except it saves a copy of its input image, and has `use_cache` set to `False` by default.
This is now the leaf node in all linear graphs, and is the only node in those graphs with `use_cache == False` _and_ the only node with `is_intermedate == False`.
## Workflow Editor
All nodes now have a footer with a new `Use Cache [ ]` checkbox. It defaults to the value set by the invocation in its python definition, but can be changed by the user.
The workflow/node validation logic has been updated to migrate old workflows to use the new default values for `use_cache`. Users may still want to review the settings that have been chosen. In the event of catastrophic failure when running this migration, the default value of `True` is applied, as this is correct for most nodes.
Users should consider saving their workflows after loading them in and having them updated.
## Future Enhancements - Callback
A future enhancement would be to provide a callback to the `use_cache` flag that would be run as the node is executed to determine, based on its own internal state, if the cache should be used or not.
This would be useful for `DynamicPromptInvocation`, where the deterministic behaviour is determined by the `combinatorial: bool` field.
## Future Enhancements - Persisted Cache
Similar to how the latents storage is backed by disk, the invocation cache could be persisted to the database or disk. We'd need to be very careful about deserializing outputs, but it's perhaps worth exploring in the future.
* fix(ui): fix queue list item width
* feat(nodes): do not send the whole node on every generator progress
* feat(ui): strip out old logic related to sessions
Things like `isProcessing` are no longer relevant with queue. Removed them all & updated everything be appropriate for queue. May be a few little quirks I've missed...
* feat(ui): fix up param collapse labels
* feat(ui): click queue count to go to queue tab
* tidy(queue): update comment, query format
* feat(ui): fix progress bar when canceling
* fix(ui): fix circular dependency
* feat(nodes): bail on node caching logic if `node_cache_size == 0`
* feat(nodes): handle KeyError on node cache pop
* feat(nodes): bypass cache codepath if caches is disabled
more better no do thing
* fix(ui): reset api cache on connect/disconnect
* feat(ui): prevent enqueue when no prompts generated
* feat(ui): add queue controls to workflow editor
* feat(ui): update floating buttons & other incidental UI tweaks
* fix(ui): fix missing/incorrect translation keys
* fix(tests): add config service to mock invocation services
invoking needs access to `node_cache_size` to occur
* optionally remove pause/resume buttons from queue UI
* option to disable prepending
* chore(ui): remove unused file
* feat(queue): remove `order_id` entirely, `item_id` is now an autoinc pk
---------
Co-authored-by: Mary Hipp <maryhipp@Marys-MacBook-Air.local>
2023-09-20 05:09:24 +00:00
"maxPrompts" : "Max Prompts" ,
2023-11-20 06:24:03 +00:00
"promptsPreview" : "Prompts Preview" ,
feat: queued generation (#4502)
* fix(config): fix typing issues in `config/`
`config/invokeai_config.py`:
- use `Optional` for things that are optional
- fix typing of `ram_cache_size()` and `vram_cache_size()`
- remove unused and incorrectly typed method `autoconvert_path`
- fix types and logic for `parse_args()`, in which `InvokeAIAppConfig.initconf` *must* be a `DictConfig`, but function would allow it to be set as a `ListConfig`, which presumably would cause issues elsewhere
`config/base.py`:
- use `cls` for first arg of class methods
- use `Optional` for things that are optional
- fix minor type issue related to setting of `env_prefix`
- remove unused `add_subparser()` method, which calls `add_parser()` on an `ArgumentParser` (method only available on the `_SubParsersAction` object, which is returned from ArgumentParser.add_subparsers()`)
* feat: queued generation and batches
Due to a very messy branch with broad addition of `isort` on `main` alongside it, some git surgery was needed to get an agreeable git history. This commit represents all of the work on queued generation. See PR for notes.
* chore: flake8, isort, black
* fix(nodes): fix incorrect service stop() method
* fix(nodes): improve names of a few variables
* fix(tests): fix up tests after changes to batches/queue
* feat(tests): add unit tests for session queue helper functions
* feat(ui): dynamic prompts is always enabled
* feat(queue): add queue_status_changed event
* feat(ui): wip queue graphs
* feat(nodes): move cleanup til after invoker startup
* feat(nodes): add cancel_by_batch_ids
* feat(ui): wip batch graphs & UI
* fix(nodes): remove `Batch.batch_id` from required
* fix(ui): cleanup and use fixedCacheKey for all mutations
* fix(ui): remove orphaned nodes from canvas graphs
* fix(nodes): fix cancel_by_batch_ids result count
* fix(ui): only show cancel batch tooltip when batches were canceled
* chore: isort
* fix(api): return `[""]` when dynamic prompts generates no prompts
Just a simple fallback so we always have a prompt.
* feat(ui): dynamicPrompts.combinatorial is always on
There seems to be little purpose in using the combinatorial generation for dynamic prompts. I've disabled it by hiding it from the UI and defaulting combinatorial to true. If we want to enable it again in the future it's straightforward to do so.
* feat: add queue_id & support logic
* feat(ui): fix upscale button
It prepends the upscale operation to queue
* feat(nodes): return queue item when enqueuing a single graph
This facilitates one-off graph async workflows in the client.
* feat(ui): move controlnet autoprocess to queue
* fix(ui): fix non-serializable DOMRect in redux state
* feat(ui): QueueTable performance tweaks
* feat(ui): update queue list
Queue items expand to show the full queue item. Just as JSON for now.
* wip threaded session_processor
* feat(nodes,ui): fully migrate queue to session_processor
* feat(nodes,ui): add processor events
* feat(ui): ui tweaks
* feat(nodes,ui): consolidate events, reduce network requests
* feat(ui): cleanup & abstract queue hooks
* feat(nodes): optimize batch permutation
Use a generator to do only as much work as is needed.
Previously, though we only ended up creating exactly as many queue items as was needed, there was still some intermediary work that calculated *all* permutations. When that number was very high, the system had a very hard time and used a lot of memory.
The logic has been refactored to use a generator. Additionally, the batch validators are optimized to return early and use less memory.
* feat(ui): add seed behaviour parameter
This dynamic prompts parameter allows the seed to be randomized per prompt or per iteration:
- Per iteration: Use the same seed for all prompts in a single dynamic prompt expansion
- Per prompt: Use a different seed for every single prompt
"Per iteration" is appropriate for exploring a the latents space with a stable starting noise, while "Per prompt" provides more variation.
* fix(ui): remove extraneous random seed nodes from linear graphs
* fix(ui): fix controlnet autoprocess not working when queue is running
* feat(queue): add timestamps to queue status updates
Also show execution time in queue list
* feat(queue): change all execution-related events to use the `queue_id` as the room, also include `queue_item_id` in InvocationQueueItem
This allows for much simpler handling of queue items.
* feat(api): deprecate sessions router
* chore(backend): tidy logging in `dependencies.py`
* fix(backend): respect `use_memory_db`
* feat(backend): add `config.log_sql` (enables sql trace logging)
* feat: add invocation cache
Supersedes #4574
The invocation cache provides simple node memoization functionality. Nodes that use the cache are memoized and not re-executed if their inputs haven't changed. Instead, the stored output is returned.
## Results
This feature provides anywhere some significant to massive performance improvement.
The improvement is most marked on large batches of generations where you only change a couple things (e.g. different seed or prompt for each iteration) and low-VRAM systems, where skipping an extraneous model load is a big deal.
## Overview
A new `invocation_cache` service is added to handle the caching. There's not much to it.
All nodes now inherit a boolean `use_cache` field from `BaseInvocation`. This is a node field and not a class attribute, because specific instances of nodes may want to opt in or out of caching.
The recently-added `invoke_internal()` method on `BaseInvocation` is used as an entrypoint for the cache logic.
To create a cache key, the invocation is first serialized using pydantic's provided `json()` method, skipping the unique `id` field. Then python's very fast builtin `hash()` is used to create an integer key. All implementations of `InvocationCacheBase` must provide a class method `create_key()` which accepts an invocation and outputs a string or integer key.
## In-Memory Implementation
An in-memory implementation is provided. In this implementation, the node outputs are stored in memory as python classes. The in-memory cache does not persist application restarts.
Max node cache size is added as `node_cache_size` under the `Generation` config category.
It defaults to 512 - this number is up for discussion, but given that these are relatively lightweight pydantic models, I think it's safe to up this even higher.
Note that the cache isn't storing the big stuff - tensors and images are store on disk, and outputs include only references to them.
## Node Definition
The default for all nodes is to use the cache. The `@invocation` decorator now accepts an optional `use_cache: bool` argument to override the default of `True`.
Non-deterministic nodes, however, should set this to `False`. Currently, all random-stuff nodes, including `dynamic_prompt`, are set to `False`.
The field name `use_cache` is now effectively a reserved field name and possibly a breaking change if any community nodes use this as a field name. In hindsight, all our reserved field names should have been prefixed with underscores or something.
## One Gotcha
Leaf nodes probably want to opt out of the cache, because if they are not cached, their outputs are not saved again.
If you run the same graph multiple times, you only end up with a single image output, because the image storage side-effects are in the `invoke()` method, which is bypassed if we have a cache hit.
## Linear UI
The linear graphs _almost_ just work, but due to the gotcha, we need to be careful about the final image-outputting node. To resolve this, a `SaveImageInvocation` node is added and used in the linear graphs.
This node is similar to `ImagePrimitive`, except it saves a copy of its input image, and has `use_cache` set to `False` by default.
This is now the leaf node in all linear graphs, and is the only node in those graphs with `use_cache == False` _and_ the only node with `is_intermedate == False`.
## Workflow Editor
All nodes now have a footer with a new `Use Cache [ ]` checkbox. It defaults to the value set by the invocation in its python definition, but can be changed by the user.
The workflow/node validation logic has been updated to migrate old workflows to use the new default values for `use_cache`. Users may still want to review the settings that have been chosen. In the event of catastrophic failure when running this migration, the default value of `True` is applied, as this is correct for most nodes.
Users should consider saving their workflows after loading them in and having them updated.
## Future Enhancements - Callback
A future enhancement would be to provide a callback to the `use_cache` flag that would be run as the node is executed to determine, based on its own internal state, if the cache should be used or not.
This would be useful for `DynamicPromptInvocation`, where the deterministic behaviour is determined by the `combinatorial: bool` field.
## Future Enhancements - Persisted Cache
Similar to how the latents storage is backed by disk, the invocation cache could be persisted to the database or disk. We'd need to be very careful about deserializing outputs, but it's perhaps worth exploring in the future.
* fix(ui): fix queue list item width
* feat(nodes): do not send the whole node on every generator progress
* feat(ui): strip out old logic related to sessions
Things like `isProcessing` are no longer relevant with queue. Removed them all & updated everything be appropriate for queue. May be a few little quirks I've missed...
* feat(ui): fix up param collapse labels
* feat(ui): click queue count to go to queue tab
* tidy(queue): update comment, query format
* feat(ui): fix progress bar when canceling
* fix(ui): fix circular dependency
* feat(nodes): bail on node caching logic if `node_cache_size == 0`
* feat(nodes): handle KeyError on node cache pop
* feat(nodes): bypass cache codepath if caches is disabled
more better no do thing
* fix(ui): reset api cache on connect/disconnect
* feat(ui): prevent enqueue when no prompts generated
* feat(ui): add queue controls to workflow editor
* feat(ui): update floating buttons & other incidental UI tweaks
* fix(ui): fix missing/incorrect translation keys
* fix(tests): add config service to mock invocation services
invoking needs access to `node_cache_size` to occur
* optionally remove pause/resume buttons from queue UI
* option to disable prepending
* chore(ui): remove unused file
* feat(queue): remove `order_id` entirely, `item_id` is now an autoinc pk
---------
Co-authored-by: Mary Hipp <maryhipp@Marys-MacBook-Air.local>
2023-09-20 05:09:24 +00:00
"promptsWithCount_one" : "{{count}} Prompt" ,
"promptsWithCount_other" : "{{count}} Prompts" ,
"seedBehaviour" : {
"label" : "Seed Behaviour" ,
"perIterationLabel" : "Seed per Iteration" ,
"perIterationDesc" : "Use a different seed for each iteration" ,
feat(ui): refactor informational popover
- Change translations to use arrays of paragraphs instead of a single paragraph.
- Change component to accept a `feature` prop to identify the feature which the popover describes.
- Add optional `wrapperProps`: passed to the wrapper element, allowing more flexibility when using the popover
- Add optional `popoverProps`: passed to the `<Popover />` component, allowing for overriding individual instances of the popover's props
- Move definitions of features and popover settings to `invokeai/frontend/web/src/common/components/IAIInformationalPopover/constants.ts`
- Add some type safety to the `feature` prop
- Edit `POPOVER_DATA` to provide `image`, `href`, `buttonLabel`, and any popover props. The popover props are applied to all instances of the popover for the given feature. Note that the component prop `popoverProps` will override settings here.
- Remove the popover's arrow. Because the popover is wrapping groups of components, sometimes the error ends up pointing to nothing, which looks kinda janky. I've just removed the arrow entirely, but feel free to add it back if you think it looks better.
- Use a `link` variant button with external link icon to better communicate that clicking the button will open a new tab.
- Default the link button label to "Learn More" (if a label is provided, that will be used instead)
- Make default position `top`, but set manually set some to `right` - namely, anything with a dropdown. This prevents the popovers from obscuring or being obscured by the dropdowns.
- Do a bit more restructuring of the Popover component itself, and how it is integrated with other components
- More ref forwarding
- Make the open delay 1s
- Set the popovers to use lazy mounting (eg do not mount until the user opens the thing)
- Update the verbiage for many popover items and add missing dynamic prompts stuff
2023-09-22 11:04:35 +00:00
"perPromptLabel" : "Seed per Image" ,
"perPromptDesc" : "Use a different seed for each image"
2024-01-02 09:43:25 +00:00
} ,
"loading" : "Generating Dynamic Prompts..."
2023-09-13 07:31:34 +00:00
} ,
"sdxl" : {
"cfgScale" : "CFG Scale" ,
2024-01-02 23:44:24 +00:00
"concatPromptStyle" : "Linking Prompt & Style" ,
2023-12-28 13:03:21 +00:00
"freePromptStyle" : "Manual Style Prompting" ,
2023-09-13 07:31:34 +00:00
"denoisingStrength" : "Denoising Strength" ,
"loading" : "Loading..." ,
"negAestheticScore" : "Negative Aesthetic Score" ,
"negStylePrompt" : "Negative Style Prompt" ,
"noModelsAvailable" : "No models available" ,
"posAestheticScore" : "Positive Aesthetic Score" ,
"posStylePrompt" : "Positive Style Prompt" ,
"refiner" : "Refiner" ,
"refinermodel" : "Refiner Model" ,
"refinerStart" : "Refiner Start" ,
2024-03-19 20:35:03 +00:00
"refinerSteps" : "Refiner Steps" ,
2023-09-13 07:31:34 +00:00
"scheduler" : "Scheduler" ,
2024-03-13 11:37:47 +00:00
"steps" : "Steps"
2023-02-18 04:23:24 +00:00
} ,
"settings" : {
2023-05-10 04:38:05 +00:00
"antialiasProgressImages" : "Antialias Progress Images" ,
2023-09-13 07:31:34 +00:00
"beta" : "Beta" ,
"confirmOnDelete" : "Confirm On Delete" ,
2023-04-29 08:25:02 +00:00
"developer" : "Developer" ,
2023-09-13 07:31:34 +00:00
"displayInProgress" : "Display Progress Images" ,
"enableImageDebugging" : "Enable Image Debugging" ,
2023-11-20 06:24:03 +00:00
"enableInformationalPopovers" : "Enable Informational Popovers" ,
2024-07-25 05:33:39 +00:00
"informationalPopoversDisabled" : "Informational Popovers Disabled" ,
"informationalPopoversDisabledDesc" : "Informational popovers have been disabled. Enable them in Settings." ,
2023-11-20 06:24:03 +00:00
"enableInvisibleWatermark" : "Enable Invisible Watermark" ,
"enableNSFWChecker" : "Enable NSFW Checker" ,
2023-09-13 07:31:34 +00:00
"general" : "General" ,
"generation" : "Generation" ,
"models" : "Models" ,
"resetComplete" : "Web UI has been reset." ,
"resetWebUI" : "Reset Web UI" ,
"resetWebUIDesc1" : "Resetting the web UI only resets the browser's local cache of your images and remembered settings. It does not delete any images from disk." ,
"resetWebUIDesc2" : "If images aren't showing up in the gallery or something else isn't working, please try resetting before submitting an issue on GitHub." ,
"shouldLogToConsole" : "Console Logging" ,
"showProgressInViewer" : "Show Progress Images in Viewer" ,
"ui" : "User Interface" ,
2023-10-15 06:27:25 +00:00
"clearIntermediatesDisabled" : "Queue must be empty to clear intermediates" ,
2023-10-12 07:35:33 +00:00
"clearIntermediatesDesc1" : "Clearing intermediates will reset your Canvas and ControlNet state." ,
"clearIntermediatesDesc2" : "Intermediate images are byproducts of generation, different from the result images in the gallery. Clearing intermediates will free disk space." ,
"clearIntermediatesDesc3" : "Your gallery images will not be deleted." ,
"clearIntermediates" : "Clear Intermediates" ,
2023-10-12 10:34:24 +00:00
"clearIntermediatesWithCount_one" : "Clear {{count}} Intermediate" ,
"clearIntermediatesWithCount_other" : "Clear {{count}} Intermediates" ,
"intermediatesCleared_one" : "Cleared {{count}} Intermediate" ,
"intermediatesCleared_other" : "Cleared {{count}} Intermediates" ,
2023-11-25 03:46:19 +00:00
"intermediatesClearedFailed" : "Problem Clearing Intermediates" ,
"reloadingIn" : "Reloading in"
2023-02-18 04:23:24 +00:00
} ,
"toast" : {
2023-09-15 03:43:32 +00:00
"addedToBoard" : "Added to board" ,
2024-05-21 09:33:44 +00:00
"baseModelChanged" : "Base Model Changed" ,
"baseModelChangedCleared_one" : "Cleared or disabled {{count}} incompatible submodel" ,
"baseModelChangedCleared_other" : "Cleared or disabled {{count}} incompatible submodels" ,
Partial migration of UI to nodes API (#3195)
* feat(ui): add axios client generator and simple example
* fix(ui): update client & nodes test code w/ new Edge type
* chore(ui): organize generated files
* chore(ui): update .eslintignore, .prettierignore
* chore(ui): update openapi.json
* feat(backend): fixes for nodes/generator
* feat(ui): generate object args for api client
* feat(ui): more nodes api prototyping
* feat(ui): nodes cancel
* chore(ui): regenerate api client
* fix(ui): disable OG web server socket connection
* fix(ui): fix scrollbar styles typing and prop
just noticed the typo, and made the types stronger.
* feat(ui): add socketio types
* feat(ui): wip nodes
- extract api client method arg types instead of manually declaring them
- update example to display images
- general tidy up
* start building out node translations from frontend state and add notes about missing features
* use reference to sampler_name
* use reference to sampler_name
* add optional apiUrl prop
* feat(ui): start hooking up dynamic txt2img node generation, create middleware for session invocation
* feat(ui): write separate nodes socket layer, txt2img generating and rendering w single node
* feat(ui): img2img implementation
* feat(ui): get intermediate images working but types are stubbed out
* chore(ui): add support for package mode
* feat(ui): add nodes mode script
* feat(ui): handle random seeds
* fix(ui): fix middleware types
* feat(ui): add rtk action type guard
* feat(ui): disable NodeAPITest
This was polluting the network/socket logs.
* feat(ui): fix parameters panel border color
This commit should be elsewhere but I don't want to break my flow
* feat(ui): make thunk types more consistent
* feat(ui): add type guards for outputs
* feat(ui): load images on socket connect
Rudimentary
* chore(ui): bump redux-toolkit
* docs(ui): update readme
* chore(ui): regenerate api client
* chore(ui): add typescript as dev dependency
I am having trouble with TS versions after vscode updated and now uses TS 5. `madge` has installed 3.9.10 and for whatever reason my vscode wants to use that. Manually specifying 4.9.5 and then setting vscode to use that as the workspace TS fixes the issue.
* feat(ui): begin migrating gallery to nodes
Along the way, migrate to use RTK `createEntityAdapter` for gallery images, and separate `results` and `uploads` into separate slices. Much cleaner this way.
* feat(ui): clean up & comment results slice
* fix(ui): separate thunk for initial gallery load so it properly gets index 0
* feat(ui): POST upload working
* fix(ui): restore removed type
* feat(ui): patch api generation for headers access
* chore(ui): regenerate api
* feat(ui): wip gallery migration
* feat(ui): wip gallery migration
* chore(ui): regenerate api
* feat(ui): wip refactor socket events
* feat(ui): disable panels based on app props
* feat(ui): invert logic to be disabled
* disable panels when app mounts
* feat(ui): add support to disableTabs
* docs(ui): organise and update docs
* lang(ui): add toast strings
* feat(ui): wip events, comments, and general refactoring
* feat(ui): add optional token for auth
* feat(ui): export StatusIndicator and ModelSelect for header use
* feat(ui) working on making socket URL dynamic
* feat(ui): dynamic middleware loading
* feat(ui): prep for socket jwt
* feat(ui): migrate cancelation
also updated action names to be event-like instead of declaration-like
sorry, i was scattered and this commit has a lot of unrelated stuff in it.
* fix(ui): fix img2img type
* chore(ui): regenerate api client
* feat(ui): improve InvocationCompleteEvent types
* feat(ui): increase StatusIndicator font size
* fix(ui): fix middleware order for multi-node graphs
* feat(ui): add exampleGraphs object w/ iterations example
* feat(ui): generate iterations graph
* feat(ui): update ModelSelect for nodes API
* feat(ui): add hi-res functionality for txt2img generations
* feat(ui): "subscribe" to particular nodes
feels like a dirty hack but oh well it works
* feat(ui): first steps to node editor ui
* fix(ui): disable event subscription
it is not fully baked just yet
* feat(ui): wip node editor
* feat(ui): remove extraneous field types
* feat(ui): nodes before deleting stuff
* feat(ui): cleanup nodes ui stuff
* feat(ui): hook up nodes to redux
* fix(ui): fix handle
* fix(ui): add basic node edges & connection validation
* feat(ui): add connection validation styling
* feat(ui): increase edge width
* feat(ui): it blends
* feat(ui): wip model handling and graph topology validation
* feat(ui): validation connections w/ graphlib
* docs(ui): update nodes doc
* feat(ui): wip node editor
* chore(ui): rebuild api, update types
* add redux-dynamic-middlewares as a dependency
* feat(ui): add url host transformation
* feat(ui): handle already-connected fields
* feat(ui): rewrite SqliteItemStore in sqlalchemy
* fix(ui): fix sqlalchemy dynamic model instantiation
* feat(ui, nodes): metadata wip
* feat(ui, nodes): models
* feat(ui, nodes): more metadata wip
* feat(ui): wip range/iterate
* fix(nodes): fix sqlite typing
* feat(ui): export new type for invoke component
* tests(nodes): fix test instantiation of ImageField
* feat(nodes): fix LoadImageInvocation
* feat(nodes): add `title` ui hint
* feat(nodes): make ImageField attrs optional
* feat(ui): wip nodes etc
* feat(nodes): roll back sqlalchemy
* fix(nodes): partially address feedback
* fix(backend): roll back changes to pngwriter
* feat(nodes): wip address metadata feedback
* feat(nodes): add seeded rng to RandomRange
* feat(nodes): address feedback
* feat(nodes): move GET images error handling to DiskImageStorage
* feat(nodes): move GET images error handling to DiskImageStorage
* fix(nodes): fix image output schema customization
* feat(ui): img2img/txt2img -> linear
- remove txt2img and img2img tabs
- add linear tab
- add initial image selection to linear parameters accordion
* feat(ui): tidy graph builders
* feat(ui): tidy misc
* feat(ui): improve invocation union types
* feat(ui): wip metadata viewer recall
* feat(ui): move fonts to normal deps
* feat(nodes): fix broken upload
* feat(nodes): add metadata module + tests, thumbnails
- `MetadataModule` is stateless and needed in places where the `InvocationContext` is not available, so have not made it a `service`
- Handles loading/parsing/building metadata, and creating png info objects
- added tests for MetadataModule
- Lifted thumbnail stuff to util
* fix(nodes): revert change to RandomRangeInvocation
* feat(nodes): address feedback
- make metadata a service
- rip out pydantic validation, implement metadata parsing as simple functions
- update tests
- address other minor feedback items
* fix(nodes): fix other tests
* fix(nodes): add metadata service to cli
* fix(nodes): fix latents/image field parsing
* feat(nodes): customise LatentsField schema
* feat(nodes): move metadata parsing to frontend
* fix(nodes): fix metadata test
---------
Co-authored-by: maryhipp <maryhipp@gmail.com>
Co-authored-by: Mary Hipp <maryhipp@Marys-MacBook-Air.local>
2023-04-22 03:10:20 +00:00
"canceled" : "Processing Canceled" ,
2023-09-15 03:43:32 +00:00
"canvasCopiedClipboard" : "Canvas Copied to Clipboard" ,
"canvasDownloaded" : "Canvas Downloaded" ,
2023-09-13 07:31:34 +00:00
"canvasMerged" : "Canvas Merged" ,
2023-09-15 03:43:32 +00:00
"canvasSavedGallery" : "Canvas Saved to Gallery" ,
"canvasSentControlnetAssets" : "Canvas Sent to ControlNet & Assets" ,
2023-09-13 07:31:34 +00:00
"connected" : "Connected to Server" ,
2023-02-18 04:23:24 +00:00
"imageCopied" : "Image Copied" ,
2023-05-03 09:27:29 +00:00
"imageNotLoadedDesc" : "Could not find image" ,
2023-09-15 03:43:32 +00:00
"imageSaved" : "Image Saved" ,
"imageSavingFailed" : "Image Saving Failed" ,
"imageUploaded" : "Image Uploaded" ,
"imageUploadFailed" : "Image Upload Failed" ,
2024-08-14 15:09:53 +00:00
"importFailed" : "Import Failed" ,
"importSuccessful" : "Import Successful" ,
2023-11-25 03:46:19 +00:00
"invalidUpload" : "Invalid Upload" ,
2023-09-15 03:43:32 +00:00
"loadedWithWarnings" : "Workflow Loaded with Warnings" ,
"maskSavedAssets" : "Mask Saved to Assets" ,
"maskSentControlnetAssets" : "Mask Sent to ControlNet & Assets" ,
2023-09-13 07:31:34 +00:00
"metadataLoadFailed" : "Failed to load metadata" ,
2024-03-11 20:30:32 +00:00
"modelAddedSimple" : "Model Added to Queue" ,
2024-02-21 22:17:34 +00:00
"modelImportCanceled" : "Model Import Canceled" ,
2024-05-21 09:33:44 +00:00
"outOfMemoryError" : "Out of Memory Error" ,
"outOfMemoryErrorDesc" : "Your current generation settings exceed system capacity. Please adjust your settings and try again." ,
2024-04-05 17:30:31 +00:00
"parameters" : "Parameters" ,
2024-05-21 09:33:44 +00:00
"parameterSet" : "Parameter Recalled" ,
"parameterSetDesc" : "Recalled {{parameter}}" ,
2024-05-25 10:32:07 +00:00
"parameterNotSet" : "Parameter Not Recalled" ,
2024-05-21 09:33:44 +00:00
"parameterNotSetDesc" : "Unable to recall {{parameter}}" ,
"parameterNotSetDescWithMessage" : "Unable to recall {{parameter}}: {{message}}" ,
"parametersSet" : "Parameters Recalled" ,
"parametersNotSet" : "Parameters Not Recalled" ,
"errorCopied" : "Error Copied" ,
2023-09-15 03:43:32 +00:00
"problemCopyingCanvas" : "Problem Copying Canvas" ,
"problemCopyingCanvasDesc" : "Unable to export base layer" ,
2023-09-13 07:31:34 +00:00
"problemCopyingImage" : "Unable to Copy Image" ,
2024-02-01 20:25:01 +00:00
"problemDownloadingImage" : "Unable to Download Image" ,
2023-09-15 03:43:32 +00:00
"problemDownloadingCanvas" : "Problem Downloading Canvas" ,
"problemDownloadingCanvasDesc" : "Unable to export base layer" ,
"problemImportingMask" : "Problem Importing Mask" ,
"problemImportingMaskDesc" : "Unable to export mask" ,
"problemMergingCanvas" : "Problem Merging Canvas" ,
"problemMergingCanvasDesc" : "Unable to export base layer" ,
"problemSavingCanvas" : "Problem Saving Canvas" ,
"problemSavingCanvasDesc" : "Unable to export base layer" ,
"problemSavingMask" : "Problem Saving Mask" ,
"problemSavingMaskDesc" : "Unable to export mask" ,
2024-02-21 22:17:34 +00:00
"prunedQueue" : "Pruned Queue" ,
2024-01-21 12:14:51 +00:00
"resetInitialImage" : "Reset Initial Image" ,
2023-09-13 07:31:34 +00:00
"sentToImageToImage" : "Sent To Image To Image" ,
"sentToUnifiedCanvas" : "Sent to Unified Canvas" ,
"serverError" : "Server Error" ,
2024-05-21 09:33:44 +00:00
"sessionRef" : "Session: {{sessionId}}" ,
2023-11-20 06:24:03 +00:00
"setAsCanvasInitialImage" : "Set as canvas initial image" ,
"setCanvasInitialImage" : "Set canvas initial image" ,
2023-09-15 03:43:32 +00:00
"setControlImage" : "Set as control image" ,
"setInitialImage" : "Set as initial image" ,
"setNodeField" : "Set as node field" ,
2024-05-21 09:33:44 +00:00
"somethingWentWrong" : "Something Went Wrong" ,
2023-09-13 07:31:34 +00:00
"uploadFailed" : "Upload failed" ,
"uploadFailedInvalidUploadDesc" : "Must be single PNG or JPEG image" ,
2024-01-21 12:14:51 +00:00
"uploadInitialImage" : "Upload Initial Image" ,
feat: workflow library (#5148)
* chore: bump pydantic to 2.5.2
This release fixes pydantic/pydantic#8175 and allows us to use `JsonValue`
* fix(ui): exclude public/en.json from prettier config
* fix(workflow_records): fix SQLite workflow insertion to ignore duplicates
* feat(backend): update workflows handling
Update workflows handling for Workflow Library.
**Updated Workflow Storage**
"Embedded Workflows" are workflows associated with images, and are now only stored in the image files. "Library Workflows" are not associated with images, and are stored only in DB.
This works out nicely. We have always saved workflows to files, but recently began saving them to the DB in addition to in image files. When that happened, we stopped reading workflows from files, so all the workflows that only existed in images were inaccessible. With this change, access to those workflows is restored, and no workflows are lost.
**Updated Workflow Handling in Nodes**
Prior to this change, workflows were embedded in images by passing the whole workflow JSON to a special workflow field on a node. In the node's `invoke()` function, the node was able to access this workflow and save it with the image. This (inaccurately) models workflows as a property of an image and is rather awkward technically.
A workflow is now a property of a batch/session queue item. It is available in the InvocationContext and therefore available to all nodes during `invoke()`.
**Database Migrations**
Added a `SQLiteMigrator` class to handle database migrations. Migrations were needed to accomodate the DB-related changes in this PR. See the code for details.
The `images`, `workflows` and `session_queue` tables required migrations for this PR, and are using the new migrator. Other tables/services are still creating tables themselves. A followup PR will adapt them to use the migrator.
**Other/Support Changes**
- Add a `has_workflow` column to `images` table to indicate that the image has an embedded workflow.
- Add handling for retrieving the workflow from an image in python. The image file must be fetched, the workflow extracted, and then sent to client, avoiding needing the browser to parse the image file. With the `has_workflow` column, the UI knows if there is a workflow to be fetched, and only fetches when the user requests to load the workflow.
- Add route to get the workflow from an image
- Add CRUD service/routes for the library workflows
- `workflow_images` table and services removed (no longer needed now that embedded workflows are not in the DB)
* feat(ui): updated workflow handling (WIP)
Clientside updates for the backend workflow changes.
Includes roughed-out workflow library UI.
* feat: revert SQLiteMigrator class
Will pursue this in a separate PR.
* feat(nodes): do not overwrite custom node module names
Use a different, simpler method to detect if a node is custom.
* feat(nodes): restore WithWorkflow as no-op class
This class is deprecated and no longer needed. Set its workflow attr value to None (meaning it is now a no-op), and issue a warning when an invocation subclasses it.
* fix(nodes): fix get_workflow from queue item dict func
* feat(backend): add WorkflowRecordListItemDTO
This is the id, name, description, created at and updated at workflow columns/attrs. Used to display lists of workflowsl
* chore(ui): typegen
* feat(ui): add workflow loading, deleting to workflow library UI
* feat(ui): workflow library pagination button styles
* wip
* feat: workflow library WIP
- Save to library
- Duplicate
- Filter/sort
- UI/queries
* feat: workflow library - system graphs - wip
* feat(backend): sync system workflows to db
* fix: merge conflicts
* feat: simplify default workflows
- Rename "system" -> "default"
- Simplify syncing logic
- Update UI to match
* feat(workflows): update default workflows
- Update TextToImage_SD15
- Add TextToImage_SDXL
- Add README
* feat(ui): refine workflow list UI
* fix(workflow_records): typo
* fix(tests): fix tests
* feat(ui): clean up workflow library hooks
* fix(db): fix mis-ordered db cleanup step
It was happening before pruning queue items - should happen afterwards, else you have to restart the app again to free disk space made available by the pruning.
* feat(ui): tweak reset workflow editor translations
* feat(ui): split out workflow redux state
The `nodes` slice is a rather complicated slice. Removing `workflow` makes it a bit more reasonable.
Also helps to flatten state out a bit.
* docs: update default workflows README
* fix: tidy up unused files, unrelated changes
* fix(backend): revert unrelated service organisational changes
* feat(backend): workflow_records.get_many arg "filter_text" -> "query"
* feat(ui): use custom hook in current image buttons
Already in use elsewhere, forgot to use it here.
* fix(ui): remove commented out property
* fix(ui): fix workflow loading
- Different handling for loading from library vs external
- Fix bug where only nodes and edges loaded
* fix(ui): fix save/save-as workflow naming
* fix(ui): fix circular dependency
* fix(db): fix bug with releasing without lock in db.clean()
* fix(db): remove extraneous lock
* chore: bump ruff
* fix(workflow_records): default `category` to `WorkflowCategory.User`
This allows old workflows to validate when reading them from the db or image files.
* hide workflow library buttons if feature is disabled
---------
Co-authored-by: Mary Hipp <maryhipp@Marys-MacBook-Air.local>
2023-12-08 22:48:38 +00:00
"workflowLoaded" : "Workflow Loaded" ,
"problemRetrievingWorkflow" : "Problem Retrieving Workflow" ,
"workflowDeleted" : "Workflow Deleted" ,
"problemDeletingWorkflow" : "Problem Deleting Workflow"
2023-02-18 04:23:24 +00:00
} ,
"tooltip" : {
"feature" : {
2023-09-13 07:31:34 +00:00
"boundingBox" : "The bounding box is the same as the Width and Height settings for Text to Image or Image to Image. Only the area in the box will be processed." ,
2023-02-18 04:23:24 +00:00
"gallery" : "Gallery displays generations from the outputs folder as they're created. Settings are stored within files and accesed by context menu." ,
2023-02-18 21:35:33 +00:00
"other" : "These options will enable alternative processing modes for Invoke. 'Seamless tiling' will create repeating patterns in the output. 'High resolution' is generation in two steps with img2img: use this setting when you want a larger and more coherent image without artifacts. It will take longer than usual txt2img." ,
2023-09-13 07:31:34 +00:00
"prompt" : "This is the prompt field. Prompt includes generation objects and stylistic terms. You can add weight (token importance) in the prompt as well, but CLI commands and parameters will not work." ,
2023-02-18 04:23:24 +00:00
"seed" : "Seed value affects the initial noise from which the image is formed. You can use the already existing seeds from previous images. 'Noise Threshold' is used to mitigate artifacts at high CFG values (try the 0-10 range), and Perlin to add Perlin noise during generation: both serve to add variation to your outputs." ,
2024-03-13 11:37:47 +00:00
"upscale" : "Use ESRGAN to enlarge the image immediately after generation."
2023-02-18 04:23:24 +00:00
}
} ,
2023-09-15 14:37:05 +00:00
"popovers" : {
"clipSkip" : {
"heading" : "CLIP Skip" ,
feat(ui): refactor informational popover
- Change translations to use arrays of paragraphs instead of a single paragraph.
- Change component to accept a `feature` prop to identify the feature which the popover describes.
- Add optional `wrapperProps`: passed to the wrapper element, allowing more flexibility when using the popover
- Add optional `popoverProps`: passed to the `<Popover />` component, allowing for overriding individual instances of the popover's props
- Move definitions of features and popover settings to `invokeai/frontend/web/src/common/components/IAIInformationalPopover/constants.ts`
- Add some type safety to the `feature` prop
- Edit `POPOVER_DATA` to provide `image`, `href`, `buttonLabel`, and any popover props. The popover props are applied to all instances of the popover for the given feature. Note that the component prop `popoverProps` will override settings here.
- Remove the popover's arrow. Because the popover is wrapping groups of components, sometimes the error ends up pointing to nothing, which looks kinda janky. I've just removed the arrow entirely, but feel free to add it back if you think it looks better.
- Use a `link` variant button with external link icon to better communicate that clicking the button will open a new tab.
- Default the link button label to "Learn More" (if a label is provided, that will be used instead)
- Make default position `top`, but set manually set some to `right` - namely, anything with a dropdown. This prevents the popovers from obscuring or being obscured by the dropdowns.
- Do a bit more restructuring of the Popover component itself, and how it is integrated with other components
- More ref forwarding
- Make the open delay 1s
- Set the popovers to use lazy mounting (eg do not mount until the user opens the thing)
- Update the verbiage for many popover items and add missing dynamic prompts stuff
2023-09-22 11:04:35 +00:00
"paragraphs" : [
2024-02-19 17:50:35 +00:00
"How many layers of the CLIP model to skip." ,
"Certain models are better suited to be used with CLIP Skip."
feat(ui): refactor informational popover
- Change translations to use arrays of paragraphs instead of a single paragraph.
- Change component to accept a `feature` prop to identify the feature which the popover describes.
- Add optional `wrapperProps`: passed to the wrapper element, allowing more flexibility when using the popover
- Add optional `popoverProps`: passed to the `<Popover />` component, allowing for overriding individual instances of the popover's props
- Move definitions of features and popover settings to `invokeai/frontend/web/src/common/components/IAIInformationalPopover/constants.ts`
- Add some type safety to the `feature` prop
- Edit `POPOVER_DATA` to provide `image`, `href`, `buttonLabel`, and any popover props. The popover props are applied to all instances of the popover for the given feature. Note that the component prop `popoverProps` will override settings here.
- Remove the popover's arrow. Because the popover is wrapping groups of components, sometimes the error ends up pointing to nothing, which looks kinda janky. I've just removed the arrow entirely, but feel free to add it back if you think it looks better.
- Use a `link` variant button with external link icon to better communicate that clicking the button will open a new tab.
- Default the link button label to "Learn More" (if a label is provided, that will be used instead)
- Make default position `top`, but set manually set some to `right` - namely, anything with a dropdown. This prevents the popovers from obscuring or being obscured by the dropdowns.
- Do a bit more restructuring of the Popover component itself, and how it is integrated with other components
- More ref forwarding
- Make the open delay 1s
- Set the popovers to use lazy mounting (eg do not mount until the user opens the thing)
- Update the verbiage for many popover items and add missing dynamic prompts stuff
2023-09-22 11:04:35 +00:00
]
} ,
"paramNegativeConditioning" : {
"heading" : "Negative Prompt" ,
"paragraphs" : [
"The generation process avoids the concepts in the negative prompt. Use this to exclude qualities or objects from the output." ,
"Supports Compel syntax and embeddings."
]
} ,
"paramPositiveConditioning" : {
"heading" : "Positive Prompt" ,
"paragraphs" : [
"Guides the generation process. You may use any words or phrases." ,
"Compel and Dynamic Prompts syntaxes and embeddings."
]
} ,
"paramScheduler" : {
"heading" : "Scheduler" ,
"paragraphs" : [
2024-02-19 17:50:35 +00:00
"Scheduler used during the generation process." ,
"Each scheduler defines how to iteratively add noise to an image or how to update a sample based on a model's output."
feat(ui): refactor informational popover
- Change translations to use arrays of paragraphs instead of a single paragraph.
- Change component to accept a `feature` prop to identify the feature which the popover describes.
- Add optional `wrapperProps`: passed to the wrapper element, allowing more flexibility when using the popover
- Add optional `popoverProps`: passed to the `<Popover />` component, allowing for overriding individual instances of the popover's props
- Move definitions of features and popover settings to `invokeai/frontend/web/src/common/components/IAIInformationalPopover/constants.ts`
- Add some type safety to the `feature` prop
- Edit `POPOVER_DATA` to provide `image`, `href`, `buttonLabel`, and any popover props. The popover props are applied to all instances of the popover for the given feature. Note that the component prop `popoverProps` will override settings here.
- Remove the popover's arrow. Because the popover is wrapping groups of components, sometimes the error ends up pointing to nothing, which looks kinda janky. I've just removed the arrow entirely, but feel free to add it back if you think it looks better.
- Use a `link` variant button with external link icon to better communicate that clicking the button will open a new tab.
- Default the link button label to "Learn More" (if a label is provided, that will be used instead)
- Make default position `top`, but set manually set some to `right` - namely, anything with a dropdown. This prevents the popovers from obscuring or being obscured by the dropdowns.
- Do a bit more restructuring of the Popover component itself, and how it is integrated with other components
- More ref forwarding
- Make the open delay 1s
- Set the popovers to use lazy mounting (eg do not mount until the user opens the thing)
- Update the verbiage for many popover items and add missing dynamic prompts stuff
2023-09-22 11:04:35 +00:00
]
2023-09-15 14:37:05 +00:00
} ,
2024-02-26 13:35:11 +00:00
"compositingMaskBlur" : {
"heading" : "Mask Blur" ,
2023-11-26 09:44:39 +00:00
"paragraphs" : [ "The blur radius of the mask." ]
2023-09-15 14:37:05 +00:00
} ,
"compositingBlurMethod" : {
"heading" : "Blur Method" ,
2023-11-26 09:44:39 +00:00
"paragraphs" : [ "The method of blur applied to the masked area." ]
2023-09-15 14:37:05 +00:00
} ,
"compositingCoherencePass" : {
"heading" : "Coherence Pass" ,
2024-01-27 09:55:55 +00:00
"paragraphs" : [ "A second round of denoising helps to composite the Inpainted/Outpainted image." ]
2023-09-15 14:37:05 +00:00
} ,
2023-09-18 15:03:02 +00:00
"compositingCoherenceMode" : {
2023-09-15 14:37:05 +00:00
"heading" : "Mode" ,
2024-02-19 17:50:35 +00:00
"paragraphs" : [ "Method used to create a coherent image with the newly generated masked area." ]
2023-09-15 14:37:05 +00:00
} ,
2024-02-21 17:32:49 +00:00
"compositingCoherenceEdgeSize" : {
"heading" : "Edge Size" ,
"paragraphs" : [ "The edge size of the coherence pass." ]
2023-09-15 14:37:05 +00:00
} ,
2024-02-21 17:32:49 +00:00
"compositingCoherenceMinDenoise" : {
"heading" : "Minimum Denoise" ,
"paragraphs" : [
"Minimum denoise strength for the Coherence mode" ,
"The minimum denoise strength for the coherence region when inpainting or outpainting"
]
2023-09-15 14:37:05 +00:00
} ,
"compositingMaskAdjustments" : {
"heading" : "Mask Adjustments" ,
2023-11-26 09:44:39 +00:00
"paragraphs" : [ "Adjust the mask." ]
2023-09-15 14:37:05 +00:00
} ,
2024-02-19 17:50:35 +00:00
"controlNet" : {
"heading" : "ControlNet" ,
"paragraphs" : [
"ControlNets provide guidance to the generation process, helping create images with controlled composition, structure, or style, depending on the model selected."
]
} ,
2023-09-15 14:37:05 +00:00
"controlNetBeginEnd" : {
"heading" : "Begin / End Step Percentage" ,
feat(ui): refactor informational popover
- Change translations to use arrays of paragraphs instead of a single paragraph.
- Change component to accept a `feature` prop to identify the feature which the popover describes.
- Add optional `wrapperProps`: passed to the wrapper element, allowing more flexibility when using the popover
- Add optional `popoverProps`: passed to the `<Popover />` component, allowing for overriding individual instances of the popover's props
- Move definitions of features and popover settings to `invokeai/frontend/web/src/common/components/IAIInformationalPopover/constants.ts`
- Add some type safety to the `feature` prop
- Edit `POPOVER_DATA` to provide `image`, `href`, `buttonLabel`, and any popover props. The popover props are applied to all instances of the popover for the given feature. Note that the component prop `popoverProps` will override settings here.
- Remove the popover's arrow. Because the popover is wrapping groups of components, sometimes the error ends up pointing to nothing, which looks kinda janky. I've just removed the arrow entirely, but feel free to add it back if you think it looks better.
- Use a `link` variant button with external link icon to better communicate that clicking the button will open a new tab.
- Default the link button label to "Learn More" (if a label is provided, that will be used instead)
- Make default position `top`, but set manually set some to `right` - namely, anything with a dropdown. This prevents the popovers from obscuring or being obscured by the dropdowns.
- Do a bit more restructuring of the Popover component itself, and how it is integrated with other components
- More ref forwarding
- Make the open delay 1s
- Set the popovers to use lazy mounting (eg do not mount until the user opens the thing)
- Update the verbiage for many popover items and add missing dynamic prompts stuff
2023-09-22 11:04:35 +00:00
"paragraphs" : [
2024-02-19 17:50:35 +00:00
"The part of the of the denoising process that will have the Control Adapter applied." ,
"Generally, Control Adapters applied at the start of the process guide composition, and Control Adapters applied at the end guide details."
feat(ui): refactor informational popover
- Change translations to use arrays of paragraphs instead of a single paragraph.
- Change component to accept a `feature` prop to identify the feature which the popover describes.
- Add optional `wrapperProps`: passed to the wrapper element, allowing more flexibility when using the popover
- Add optional `popoverProps`: passed to the `<Popover />` component, allowing for overriding individual instances of the popover's props
- Move definitions of features and popover settings to `invokeai/frontend/web/src/common/components/IAIInformationalPopover/constants.ts`
- Add some type safety to the `feature` prop
- Edit `POPOVER_DATA` to provide `image`, `href`, `buttonLabel`, and any popover props. The popover props are applied to all instances of the popover for the given feature. Note that the component prop `popoverProps` will override settings here.
- Remove the popover's arrow. Because the popover is wrapping groups of components, sometimes the error ends up pointing to nothing, which looks kinda janky. I've just removed the arrow entirely, but feel free to add it back if you think it looks better.
- Use a `link` variant button with external link icon to better communicate that clicking the button will open a new tab.
- Default the link button label to "Learn More" (if a label is provided, that will be used instead)
- Make default position `top`, but set manually set some to `right` - namely, anything with a dropdown. This prevents the popovers from obscuring or being obscured by the dropdowns.
- Do a bit more restructuring of the Popover component itself, and how it is integrated with other components
- More ref forwarding
- Make the open delay 1s
- Set the popovers to use lazy mounting (eg do not mount until the user opens the thing)
- Update the verbiage for many popover items and add missing dynamic prompts stuff
2023-09-22 11:04:35 +00:00
]
2023-09-15 14:37:05 +00:00
} ,
"controlNetControlMode" : {
"heading" : "Control Mode" ,
2024-02-19 17:50:35 +00:00
"paragraphs" : [ "Lend more weight to either the prompt or ControlNet." ]
2023-09-15 14:37:05 +00:00
} ,
2024-02-19 17:50:35 +00:00
"controlNetProcessor" : {
"heading" : "Processor" ,
feat(ui): refactor informational popover
- Change translations to use arrays of paragraphs instead of a single paragraph.
- Change component to accept a `feature` prop to identify the feature which the popover describes.
- Add optional `wrapperProps`: passed to the wrapper element, allowing more flexibility when using the popover
- Add optional `popoverProps`: passed to the `<Popover />` component, allowing for overriding individual instances of the popover's props
- Move definitions of features and popover settings to `invokeai/frontend/web/src/common/components/IAIInformationalPopover/constants.ts`
- Add some type safety to the `feature` prop
- Edit `POPOVER_DATA` to provide `image`, `href`, `buttonLabel`, and any popover props. The popover props are applied to all instances of the popover for the given feature. Note that the component prop `popoverProps` will override settings here.
- Remove the popover's arrow. Because the popover is wrapping groups of components, sometimes the error ends up pointing to nothing, which looks kinda janky. I've just removed the arrow entirely, but feel free to add it back if you think it looks better.
- Use a `link` variant button with external link icon to better communicate that clicking the button will open a new tab.
- Default the link button label to "Learn More" (if a label is provided, that will be used instead)
- Make default position `top`, but set manually set some to `right` - namely, anything with a dropdown. This prevents the popovers from obscuring or being obscured by the dropdowns.
- Do a bit more restructuring of the Popover component itself, and how it is integrated with other components
- More ref forwarding
- Make the open delay 1s
- Set the popovers to use lazy mounting (eg do not mount until the user opens the thing)
- Update the verbiage for many popover items and add missing dynamic prompts stuff
2023-09-22 11:04:35 +00:00
"paragraphs" : [
2024-02-19 17:50:35 +00:00
"Method of processing the input image to guide the generation process. Different processors will providedifferent effects or styles in your generated images."
feat(ui): refactor informational popover
- Change translations to use arrays of paragraphs instead of a single paragraph.
- Change component to accept a `feature` prop to identify the feature which the popover describes.
- Add optional `wrapperProps`: passed to the wrapper element, allowing more flexibility when using the popover
- Add optional `popoverProps`: passed to the `<Popover />` component, allowing for overriding individual instances of the popover's props
- Move definitions of features and popover settings to `invokeai/frontend/web/src/common/components/IAIInformationalPopover/constants.ts`
- Add some type safety to the `feature` prop
- Edit `POPOVER_DATA` to provide `image`, `href`, `buttonLabel`, and any popover props. The popover props are applied to all instances of the popover for the given feature. Note that the component prop `popoverProps` will override settings here.
- Remove the popover's arrow. Because the popover is wrapping groups of components, sometimes the error ends up pointing to nothing, which looks kinda janky. I've just removed the arrow entirely, but feel free to add it back if you think it looks better.
- Use a `link` variant button with external link icon to better communicate that clicking the button will open a new tab.
- Default the link button label to "Learn More" (if a label is provided, that will be used instead)
- Make default position `top`, but set manually set some to `right` - namely, anything with a dropdown. This prevents the popovers from obscuring or being obscured by the dropdowns.
- Do a bit more restructuring of the Popover component itself, and how it is integrated with other components
- More ref forwarding
- Make the open delay 1s
- Set the popovers to use lazy mounting (eg do not mount until the user opens the thing)
- Update the verbiage for many popover items and add missing dynamic prompts stuff
2023-09-22 11:04:35 +00:00
]
2023-09-15 14:37:05 +00:00
} ,
2024-02-19 17:50:35 +00:00
"controlNetResizeMode" : {
"heading" : "Resize Mode" ,
"paragraphs" : [ "Method to fit Control Adapter's input image size to the output generation size." ]
} ,
2024-04-18 15:29:09 +00:00
"ipAdapterMethod" : {
"heading" : "Method" ,
"paragraphs" : [ "Method by which to apply the current IP Adapter." ]
} ,
2023-09-15 14:37:05 +00:00
"controlNetWeight" : {
"heading" : "Weight" ,
2024-02-19 17:50:35 +00:00
"paragraphs" : [
"Weight of the Control Adapter. Higher weight will lead to larger impacts on the final image."
]
feat(ui): refactor informational popover
- Change translations to use arrays of paragraphs instead of a single paragraph.
- Change component to accept a `feature` prop to identify the feature which the popover describes.
- Add optional `wrapperProps`: passed to the wrapper element, allowing more flexibility when using the popover
- Add optional `popoverProps`: passed to the `<Popover />` component, allowing for overriding individual instances of the popover's props
- Move definitions of features and popover settings to `invokeai/frontend/web/src/common/components/IAIInformationalPopover/constants.ts`
- Add some type safety to the `feature` prop
- Edit `POPOVER_DATA` to provide `image`, `href`, `buttonLabel`, and any popover props. The popover props are applied to all instances of the popover for the given feature. Note that the component prop `popoverProps` will override settings here.
- Remove the popover's arrow. Because the popover is wrapping groups of components, sometimes the error ends up pointing to nothing, which looks kinda janky. I've just removed the arrow entirely, but feel free to add it back if you think it looks better.
- Use a `link` variant button with external link icon to better communicate that clicking the button will open a new tab.
- Default the link button label to "Learn More" (if a label is provided, that will be used instead)
- Make default position `top`, but set manually set some to `right` - namely, anything with a dropdown. This prevents the popovers from obscuring or being obscured by the dropdowns.
- Do a bit more restructuring of the Popover component itself, and how it is integrated with other components
- More ref forwarding
- Make the open delay 1s
- Set the popovers to use lazy mounting (eg do not mount until the user opens the thing)
- Update the verbiage for many popover items and add missing dynamic prompts stuff
2023-09-22 11:04:35 +00:00
} ,
"dynamicPrompts" : {
"heading" : "Dynamic Prompts" ,
"paragraphs" : [
"Dynamic Prompts parses a single prompt into many." ,
"The basic syntax is \"a {red|green|blue} ball\". This will produce three prompts: \"a red ball\", \"a green ball\" and \"a blue ball\"." ,
"You can use the syntax as many times as you like in a single prompt, but be sure to keep the number of prompts generated in check with the Max Prompts setting."
]
} ,
"dynamicPromptsMaxPrompts" : {
"heading" : "Max Prompts" ,
2024-01-27 09:55:55 +00:00
"paragraphs" : [ "Limits the number of prompts that can be generated by Dynamic Prompts." ]
feat(ui): refactor informational popover
- Change translations to use arrays of paragraphs instead of a single paragraph.
- Change component to accept a `feature` prop to identify the feature which the popover describes.
- Add optional `wrapperProps`: passed to the wrapper element, allowing more flexibility when using the popover
- Add optional `popoverProps`: passed to the `<Popover />` component, allowing for overriding individual instances of the popover's props
- Move definitions of features and popover settings to `invokeai/frontend/web/src/common/components/IAIInformationalPopover/constants.ts`
- Add some type safety to the `feature` prop
- Edit `POPOVER_DATA` to provide `image`, `href`, `buttonLabel`, and any popover props. The popover props are applied to all instances of the popover for the given feature. Note that the component prop `popoverProps` will override settings here.
- Remove the popover's arrow. Because the popover is wrapping groups of components, sometimes the error ends up pointing to nothing, which looks kinda janky. I've just removed the arrow entirely, but feel free to add it back if you think it looks better.
- Use a `link` variant button with external link icon to better communicate that clicking the button will open a new tab.
- Default the link button label to "Learn More" (if a label is provided, that will be used instead)
- Make default position `top`, but set manually set some to `right` - namely, anything with a dropdown. This prevents the popovers from obscuring or being obscured by the dropdowns.
- Do a bit more restructuring of the Popover component itself, and how it is integrated with other components
- More ref forwarding
- Make the open delay 1s
- Set the popovers to use lazy mounting (eg do not mount until the user opens the thing)
- Update the verbiage for many popover items and add missing dynamic prompts stuff
2023-09-22 11:04:35 +00:00
} ,
"dynamicPromptsSeedBehaviour" : {
"heading" : "Seed Behaviour" ,
"paragraphs" : [
"Controls how the seed is used when generating prompts." ,
"Per Iteration will use a unique seed for each iteration. Use this to explore prompt variations on a single seed." ,
"For example, if you have 5 prompts, each image will use the same seed." ,
"Per Image will use a unique seed for each image. This provides more variation."
]
2023-09-15 14:37:05 +00:00
} ,
2024-02-19 17:50:35 +00:00
"imageFit" : {
"heading" : "Fit Initial Image to Output Size" ,
"paragraphs" : [
"Resizes the initial image to the width and height of the output image. Recommended to enable."
]
} ,
2023-09-15 14:37:05 +00:00
"infillMethod" : {
"heading" : "Infill Method" ,
2024-02-19 17:50:35 +00:00
"paragraphs" : [ "Method of infilling during the Outpainting or Inpainting process." ]
2023-09-15 14:37:05 +00:00
} ,
"lora" : {
2024-02-19 17:50:35 +00:00
"heading" : "LoRA" ,
"paragraphs" : [ "Lightweight models that are used in conjunction with base models." ]
} ,
"loraWeight" : {
"heading" : "Weight" ,
"paragraphs" : [ "Weight of the LoRA. Higher weight will lead to larger impacts on the final image." ]
2023-09-15 14:37:05 +00:00
} ,
"noiseUseCPU" : {
"heading" : "Use CPU Noise" ,
feat(ui): refactor informational popover
- Change translations to use arrays of paragraphs instead of a single paragraph.
- Change component to accept a `feature` prop to identify the feature which the popover describes.
- Add optional `wrapperProps`: passed to the wrapper element, allowing more flexibility when using the popover
- Add optional `popoverProps`: passed to the `<Popover />` component, allowing for overriding individual instances of the popover's props
- Move definitions of features and popover settings to `invokeai/frontend/web/src/common/components/IAIInformationalPopover/constants.ts`
- Add some type safety to the `feature` prop
- Edit `POPOVER_DATA` to provide `image`, `href`, `buttonLabel`, and any popover props. The popover props are applied to all instances of the popover for the given feature. Note that the component prop `popoverProps` will override settings here.
- Remove the popover's arrow. Because the popover is wrapping groups of components, sometimes the error ends up pointing to nothing, which looks kinda janky. I've just removed the arrow entirely, but feel free to add it back if you think it looks better.
- Use a `link` variant button with external link icon to better communicate that clicking the button will open a new tab.
- Default the link button label to "Learn More" (if a label is provided, that will be used instead)
- Make default position `top`, but set manually set some to `right` - namely, anything with a dropdown. This prevents the popovers from obscuring or being obscured by the dropdowns.
- Do a bit more restructuring of the Popover component itself, and how it is integrated with other components
- More ref forwarding
- Make the open delay 1s
- Set the popovers to use lazy mounting (eg do not mount until the user opens the thing)
- Update the verbiage for many popover items and add missing dynamic prompts stuff
2023-09-22 11:04:35 +00:00
"paragraphs" : [
"Controls whether noise is generated on the CPU or GPU." ,
"With CPU Noise enabled, a particular seed will produce the same image on any machine." ,
"There is no performance impact to enabling CPU Noise."
]
2023-09-15 14:37:05 +00:00
} ,
2024-02-19 17:50:35 +00:00
"paramAspect" : {
"heading" : "Aspect" ,
"paragraphs" : [
"Aspect ratio of the generated image. Changing the ratio will update the Width and Height accordingly." ,
2024-03-13 11:15:09 +00:00
"\"Optimize\" will set the Width and Height to optimal dimensions for the chosen model."
2024-02-19 17:50:35 +00:00
]
} ,
2023-09-15 18:48:36 +00:00
"paramCFGScale" : {
2023-09-15 14:37:05 +00:00
"heading" : "CFG Scale" ,
2024-02-19 17:50:35 +00:00
"paragraphs" : [
"Controls how much the prompt influences the generation process." ,
"High CFG Scale values can result in over-saturation and distorted generation results. "
]
2023-09-15 14:37:05 +00:00
} ,
2023-11-30 09:55:20 +00:00
"paramCFGRescaleMultiplier" : {
"heading" : "CFG Rescale Multiplier" ,
"paragraphs" : [
2024-02-19 17:50:35 +00:00
"Rescale multiplier for CFG guidance, used for models trained using zero-terminal SNR (ztsnr)." ,
"Suggested value of 0.7 for these models."
2023-11-30 09:55:20 +00:00
]
} ,
2023-09-15 14:37:05 +00:00
"paramDenoisingStrength" : {
"heading" : "Denoising Strength" ,
feat(ui): refactor informational popover
- Change translations to use arrays of paragraphs instead of a single paragraph.
- Change component to accept a `feature` prop to identify the feature which the popover describes.
- Add optional `wrapperProps`: passed to the wrapper element, allowing more flexibility when using the popover
- Add optional `popoverProps`: passed to the `<Popover />` component, allowing for overriding individual instances of the popover's props
- Move definitions of features and popover settings to `invokeai/frontend/web/src/common/components/IAIInformationalPopover/constants.ts`
- Add some type safety to the `feature` prop
- Edit `POPOVER_DATA` to provide `image`, `href`, `buttonLabel`, and any popover props. The popover props are applied to all instances of the popover for the given feature. Note that the component prop `popoverProps` will override settings here.
- Remove the popover's arrow. Because the popover is wrapping groups of components, sometimes the error ends up pointing to nothing, which looks kinda janky. I've just removed the arrow entirely, but feel free to add it back if you think it looks better.
- Use a `link` variant button with external link icon to better communicate that clicking the button will open a new tab.
- Default the link button label to "Learn More" (if a label is provided, that will be used instead)
- Make default position `top`, but set manually set some to `right` - namely, anything with a dropdown. This prevents the popovers from obscuring or being obscured by the dropdowns.
- Do a bit more restructuring of the Popover component itself, and how it is integrated with other components
- More ref forwarding
- Make the open delay 1s
- Set the popovers to use lazy mounting (eg do not mount until the user opens the thing)
- Update the verbiage for many popover items and add missing dynamic prompts stuff
2023-09-22 11:04:35 +00:00
"paragraphs" : [
"How much noise is added to the input image." ,
"0 will result in an identical image, while 1 will result in a completely new image."
]
2023-09-15 14:37:05 +00:00
} ,
2024-02-19 17:50:35 +00:00
"paramHeight" : {
"heading" : "Height" ,
"paragraphs" : [ "Height of the generated image. Must be a multiple of 8." ]
} ,
"paramHrf" : {
"heading" : "Enable High Resolution Fix" ,
"paragraphs" : [
"Generate high quality images at a larger resolution than optimal for the model. Generally used to prevent duplication in the generated image."
]
} ,
2023-09-21 10:01:00 +00:00
"paramIterations" : {
"heading" : "Iterations" ,
feat(ui): refactor informational popover
- Change translations to use arrays of paragraphs instead of a single paragraph.
- Change component to accept a `feature` prop to identify the feature which the popover describes.
- Add optional `wrapperProps`: passed to the wrapper element, allowing more flexibility when using the popover
- Add optional `popoverProps`: passed to the `<Popover />` component, allowing for overriding individual instances of the popover's props
- Move definitions of features and popover settings to `invokeai/frontend/web/src/common/components/IAIInformationalPopover/constants.ts`
- Add some type safety to the `feature` prop
- Edit `POPOVER_DATA` to provide `image`, `href`, `buttonLabel`, and any popover props. The popover props are applied to all instances of the popover for the given feature. Note that the component prop `popoverProps` will override settings here.
- Remove the popover's arrow. Because the popover is wrapping groups of components, sometimes the error ends up pointing to nothing, which looks kinda janky. I've just removed the arrow entirely, but feel free to add it back if you think it looks better.
- Use a `link` variant button with external link icon to better communicate that clicking the button will open a new tab.
- Default the link button label to "Learn More" (if a label is provided, that will be used instead)
- Make default position `top`, but set manually set some to `right` - namely, anything with a dropdown. This prevents the popovers from obscuring or being obscured by the dropdowns.
- Do a bit more restructuring of the Popover component itself, and how it is integrated with other components
- More ref forwarding
- Make the open delay 1s
- Set the popovers to use lazy mounting (eg do not mount until the user opens the thing)
- Update the verbiage for many popover items and add missing dynamic prompts stuff
2023-09-22 11:04:35 +00:00
"paragraphs" : [
"The number of images to generate." ,
"If Dynamic Prompts is enabled, each of the prompts will be generated this many times."
]
2023-09-15 14:37:05 +00:00
} ,
"paramModel" : {
"heading" : "Model" ,
feat(ui): refactor informational popover
- Change translations to use arrays of paragraphs instead of a single paragraph.
- Change component to accept a `feature` prop to identify the feature which the popover describes.
- Add optional `wrapperProps`: passed to the wrapper element, allowing more flexibility when using the popover
- Add optional `popoverProps`: passed to the `<Popover />` component, allowing for overriding individual instances of the popover's props
- Move definitions of features and popover settings to `invokeai/frontend/web/src/common/components/IAIInformationalPopover/constants.ts`
- Add some type safety to the `feature` prop
- Edit `POPOVER_DATA` to provide `image`, `href`, `buttonLabel`, and any popover props. The popover props are applied to all instances of the popover for the given feature. Note that the component prop `popoverProps` will override settings here.
- Remove the popover's arrow. Because the popover is wrapping groups of components, sometimes the error ends up pointing to nothing, which looks kinda janky. I've just removed the arrow entirely, but feel free to add it back if you think it looks better.
- Use a `link` variant button with external link icon to better communicate that clicking the button will open a new tab.
- Default the link button label to "Learn More" (if a label is provided, that will be used instead)
- Make default position `top`, but set manually set some to `right` - namely, anything with a dropdown. This prevents the popovers from obscuring or being obscured by the dropdowns.
- Do a bit more restructuring of the Popover component itself, and how it is integrated with other components
- More ref forwarding
- Make the open delay 1s
- Set the popovers to use lazy mounting (eg do not mount until the user opens the thing)
- Update the verbiage for many popover items and add missing dynamic prompts stuff
2023-09-22 11:04:35 +00:00
"paragraphs" : [
2024-02-19 17:50:35 +00:00
"Model used for generation. Different models are trained to specialize in producing different aesthetic results and content."
feat(ui): refactor informational popover
- Change translations to use arrays of paragraphs instead of a single paragraph.
- Change component to accept a `feature` prop to identify the feature which the popover describes.
- Add optional `wrapperProps`: passed to the wrapper element, allowing more flexibility when using the popover
- Add optional `popoverProps`: passed to the `<Popover />` component, allowing for overriding individual instances of the popover's props
- Move definitions of features and popover settings to `invokeai/frontend/web/src/common/components/IAIInformationalPopover/constants.ts`
- Add some type safety to the `feature` prop
- Edit `POPOVER_DATA` to provide `image`, `href`, `buttonLabel`, and any popover props. The popover props are applied to all instances of the popover for the given feature. Note that the component prop `popoverProps` will override settings here.
- Remove the popover's arrow. Because the popover is wrapping groups of components, sometimes the error ends up pointing to nothing, which looks kinda janky. I've just removed the arrow entirely, but feel free to add it back if you think it looks better.
- Use a `link` variant button with external link icon to better communicate that clicking the button will open a new tab.
- Default the link button label to "Learn More" (if a label is provided, that will be used instead)
- Make default position `top`, but set manually set some to `right` - namely, anything with a dropdown. This prevents the popovers from obscuring or being obscured by the dropdowns.
- Do a bit more restructuring of the Popover component itself, and how it is integrated with other components
- More ref forwarding
- Make the open delay 1s
- Set the popovers to use lazy mounting (eg do not mount until the user opens the thing)
- Update the verbiage for many popover items and add missing dynamic prompts stuff
2023-09-22 11:04:35 +00:00
]
2023-09-15 14:37:05 +00:00
} ,
2023-09-15 18:48:36 +00:00
"paramRatio" : {
feat(ui): refactor informational popover
- Change translations to use arrays of paragraphs instead of a single paragraph.
- Change component to accept a `feature` prop to identify the feature which the popover describes.
- Add optional `wrapperProps`: passed to the wrapper element, allowing more flexibility when using the popover
- Add optional `popoverProps`: passed to the `<Popover />` component, allowing for overriding individual instances of the popover's props
- Move definitions of features and popover settings to `invokeai/frontend/web/src/common/components/IAIInformationalPopover/constants.ts`
- Add some type safety to the `feature` prop
- Edit `POPOVER_DATA` to provide `image`, `href`, `buttonLabel`, and any popover props. The popover props are applied to all instances of the popover for the given feature. Note that the component prop `popoverProps` will override settings here.
- Remove the popover's arrow. Because the popover is wrapping groups of components, sometimes the error ends up pointing to nothing, which looks kinda janky. I've just removed the arrow entirely, but feel free to add it back if you think it looks better.
- Use a `link` variant button with external link icon to better communicate that clicking the button will open a new tab.
- Default the link button label to "Learn More" (if a label is provided, that will be used instead)
- Make default position `top`, but set manually set some to `right` - namely, anything with a dropdown. This prevents the popovers from obscuring or being obscured by the dropdowns.
- Do a bit more restructuring of the Popover component itself, and how it is integrated with other components
- More ref forwarding
- Make the open delay 1s
- Set the popovers to use lazy mounting (eg do not mount until the user opens the thing)
- Update the verbiage for many popover items and add missing dynamic prompts stuff
2023-09-22 11:04:35 +00:00
"heading" : "Aspect Ratio" ,
"paragraphs" : [
"The aspect ratio of the dimensions of the image generated." ,
"An image size (in number of pixels) equivalent to 512x512 is recommended for SD1.5 models and a size equivalent to 1024x1024 is recommended for SDXL models."
]
2023-09-15 14:37:05 +00:00
} ,
"paramSeed" : {
"heading" : "Seed" ,
feat(ui): refactor informational popover
- Change translations to use arrays of paragraphs instead of a single paragraph.
- Change component to accept a `feature` prop to identify the feature which the popover describes.
- Add optional `wrapperProps`: passed to the wrapper element, allowing more flexibility when using the popover
- Add optional `popoverProps`: passed to the `<Popover />` component, allowing for overriding individual instances of the popover's props
- Move definitions of features and popover settings to `invokeai/frontend/web/src/common/components/IAIInformationalPopover/constants.ts`
- Add some type safety to the `feature` prop
- Edit `POPOVER_DATA` to provide `image`, `href`, `buttonLabel`, and any popover props. The popover props are applied to all instances of the popover for the given feature. Note that the component prop `popoverProps` will override settings here.
- Remove the popover's arrow. Because the popover is wrapping groups of components, sometimes the error ends up pointing to nothing, which looks kinda janky. I've just removed the arrow entirely, but feel free to add it back if you think it looks better.
- Use a `link` variant button with external link icon to better communicate that clicking the button will open a new tab.
- Default the link button label to "Learn More" (if a label is provided, that will be used instead)
- Make default position `top`, but set manually set some to `right` - namely, anything with a dropdown. This prevents the popovers from obscuring or being obscured by the dropdowns.
- Do a bit more restructuring of the Popover component itself, and how it is integrated with other components
- More ref forwarding
- Make the open delay 1s
- Set the popovers to use lazy mounting (eg do not mount until the user opens the thing)
- Update the verbiage for many popover items and add missing dynamic prompts stuff
2023-09-22 11:04:35 +00:00
"paragraphs" : [
"Controls the starting noise used for generation." ,
2024-03-13 11:15:09 +00:00
"Disable the \"Random\" option to produce identical results with the same generation settings."
feat(ui): refactor informational popover
- Change translations to use arrays of paragraphs instead of a single paragraph.
- Change component to accept a `feature` prop to identify the feature which the popover describes.
- Add optional `wrapperProps`: passed to the wrapper element, allowing more flexibility when using the popover
- Add optional `popoverProps`: passed to the `<Popover />` component, allowing for overriding individual instances of the popover's props
- Move definitions of features and popover settings to `invokeai/frontend/web/src/common/components/IAIInformationalPopover/constants.ts`
- Add some type safety to the `feature` prop
- Edit `POPOVER_DATA` to provide `image`, `href`, `buttonLabel`, and any popover props. The popover props are applied to all instances of the popover for the given feature. Note that the component prop `popoverProps` will override settings here.
- Remove the popover's arrow. Because the popover is wrapping groups of components, sometimes the error ends up pointing to nothing, which looks kinda janky. I've just removed the arrow entirely, but feel free to add it back if you think it looks better.
- Use a `link` variant button with external link icon to better communicate that clicking the button will open a new tab.
- Default the link button label to "Learn More" (if a label is provided, that will be used instead)
- Make default position `top`, but set manually set some to `right` - namely, anything with a dropdown. This prevents the popovers from obscuring or being obscured by the dropdowns.
- Do a bit more restructuring of the Popover component itself, and how it is integrated with other components
- More ref forwarding
- Make the open delay 1s
- Set the popovers to use lazy mounting (eg do not mount until the user opens the thing)
- Update the verbiage for many popover items and add missing dynamic prompts stuff
2023-09-22 11:04:35 +00:00
]
2023-09-15 14:37:05 +00:00
} ,
"paramSteps" : {
"heading" : "Steps" ,
feat(ui): refactor informational popover
- Change translations to use arrays of paragraphs instead of a single paragraph.
- Change component to accept a `feature` prop to identify the feature which the popover describes.
- Add optional `wrapperProps`: passed to the wrapper element, allowing more flexibility when using the popover
- Add optional `popoverProps`: passed to the `<Popover />` component, allowing for overriding individual instances of the popover's props
- Move definitions of features and popover settings to `invokeai/frontend/web/src/common/components/IAIInformationalPopover/constants.ts`
- Add some type safety to the `feature` prop
- Edit `POPOVER_DATA` to provide `image`, `href`, `buttonLabel`, and any popover props. The popover props are applied to all instances of the popover for the given feature. Note that the component prop `popoverProps` will override settings here.
- Remove the popover's arrow. Because the popover is wrapping groups of components, sometimes the error ends up pointing to nothing, which looks kinda janky. I've just removed the arrow entirely, but feel free to add it back if you think it looks better.
- Use a `link` variant button with external link icon to better communicate that clicking the button will open a new tab.
- Default the link button label to "Learn More" (if a label is provided, that will be used instead)
- Make default position `top`, but set manually set some to `right` - namely, anything with a dropdown. This prevents the popovers from obscuring or being obscured by the dropdowns.
- Do a bit more restructuring of the Popover component itself, and how it is integrated with other components
- More ref forwarding
- Make the open delay 1s
- Set the popovers to use lazy mounting (eg do not mount until the user opens the thing)
- Update the verbiage for many popover items and add missing dynamic prompts stuff
2023-09-22 11:04:35 +00:00
"paragraphs" : [
"Number of steps that will be performed in each generation." ,
"Higher step counts will typically create better images but will require more generation time."
]
2023-09-15 14:37:05 +00:00
} ,
2024-02-19 17:50:35 +00:00
"paramUpscaleMethod" : {
"heading" : "Upscale Method" ,
"paragraphs" : [ "Method used to upscale the image for High Resolution Fix." ]
} ,
2023-09-15 14:37:05 +00:00
"paramVAE" : {
"heading" : "VAE" ,
2024-01-27 09:55:55 +00:00
"paragraphs" : [ "Model used for translating AI output into the final image." ]
2023-09-15 14:37:05 +00:00
} ,
"paramVAEPrecision" : {
"heading" : "VAE Precision" ,
feat(ui): refactor informational popover
- Change translations to use arrays of paragraphs instead of a single paragraph.
- Change component to accept a `feature` prop to identify the feature which the popover describes.
- Add optional `wrapperProps`: passed to the wrapper element, allowing more flexibility when using the popover
- Add optional `popoverProps`: passed to the `<Popover />` component, allowing for overriding individual instances of the popover's props
- Move definitions of features and popover settings to `invokeai/frontend/web/src/common/components/IAIInformationalPopover/constants.ts`
- Add some type safety to the `feature` prop
- Edit `POPOVER_DATA` to provide `image`, `href`, `buttonLabel`, and any popover props. The popover props are applied to all instances of the popover for the given feature. Note that the component prop `popoverProps` will override settings here.
- Remove the popover's arrow. Because the popover is wrapping groups of components, sometimes the error ends up pointing to nothing, which looks kinda janky. I've just removed the arrow entirely, but feel free to add it back if you think it looks better.
- Use a `link` variant button with external link icon to better communicate that clicking the button will open a new tab.
- Default the link button label to "Learn More" (if a label is provided, that will be used instead)
- Make default position `top`, but set manually set some to `right` - namely, anything with a dropdown. This prevents the popovers from obscuring or being obscured by the dropdowns.
- Do a bit more restructuring of the Popover component itself, and how it is integrated with other components
- More ref forwarding
- Make the open delay 1s
- Set the popovers to use lazy mounting (eg do not mount until the user opens the thing)
- Update the verbiage for many popover items and add missing dynamic prompts stuff
2023-09-22 11:04:35 +00:00
"paragraphs" : [
2024-02-19 17:50:35 +00:00
"The precision used during VAE encoding and decoding." ,
"Fp16/Half precision is more efficient, at the expense of minor image variations."
]
} ,
"paramWidth" : {
"heading" : "Width" ,
"paragraphs" : [ "Width of the generated image. Must be a multiple of 8." ]
} ,
"patchmatchDownScaleSize" : {
"heading" : "Downscale" ,
"paragraphs" : [
"How much downscaling occurs before infilling." ,
"Higher downscaling will improve performance and reduce quality."
]
} ,
"refinerModel" : {
"heading" : "Refiner Model" ,
"paragraphs" : [
"Model used during the refiner portion of the generation process." ,
"Similar to the Generation Model."
]
} ,
"refinerPositiveAestheticScore" : {
"heading" : "Positive Aesthetic Score" ,
"paragraphs" : [
"Weight generations to be more similar to images with a high aesthetic score, based on the training data."
]
} ,
"refinerNegativeAestheticScore" : {
"heading" : "Negative Aesthetic Score" ,
"paragraphs" : [
"Weight generations to be more similar to images with a low aesthetic score, based on the training data."
]
} ,
"refinerScheduler" : {
"heading" : "Scheduler" ,
"paragraphs" : [
"Scheduler used during the refiner portion of the generation process." ,
"Similar to the Generation Scheduler."
]
} ,
"refinerStart" : {
"heading" : "Refiner Start" ,
"paragraphs" : [
"Where in the generation process the refiner will start to be used." ,
"0 means the refiner will be used for the entire generation process, 0.8 means the refiner will be used for the last 20% of the generation process."
]
} ,
"refinerSteps" : {
"heading" : "Steps" ,
"paragraphs" : [
"Number of steps that will be performed during the refiner portion of the generation process." ,
"Similar to the Generation Steps."
]
} ,
"refinerCfgScale" : {
"heading" : "CFG Scale" ,
"paragraphs" : [
"Controls how much the prompt influences the generation process." ,
"Similar to the Generation CFG Scale."
feat(ui): refactor informational popover
- Change translations to use arrays of paragraphs instead of a single paragraph.
- Change component to accept a `feature` prop to identify the feature which the popover describes.
- Add optional `wrapperProps`: passed to the wrapper element, allowing more flexibility when using the popover
- Add optional `popoverProps`: passed to the `<Popover />` component, allowing for overriding individual instances of the popover's props
- Move definitions of features and popover settings to `invokeai/frontend/web/src/common/components/IAIInformationalPopover/constants.ts`
- Add some type safety to the `feature` prop
- Edit `POPOVER_DATA` to provide `image`, `href`, `buttonLabel`, and any popover props. The popover props are applied to all instances of the popover for the given feature. Note that the component prop `popoverProps` will override settings here.
- Remove the popover's arrow. Because the popover is wrapping groups of components, sometimes the error ends up pointing to nothing, which looks kinda janky. I've just removed the arrow entirely, but feel free to add it back if you think it looks better.
- Use a `link` variant button with external link icon to better communicate that clicking the button will open a new tab.
- Default the link button label to "Learn More" (if a label is provided, that will be used instead)
- Make default position `top`, but set manually set some to `right` - namely, anything with a dropdown. This prevents the popovers from obscuring or being obscured by the dropdowns.
- Do a bit more restructuring of the Popover component itself, and how it is integrated with other components
- More ref forwarding
- Make the open delay 1s
- Set the popovers to use lazy mounting (eg do not mount until the user opens the thing)
- Update the verbiage for many popover items and add missing dynamic prompts stuff
2023-09-22 11:04:35 +00:00
]
2023-09-15 14:37:05 +00:00
} ,
"scaleBeforeProcessing" : {
"heading" : "Scale Before Processing" ,
feat(ui): refactor informational popover
- Change translations to use arrays of paragraphs instead of a single paragraph.
- Change component to accept a `feature` prop to identify the feature which the popover describes.
- Add optional `wrapperProps`: passed to the wrapper element, allowing more flexibility when using the popover
- Add optional `popoverProps`: passed to the `<Popover />` component, allowing for overriding individual instances of the popover's props
- Move definitions of features and popover settings to `invokeai/frontend/web/src/common/components/IAIInformationalPopover/constants.ts`
- Add some type safety to the `feature` prop
- Edit `POPOVER_DATA` to provide `image`, `href`, `buttonLabel`, and any popover props. The popover props are applied to all instances of the popover for the given feature. Note that the component prop `popoverProps` will override settings here.
- Remove the popover's arrow. Because the popover is wrapping groups of components, sometimes the error ends up pointing to nothing, which looks kinda janky. I've just removed the arrow entirely, but feel free to add it back if you think it looks better.
- Use a `link` variant button with external link icon to better communicate that clicking the button will open a new tab.
- Default the link button label to "Learn More" (if a label is provided, that will be used instead)
- Make default position `top`, but set manually set some to `right` - namely, anything with a dropdown. This prevents the popovers from obscuring or being obscured by the dropdowns.
- Do a bit more restructuring of the Popover component itself, and how it is integrated with other components
- More ref forwarding
- Make the open delay 1s
- Set the popovers to use lazy mounting (eg do not mount until the user opens the thing)
- Update the verbiage for many popover items and add missing dynamic prompts stuff
2023-09-22 11:04:35 +00:00
"paragraphs" : [
2024-03-13 11:15:09 +00:00
"\"Auto\" scales the selected area to the size best suited for the model before the image generation process." ,
"\"Manual\" allows you to choose the width and height the selected area will be scaled to before the image generation process."
feat(ui): refactor informational popover
- Change translations to use arrays of paragraphs instead of a single paragraph.
- Change component to accept a `feature` prop to identify the feature which the popover describes.
- Add optional `wrapperProps`: passed to the wrapper element, allowing more flexibility when using the popover
- Add optional `popoverProps`: passed to the `<Popover />` component, allowing for overriding individual instances of the popover's props
- Move definitions of features and popover settings to `invokeai/frontend/web/src/common/components/IAIInformationalPopover/constants.ts`
- Add some type safety to the `feature` prop
- Edit `POPOVER_DATA` to provide `image`, `href`, `buttonLabel`, and any popover props. The popover props are applied to all instances of the popover for the given feature. Note that the component prop `popoverProps` will override settings here.
- Remove the popover's arrow. Because the popover is wrapping groups of components, sometimes the error ends up pointing to nothing, which looks kinda janky. I've just removed the arrow entirely, but feel free to add it back if you think it looks better.
- Use a `link` variant button with external link icon to better communicate that clicking the button will open a new tab.
- Default the link button label to "Learn More" (if a label is provided, that will be used instead)
- Make default position `top`, but set manually set some to `right` - namely, anything with a dropdown. This prevents the popovers from obscuring or being obscured by the dropdowns.
- Do a bit more restructuring of the Popover component itself, and how it is integrated with other components
- More ref forwarding
- Make the open delay 1s
- Set the popovers to use lazy mounting (eg do not mount until the user opens the thing)
- Update the verbiage for many popover items and add missing dynamic prompts stuff
2023-09-22 11:04:35 +00:00
]
2024-02-19 17:50:35 +00:00
} ,
"seamlessTilingXAxis" : {
"heading" : "Seamless Tiling X Axis" ,
"paragraphs" : [ "Seamlessly tile an image along the horizontal axis." ]
} ,
"seamlessTilingYAxis" : {
"heading" : "Seamless Tiling Y Axis" ,
"paragraphs" : [ "Seamlessly tile an image along the vertical axis." ]
2024-07-24 23:23:14 +00:00
} ,
"upscaleModel" : {
"heading" : "Upscale Model" ,
"paragraphs" : [
2024-07-25 13:30:21 +00:00
"The upscale model scales the image to the output size before details are added. Any supported upscale model may be used, but some are specialized for different kinds of images, like photos or line drawings."
2024-07-24 23:23:14 +00:00
]
} ,
"scale" : {
"heading" : "Scale" ,
"paragraphs" : [
"Scale controls the output image size, and is based on a multiple of the input image resolution. For example a 2x upscale on a 1024x1024 image would produce a 2048 x 2048 output."
]
} ,
"creativity" : {
"heading" : "Creativity" ,
"paragraphs" : [
2024-07-25 13:27:33 +00:00
"Creativity controls the amount of freedom granted to the model when adding details. Low creativity stays close to the original image, while high creativity allows for more change. When using a prompt, high creativity increases the influence of the prompt."
2024-07-24 23:23:14 +00:00
]
} ,
"structure" : {
"heading" : "Structure" ,
"paragraphs" : [
2024-07-25 13:27:33 +00:00
"Structure controls how closely the output image will keep to the layout of the original. Low structure allows major changes, while high structure strictly maintains the original composition and layout."
2024-07-24 23:23:14 +00:00
]
2023-09-15 14:37:05 +00:00
}
} ,
2023-02-18 04:23:24 +00:00
"unifiedCanvas" : {
2023-09-13 07:31:34 +00:00
"accept" : "Accept" ,
"activeLayer" : "Active Layer" ,
"antialiasing" : "Antialiasing" ,
"autoSaveToGallery" : "Auto Save to Gallery" ,
2023-02-18 04:23:24 +00:00
"base" : "Base" ,
2023-09-13 07:31:34 +00:00
"boundingBox" : "Bounding Box" ,
"boundingBoxPosition" : "Bounding Box Position" ,
2023-02-18 04:23:24 +00:00
"brush" : "Brush" ,
"brushOptions" : "Brush Options" ,
"brushSize" : "Size" ,
2023-09-13 07:31:34 +00:00
"canvasDimensions" : "Canvas Dimensions" ,
"canvasPosition" : "Canvas Position" ,
"canvasScale" : "Canvas Scale" ,
2023-02-18 04:23:24 +00:00
"canvasSettings" : "Canvas Settings" ,
2023-09-13 07:31:34 +00:00
"clearCanvas" : "Clear Canvas" ,
2023-02-18 04:23:24 +00:00
"clearCanvasHistory" : "Clear Canvas History" ,
"clearCanvasHistoryConfirm" : "Are you sure you want to clear the canvas history?" ,
2023-09-13 07:31:34 +00:00
"clearCanvasHistoryMessage" : "Clearing the canvas history leaves your current canvas intact, but irreversibly clears the undo and redo history." ,
"clearHistory" : "Clear History" ,
2023-11-25 03:46:19 +00:00
"clearMask" : "Clear Mask (Shift+C)" ,
2024-02-21 17:32:49 +00:00
"coherenceModeGaussianBlur" : "Gaussian Blur" ,
"coherenceModeBoxBlur" : "Box Blur" ,
"coherenceModeStaged" : "Staged" ,
2023-09-13 07:31:34 +00:00
"colorPicker" : "Color Picker" ,
"copyToClipboard" : "Copy to Clipboard" ,
"cursorPosition" : "Cursor Position" ,
"darkenOutsideSelection" : "Darken Outside Selection" ,
"discardAll" : "Discard All" ,
2024-03-06 02:12:11 +00:00
"discardCurrent" : "Discard Current" ,
2023-09-13 07:31:34 +00:00
"downloadAsImage" : "Download As Image" ,
"enableMask" : "Enable Mask" ,
"eraseBoundingBox" : "Erase Bounding Box" ,
"eraser" : "Eraser" ,
"fillBoundingBox" : "Fill Bounding Box" ,
2024-04-11 23:53:02 +00:00
"hideBoundingBox" : "Hide Bounding Box" ,
2024-04-04 10:41:22 +00:00
"initialFitImageSize" : "Fit Image Size on Drop" ,
2024-03-06 03:16:39 +00:00
"invertBrushSizeScrollDirection" : "Invert Scroll for Brush Size" ,
2023-09-13 07:31:34 +00:00
"layer" : "Layer" ,
"limitStrokesToBox" : "Limit Strokes to Box" ,
"mask" : "Mask" ,
"maskingOptions" : "Masking Options" ,
"mergeVisible" : "Merge Visible" ,
"move" : "Move" ,
2023-02-18 04:23:24 +00:00
"next" : "Next" ,
2023-09-13 07:31:34 +00:00
"preserveMaskedArea" : "Preserve Masked Area" ,
"previous" : "Previous" ,
"redo" : "Redo" ,
"resetView" : "Reset View" ,
"saveBoxRegionOnly" : "Save Box Region Only" ,
2023-11-26 09:44:39 +00:00
"saveMask" : "Save $t(unifiedCanvas.mask)" ,
2023-09-13 07:31:34 +00:00
"saveToGallery" : "Save To Gallery" ,
"scaledBoundingBox" : "Scaled Bounding Box" ,
2024-04-11 23:53:02 +00:00
"showBoundingBox" : "Show Bounding Box" ,
2023-09-13 07:31:34 +00:00
"showCanvasDebugInfo" : "Show Additional Canvas Info" ,
"showGrid" : "Show Grid" ,
2023-09-27 07:12:57 +00:00
"showResultsOn" : "Show Results (On)" ,
"showResultsOff" : "Show Results (Off)" ,
2023-09-13 07:31:34 +00:00
"showIntermediates" : "Show Intermediates" ,
"snapToGrid" : "Snap to Grid" ,
"undo" : "Undo"
feat: workflow library (#5148)
* chore: bump pydantic to 2.5.2
This release fixes pydantic/pydantic#8175 and allows us to use `JsonValue`
* fix(ui): exclude public/en.json from prettier config
* fix(workflow_records): fix SQLite workflow insertion to ignore duplicates
* feat(backend): update workflows handling
Update workflows handling for Workflow Library.
**Updated Workflow Storage**
"Embedded Workflows" are workflows associated with images, and are now only stored in the image files. "Library Workflows" are not associated with images, and are stored only in DB.
This works out nicely. We have always saved workflows to files, but recently began saving them to the DB in addition to in image files. When that happened, we stopped reading workflows from files, so all the workflows that only existed in images were inaccessible. With this change, access to those workflows is restored, and no workflows are lost.
**Updated Workflow Handling in Nodes**
Prior to this change, workflows were embedded in images by passing the whole workflow JSON to a special workflow field on a node. In the node's `invoke()` function, the node was able to access this workflow and save it with the image. This (inaccurately) models workflows as a property of an image and is rather awkward technically.
A workflow is now a property of a batch/session queue item. It is available in the InvocationContext and therefore available to all nodes during `invoke()`.
**Database Migrations**
Added a `SQLiteMigrator` class to handle database migrations. Migrations were needed to accomodate the DB-related changes in this PR. See the code for details.
The `images`, `workflows` and `session_queue` tables required migrations for this PR, and are using the new migrator. Other tables/services are still creating tables themselves. A followup PR will adapt them to use the migrator.
**Other/Support Changes**
- Add a `has_workflow` column to `images` table to indicate that the image has an embedded workflow.
- Add handling for retrieving the workflow from an image in python. The image file must be fetched, the workflow extracted, and then sent to client, avoiding needing the browser to parse the image file. With the `has_workflow` column, the UI knows if there is a workflow to be fetched, and only fetches when the user requests to load the workflow.
- Add route to get the workflow from an image
- Add CRUD service/routes for the library workflows
- `workflow_images` table and services removed (no longer needed now that embedded workflows are not in the DB)
* feat(ui): updated workflow handling (WIP)
Clientside updates for the backend workflow changes.
Includes roughed-out workflow library UI.
* feat: revert SQLiteMigrator class
Will pursue this in a separate PR.
* feat(nodes): do not overwrite custom node module names
Use a different, simpler method to detect if a node is custom.
* feat(nodes): restore WithWorkflow as no-op class
This class is deprecated and no longer needed. Set its workflow attr value to None (meaning it is now a no-op), and issue a warning when an invocation subclasses it.
* fix(nodes): fix get_workflow from queue item dict func
* feat(backend): add WorkflowRecordListItemDTO
This is the id, name, description, created at and updated at workflow columns/attrs. Used to display lists of workflowsl
* chore(ui): typegen
* feat(ui): add workflow loading, deleting to workflow library UI
* feat(ui): workflow library pagination button styles
* wip
* feat: workflow library WIP
- Save to library
- Duplicate
- Filter/sort
- UI/queries
* feat: workflow library - system graphs - wip
* feat(backend): sync system workflows to db
* fix: merge conflicts
* feat: simplify default workflows
- Rename "system" -> "default"
- Simplify syncing logic
- Update UI to match
* feat(workflows): update default workflows
- Update TextToImage_SD15
- Add TextToImage_SDXL
- Add README
* feat(ui): refine workflow list UI
* fix(workflow_records): typo
* fix(tests): fix tests
* feat(ui): clean up workflow library hooks
* fix(db): fix mis-ordered db cleanup step
It was happening before pruning queue items - should happen afterwards, else you have to restart the app again to free disk space made available by the pruning.
* feat(ui): tweak reset workflow editor translations
* feat(ui): split out workflow redux state
The `nodes` slice is a rather complicated slice. Removing `workflow` makes it a bit more reasonable.
Also helps to flatten state out a bit.
* docs: update default workflows README
* fix: tidy up unused files, unrelated changes
* fix(backend): revert unrelated service organisational changes
* feat(backend): workflow_records.get_many arg "filter_text" -> "query"
* feat(ui): use custom hook in current image buttons
Already in use elsewhere, forgot to use it here.
* fix(ui): remove commented out property
* fix(ui): fix workflow loading
- Different handling for loading from library vs external
- Fix bug where only nodes and edges loaded
* fix(ui): fix save/save-as workflow naming
* fix(ui): fix circular dependency
* fix(db): fix bug with releasing without lock in db.clean()
* fix(db): remove extraneous lock
* chore: bump ruff
* fix(workflow_records): default `category` to `WorkflowCategory.User`
This allows old workflows to validate when reading them from the db or image files.
* hide workflow library buttons if feature is disabled
---------
Co-authored-by: Mary Hipp <maryhipp@Marys-MacBook-Air.local>
2023-12-08 22:48:38 +00:00
} ,
"workflows" : {
2024-03-13 18:57:56 +00:00
"ascending" : "Ascending" ,
"created" : "Created" ,
2024-03-20 05:21:49 +00:00
"descending" : "Descending" ,
feat: workflow library (#5148)
* chore: bump pydantic to 2.5.2
This release fixes pydantic/pydantic#8175 and allows us to use `JsonValue`
* fix(ui): exclude public/en.json from prettier config
* fix(workflow_records): fix SQLite workflow insertion to ignore duplicates
* feat(backend): update workflows handling
Update workflows handling for Workflow Library.
**Updated Workflow Storage**
"Embedded Workflows" are workflows associated with images, and are now only stored in the image files. "Library Workflows" are not associated with images, and are stored only in DB.
This works out nicely. We have always saved workflows to files, but recently began saving them to the DB in addition to in image files. When that happened, we stopped reading workflows from files, so all the workflows that only existed in images were inaccessible. With this change, access to those workflows is restored, and no workflows are lost.
**Updated Workflow Handling in Nodes**
Prior to this change, workflows were embedded in images by passing the whole workflow JSON to a special workflow field on a node. In the node's `invoke()` function, the node was able to access this workflow and save it with the image. This (inaccurately) models workflows as a property of an image and is rather awkward technically.
A workflow is now a property of a batch/session queue item. It is available in the InvocationContext and therefore available to all nodes during `invoke()`.
**Database Migrations**
Added a `SQLiteMigrator` class to handle database migrations. Migrations were needed to accomodate the DB-related changes in this PR. See the code for details.
The `images`, `workflows` and `session_queue` tables required migrations for this PR, and are using the new migrator. Other tables/services are still creating tables themselves. A followup PR will adapt them to use the migrator.
**Other/Support Changes**
- Add a `has_workflow` column to `images` table to indicate that the image has an embedded workflow.
- Add handling for retrieving the workflow from an image in python. The image file must be fetched, the workflow extracted, and then sent to client, avoiding needing the browser to parse the image file. With the `has_workflow` column, the UI knows if there is a workflow to be fetched, and only fetches when the user requests to load the workflow.
- Add route to get the workflow from an image
- Add CRUD service/routes for the library workflows
- `workflow_images` table and services removed (no longer needed now that embedded workflows are not in the DB)
* feat(ui): updated workflow handling (WIP)
Clientside updates for the backend workflow changes.
Includes roughed-out workflow library UI.
* feat: revert SQLiteMigrator class
Will pursue this in a separate PR.
* feat(nodes): do not overwrite custom node module names
Use a different, simpler method to detect if a node is custom.
* feat(nodes): restore WithWorkflow as no-op class
This class is deprecated and no longer needed. Set its workflow attr value to None (meaning it is now a no-op), and issue a warning when an invocation subclasses it.
* fix(nodes): fix get_workflow from queue item dict func
* feat(backend): add WorkflowRecordListItemDTO
This is the id, name, description, created at and updated at workflow columns/attrs. Used to display lists of workflowsl
* chore(ui): typegen
* feat(ui): add workflow loading, deleting to workflow library UI
* feat(ui): workflow library pagination button styles
* wip
* feat: workflow library WIP
- Save to library
- Duplicate
- Filter/sort
- UI/queries
* feat: workflow library - system graphs - wip
* feat(backend): sync system workflows to db
* fix: merge conflicts
* feat: simplify default workflows
- Rename "system" -> "default"
- Simplify syncing logic
- Update UI to match
* feat(workflows): update default workflows
- Update TextToImage_SD15
- Add TextToImage_SDXL
- Add README
* feat(ui): refine workflow list UI
* fix(workflow_records): typo
* fix(tests): fix tests
* feat(ui): clean up workflow library hooks
* fix(db): fix mis-ordered db cleanup step
It was happening before pruning queue items - should happen afterwards, else you have to restart the app again to free disk space made available by the pruning.
* feat(ui): tweak reset workflow editor translations
* feat(ui): split out workflow redux state
The `nodes` slice is a rather complicated slice. Removing `workflow` makes it a bit more reasonable.
Also helps to flatten state out a bit.
* docs: update default workflows README
* fix: tidy up unused files, unrelated changes
* fix(backend): revert unrelated service organisational changes
* feat(backend): workflow_records.get_many arg "filter_text" -> "query"
* feat(ui): use custom hook in current image buttons
Already in use elsewhere, forgot to use it here.
* fix(ui): remove commented out property
* fix(ui): fix workflow loading
- Different handling for loading from library vs external
- Fix bug where only nodes and edges loaded
* fix(ui): fix save/save-as workflow naming
* fix(ui): fix circular dependency
* fix(db): fix bug with releasing without lock in db.clean()
* fix(db): remove extraneous lock
* chore: bump ruff
* fix(workflow_records): default `category` to `WorkflowCategory.User`
This allows old workflows to validate when reading them from the db or image files.
* hide workflow library buttons if feature is disabled
---------
Co-authored-by: Mary Hipp <maryhipp@Marys-MacBook-Air.local>
2023-12-08 22:48:38 +00:00
"workflows" : "Workflows" ,
2023-12-06 09:15:26 +00:00
"workflowLibrary" : "Library" ,
2024-03-13 18:57:56 +00:00
"userWorkflows" : "My Workflows" ,
"defaultWorkflows" : "Default Workflows" ,
"projectWorkflows" : "Project Workflows" ,
"opened" : "Opened" ,
feat: workflow library (#5148)
* chore: bump pydantic to 2.5.2
This release fixes pydantic/pydantic#8175 and allows us to use `JsonValue`
* fix(ui): exclude public/en.json from prettier config
* fix(workflow_records): fix SQLite workflow insertion to ignore duplicates
* feat(backend): update workflows handling
Update workflows handling for Workflow Library.
**Updated Workflow Storage**
"Embedded Workflows" are workflows associated with images, and are now only stored in the image files. "Library Workflows" are not associated with images, and are stored only in DB.
This works out nicely. We have always saved workflows to files, but recently began saving them to the DB in addition to in image files. When that happened, we stopped reading workflows from files, so all the workflows that only existed in images were inaccessible. With this change, access to those workflows is restored, and no workflows are lost.
**Updated Workflow Handling in Nodes**
Prior to this change, workflows were embedded in images by passing the whole workflow JSON to a special workflow field on a node. In the node's `invoke()` function, the node was able to access this workflow and save it with the image. This (inaccurately) models workflows as a property of an image and is rather awkward technically.
A workflow is now a property of a batch/session queue item. It is available in the InvocationContext and therefore available to all nodes during `invoke()`.
**Database Migrations**
Added a `SQLiteMigrator` class to handle database migrations. Migrations were needed to accomodate the DB-related changes in this PR. See the code for details.
The `images`, `workflows` and `session_queue` tables required migrations for this PR, and are using the new migrator. Other tables/services are still creating tables themselves. A followup PR will adapt them to use the migrator.
**Other/Support Changes**
- Add a `has_workflow` column to `images` table to indicate that the image has an embedded workflow.
- Add handling for retrieving the workflow from an image in python. The image file must be fetched, the workflow extracted, and then sent to client, avoiding needing the browser to parse the image file. With the `has_workflow` column, the UI knows if there is a workflow to be fetched, and only fetches when the user requests to load the workflow.
- Add route to get the workflow from an image
- Add CRUD service/routes for the library workflows
- `workflow_images` table and services removed (no longer needed now that embedded workflows are not in the DB)
* feat(ui): updated workflow handling (WIP)
Clientside updates for the backend workflow changes.
Includes roughed-out workflow library UI.
* feat: revert SQLiteMigrator class
Will pursue this in a separate PR.
* feat(nodes): do not overwrite custom node module names
Use a different, simpler method to detect if a node is custom.
* feat(nodes): restore WithWorkflow as no-op class
This class is deprecated and no longer needed. Set its workflow attr value to None (meaning it is now a no-op), and issue a warning when an invocation subclasses it.
* fix(nodes): fix get_workflow from queue item dict func
* feat(backend): add WorkflowRecordListItemDTO
This is the id, name, description, created at and updated at workflow columns/attrs. Used to display lists of workflowsl
* chore(ui): typegen
* feat(ui): add workflow loading, deleting to workflow library UI
* feat(ui): workflow library pagination button styles
* wip
* feat: workflow library WIP
- Save to library
- Duplicate
- Filter/sort
- UI/queries
* feat: workflow library - system graphs - wip
* feat(backend): sync system workflows to db
* fix: merge conflicts
* feat: simplify default workflows
- Rename "system" -> "default"
- Simplify syncing logic
- Update UI to match
* feat(workflows): update default workflows
- Update TextToImage_SD15
- Add TextToImage_SDXL
- Add README
* feat(ui): refine workflow list UI
* fix(workflow_records): typo
* fix(tests): fix tests
* feat(ui): clean up workflow library hooks
* fix(db): fix mis-ordered db cleanup step
It was happening before pruning queue items - should happen afterwards, else you have to restart the app again to free disk space made available by the pruning.
* feat(ui): tweak reset workflow editor translations
* feat(ui): split out workflow redux state
The `nodes` slice is a rather complicated slice. Removing `workflow` makes it a bit more reasonable.
Also helps to flatten state out a bit.
* docs: update default workflows README
* fix: tidy up unused files, unrelated changes
* fix(backend): revert unrelated service organisational changes
* feat(backend): workflow_records.get_many arg "filter_text" -> "query"
* feat(ui): use custom hook in current image buttons
Already in use elsewhere, forgot to use it here.
* fix(ui): remove commented out property
* fix(ui): fix workflow loading
- Different handling for loading from library vs external
- Fix bug where only nodes and edges loaded
* fix(ui): fix save/save-as workflow naming
* fix(ui): fix circular dependency
* fix(db): fix bug with releasing without lock in db.clean()
* fix(db): remove extraneous lock
* chore: bump ruff
* fix(workflow_records): default `category` to `WorkflowCategory.User`
This allows old workflows to validate when reading them from the db or image files.
* hide workflow library buttons if feature is disabled
---------
Co-authored-by: Mary Hipp <maryhipp@Marys-MacBook-Air.local>
2023-12-08 22:48:38 +00:00
"openWorkflow" : "Open Workflow" ,
2024-03-13 18:57:56 +00:00
"updated" : "Updated" ,
2023-12-14 08:55:47 +00:00
"uploadWorkflow" : "Load from File" ,
feat: workflow library (#5148)
* chore: bump pydantic to 2.5.2
This release fixes pydantic/pydantic#8175 and allows us to use `JsonValue`
* fix(ui): exclude public/en.json from prettier config
* fix(workflow_records): fix SQLite workflow insertion to ignore duplicates
* feat(backend): update workflows handling
Update workflows handling for Workflow Library.
**Updated Workflow Storage**
"Embedded Workflows" are workflows associated with images, and are now only stored in the image files. "Library Workflows" are not associated with images, and are stored only in DB.
This works out nicely. We have always saved workflows to files, but recently began saving them to the DB in addition to in image files. When that happened, we stopped reading workflows from files, so all the workflows that only existed in images were inaccessible. With this change, access to those workflows is restored, and no workflows are lost.
**Updated Workflow Handling in Nodes**
Prior to this change, workflows were embedded in images by passing the whole workflow JSON to a special workflow field on a node. In the node's `invoke()` function, the node was able to access this workflow and save it with the image. This (inaccurately) models workflows as a property of an image and is rather awkward technically.
A workflow is now a property of a batch/session queue item. It is available in the InvocationContext and therefore available to all nodes during `invoke()`.
**Database Migrations**
Added a `SQLiteMigrator` class to handle database migrations. Migrations were needed to accomodate the DB-related changes in this PR. See the code for details.
The `images`, `workflows` and `session_queue` tables required migrations for this PR, and are using the new migrator. Other tables/services are still creating tables themselves. A followup PR will adapt them to use the migrator.
**Other/Support Changes**
- Add a `has_workflow` column to `images` table to indicate that the image has an embedded workflow.
- Add handling for retrieving the workflow from an image in python. The image file must be fetched, the workflow extracted, and then sent to client, avoiding needing the browser to parse the image file. With the `has_workflow` column, the UI knows if there is a workflow to be fetched, and only fetches when the user requests to load the workflow.
- Add route to get the workflow from an image
- Add CRUD service/routes for the library workflows
- `workflow_images` table and services removed (no longer needed now that embedded workflows are not in the DB)
* feat(ui): updated workflow handling (WIP)
Clientside updates for the backend workflow changes.
Includes roughed-out workflow library UI.
* feat: revert SQLiteMigrator class
Will pursue this in a separate PR.
* feat(nodes): do not overwrite custom node module names
Use a different, simpler method to detect if a node is custom.
* feat(nodes): restore WithWorkflow as no-op class
This class is deprecated and no longer needed. Set its workflow attr value to None (meaning it is now a no-op), and issue a warning when an invocation subclasses it.
* fix(nodes): fix get_workflow from queue item dict func
* feat(backend): add WorkflowRecordListItemDTO
This is the id, name, description, created at and updated at workflow columns/attrs. Used to display lists of workflowsl
* chore(ui): typegen
* feat(ui): add workflow loading, deleting to workflow library UI
* feat(ui): workflow library pagination button styles
* wip
* feat: workflow library WIP
- Save to library
- Duplicate
- Filter/sort
- UI/queries
* feat: workflow library - system graphs - wip
* feat(backend): sync system workflows to db
* fix: merge conflicts
* feat: simplify default workflows
- Rename "system" -> "default"
- Simplify syncing logic
- Update UI to match
* feat(workflows): update default workflows
- Update TextToImage_SD15
- Add TextToImage_SDXL
- Add README
* feat(ui): refine workflow list UI
* fix(workflow_records): typo
* fix(tests): fix tests
* feat(ui): clean up workflow library hooks
* fix(db): fix mis-ordered db cleanup step
It was happening before pruning queue items - should happen afterwards, else you have to restart the app again to free disk space made available by the pruning.
* feat(ui): tweak reset workflow editor translations
* feat(ui): split out workflow redux state
The `nodes` slice is a rather complicated slice. Removing `workflow` makes it a bit more reasonable.
Also helps to flatten state out a bit.
* docs: update default workflows README
* fix: tidy up unused files, unrelated changes
* fix(backend): revert unrelated service organisational changes
* feat(backend): workflow_records.get_many arg "filter_text" -> "query"
* feat(ui): use custom hook in current image buttons
Already in use elsewhere, forgot to use it here.
* fix(ui): remove commented out property
* fix(ui): fix workflow loading
- Different handling for loading from library vs external
- Fix bug where only nodes and edges loaded
* fix(ui): fix save/save-as workflow naming
* fix(ui): fix circular dependency
* fix(db): fix bug with releasing without lock in db.clean()
* fix(db): remove extraneous lock
* chore: bump ruff
* fix(workflow_records): default `category` to `WorkflowCategory.User`
This allows old workflows to validate when reading them from the db or image files.
* hide workflow library buttons if feature is disabled
---------
Co-authored-by: Mary Hipp <maryhipp@Marys-MacBook-Air.local>
2023-12-08 22:48:38 +00:00
"deleteWorkflow" : "Delete Workflow" ,
"unnamedWorkflow" : "Unnamed Workflow" ,
2023-12-14 08:55:47 +00:00
"downloadWorkflow" : "Save to File" ,
feat: workflow library (#5148)
* chore: bump pydantic to 2.5.2
This release fixes pydantic/pydantic#8175 and allows us to use `JsonValue`
* fix(ui): exclude public/en.json from prettier config
* fix(workflow_records): fix SQLite workflow insertion to ignore duplicates
* feat(backend): update workflows handling
Update workflows handling for Workflow Library.
**Updated Workflow Storage**
"Embedded Workflows" are workflows associated with images, and are now only stored in the image files. "Library Workflows" are not associated with images, and are stored only in DB.
This works out nicely. We have always saved workflows to files, but recently began saving them to the DB in addition to in image files. When that happened, we stopped reading workflows from files, so all the workflows that only existed in images were inaccessible. With this change, access to those workflows is restored, and no workflows are lost.
**Updated Workflow Handling in Nodes**
Prior to this change, workflows were embedded in images by passing the whole workflow JSON to a special workflow field on a node. In the node's `invoke()` function, the node was able to access this workflow and save it with the image. This (inaccurately) models workflows as a property of an image and is rather awkward technically.
A workflow is now a property of a batch/session queue item. It is available in the InvocationContext and therefore available to all nodes during `invoke()`.
**Database Migrations**
Added a `SQLiteMigrator` class to handle database migrations. Migrations were needed to accomodate the DB-related changes in this PR. See the code for details.
The `images`, `workflows` and `session_queue` tables required migrations for this PR, and are using the new migrator. Other tables/services are still creating tables themselves. A followup PR will adapt them to use the migrator.
**Other/Support Changes**
- Add a `has_workflow` column to `images` table to indicate that the image has an embedded workflow.
- Add handling for retrieving the workflow from an image in python. The image file must be fetched, the workflow extracted, and then sent to client, avoiding needing the browser to parse the image file. With the `has_workflow` column, the UI knows if there is a workflow to be fetched, and only fetches when the user requests to load the workflow.
- Add route to get the workflow from an image
- Add CRUD service/routes for the library workflows
- `workflow_images` table and services removed (no longer needed now that embedded workflows are not in the DB)
* feat(ui): updated workflow handling (WIP)
Clientside updates for the backend workflow changes.
Includes roughed-out workflow library UI.
* feat: revert SQLiteMigrator class
Will pursue this in a separate PR.
* feat(nodes): do not overwrite custom node module names
Use a different, simpler method to detect if a node is custom.
* feat(nodes): restore WithWorkflow as no-op class
This class is deprecated and no longer needed. Set its workflow attr value to None (meaning it is now a no-op), and issue a warning when an invocation subclasses it.
* fix(nodes): fix get_workflow from queue item dict func
* feat(backend): add WorkflowRecordListItemDTO
This is the id, name, description, created at and updated at workflow columns/attrs. Used to display lists of workflowsl
* chore(ui): typegen
* feat(ui): add workflow loading, deleting to workflow library UI
* feat(ui): workflow library pagination button styles
* wip
* feat: workflow library WIP
- Save to library
- Duplicate
- Filter/sort
- UI/queries
* feat: workflow library - system graphs - wip
* feat(backend): sync system workflows to db
* fix: merge conflicts
* feat: simplify default workflows
- Rename "system" -> "default"
- Simplify syncing logic
- Update UI to match
* feat(workflows): update default workflows
- Update TextToImage_SD15
- Add TextToImage_SDXL
- Add README
* feat(ui): refine workflow list UI
* fix(workflow_records): typo
* fix(tests): fix tests
* feat(ui): clean up workflow library hooks
* fix(db): fix mis-ordered db cleanup step
It was happening before pruning queue items - should happen afterwards, else you have to restart the app again to free disk space made available by the pruning.
* feat(ui): tweak reset workflow editor translations
* feat(ui): split out workflow redux state
The `nodes` slice is a rather complicated slice. Removing `workflow` makes it a bit more reasonable.
Also helps to flatten state out a bit.
* docs: update default workflows README
* fix: tidy up unused files, unrelated changes
* fix(backend): revert unrelated service organisational changes
* feat(backend): workflow_records.get_many arg "filter_text" -> "query"
* feat(ui): use custom hook in current image buttons
Already in use elsewhere, forgot to use it here.
* fix(ui): remove commented out property
* fix(ui): fix workflow loading
- Different handling for loading from library vs external
- Fix bug where only nodes and edges loaded
* fix(ui): fix save/save-as workflow naming
* fix(ui): fix circular dependency
* fix(db): fix bug with releasing without lock in db.clean()
* fix(db): remove extraneous lock
* chore: bump ruff
* fix(workflow_records): default `category` to `WorkflowCategory.User`
This allows old workflows to validate when reading them from the db or image files.
* hide workflow library buttons if feature is disabled
---------
Co-authored-by: Mary Hipp <maryhipp@Marys-MacBook-Air.local>
2023-12-08 22:48:38 +00:00
"saveWorkflow" : "Save Workflow" ,
"saveWorkflowAs" : "Save Workflow As" ,
2024-01-31 13:32:31 +00:00
"saveWorkflowToProject" : "Save Workflow to Project" ,
2023-12-06 09:15:26 +00:00
"savingWorkflow" : "Saving Workflow..." ,
feat: workflow library (#5148)
* chore: bump pydantic to 2.5.2
This release fixes pydantic/pydantic#8175 and allows us to use `JsonValue`
* fix(ui): exclude public/en.json from prettier config
* fix(workflow_records): fix SQLite workflow insertion to ignore duplicates
* feat(backend): update workflows handling
Update workflows handling for Workflow Library.
**Updated Workflow Storage**
"Embedded Workflows" are workflows associated with images, and are now only stored in the image files. "Library Workflows" are not associated with images, and are stored only in DB.
This works out nicely. We have always saved workflows to files, but recently began saving them to the DB in addition to in image files. When that happened, we stopped reading workflows from files, so all the workflows that only existed in images were inaccessible. With this change, access to those workflows is restored, and no workflows are lost.
**Updated Workflow Handling in Nodes**
Prior to this change, workflows were embedded in images by passing the whole workflow JSON to a special workflow field on a node. In the node's `invoke()` function, the node was able to access this workflow and save it with the image. This (inaccurately) models workflows as a property of an image and is rather awkward technically.
A workflow is now a property of a batch/session queue item. It is available in the InvocationContext and therefore available to all nodes during `invoke()`.
**Database Migrations**
Added a `SQLiteMigrator` class to handle database migrations. Migrations were needed to accomodate the DB-related changes in this PR. See the code for details.
The `images`, `workflows` and `session_queue` tables required migrations for this PR, and are using the new migrator. Other tables/services are still creating tables themselves. A followup PR will adapt them to use the migrator.
**Other/Support Changes**
- Add a `has_workflow` column to `images` table to indicate that the image has an embedded workflow.
- Add handling for retrieving the workflow from an image in python. The image file must be fetched, the workflow extracted, and then sent to client, avoiding needing the browser to parse the image file. With the `has_workflow` column, the UI knows if there is a workflow to be fetched, and only fetches when the user requests to load the workflow.
- Add route to get the workflow from an image
- Add CRUD service/routes for the library workflows
- `workflow_images` table and services removed (no longer needed now that embedded workflows are not in the DB)
* feat(ui): updated workflow handling (WIP)
Clientside updates for the backend workflow changes.
Includes roughed-out workflow library UI.
* feat: revert SQLiteMigrator class
Will pursue this in a separate PR.
* feat(nodes): do not overwrite custom node module names
Use a different, simpler method to detect if a node is custom.
* feat(nodes): restore WithWorkflow as no-op class
This class is deprecated and no longer needed. Set its workflow attr value to None (meaning it is now a no-op), and issue a warning when an invocation subclasses it.
* fix(nodes): fix get_workflow from queue item dict func
* feat(backend): add WorkflowRecordListItemDTO
This is the id, name, description, created at and updated at workflow columns/attrs. Used to display lists of workflowsl
* chore(ui): typegen
* feat(ui): add workflow loading, deleting to workflow library UI
* feat(ui): workflow library pagination button styles
* wip
* feat: workflow library WIP
- Save to library
- Duplicate
- Filter/sort
- UI/queries
* feat: workflow library - system graphs - wip
* feat(backend): sync system workflows to db
* fix: merge conflicts
* feat: simplify default workflows
- Rename "system" -> "default"
- Simplify syncing logic
- Update UI to match
* feat(workflows): update default workflows
- Update TextToImage_SD15
- Add TextToImage_SDXL
- Add README
* feat(ui): refine workflow list UI
* fix(workflow_records): typo
* fix(tests): fix tests
* feat(ui): clean up workflow library hooks
* fix(db): fix mis-ordered db cleanup step
It was happening before pruning queue items - should happen afterwards, else you have to restart the app again to free disk space made available by the pruning.
* feat(ui): tweak reset workflow editor translations
* feat(ui): split out workflow redux state
The `nodes` slice is a rather complicated slice. Removing `workflow` makes it a bit more reasonable.
Also helps to flatten state out a bit.
* docs: update default workflows README
* fix: tidy up unused files, unrelated changes
* fix(backend): revert unrelated service organisational changes
* feat(backend): workflow_records.get_many arg "filter_text" -> "query"
* feat(ui): use custom hook in current image buttons
Already in use elsewhere, forgot to use it here.
* fix(ui): remove commented out property
* fix(ui): fix workflow loading
- Different handling for loading from library vs external
- Fix bug where only nodes and edges loaded
* fix(ui): fix save/save-as workflow naming
* fix(ui): fix circular dependency
* fix(db): fix bug with releasing without lock in db.clean()
* fix(db): remove extraneous lock
* chore: bump ruff
* fix(workflow_records): default `category` to `WorkflowCategory.User`
This allows old workflows to validate when reading them from the db or image files.
* hide workflow library buttons if feature is disabled
---------
Co-authored-by: Mary Hipp <maryhipp@Marys-MacBook-Air.local>
2023-12-08 22:48:38 +00:00
"problemSavingWorkflow" : "Problem Saving Workflow" ,
"workflowSaved" : "Workflow Saved" ,
2024-03-13 18:57:56 +00:00
"name" : "Name" ,
"noRecentWorkflows" : "No Recent Workflows" ,
"noUserWorkflows" : "No User Workflows" ,
2024-01-23 14:49:50 +00:00
"noWorkflows" : "No Workflows" ,
feat: workflow library (#5148)
* chore: bump pydantic to 2.5.2
This release fixes pydantic/pydantic#8175 and allows us to use `JsonValue`
* fix(ui): exclude public/en.json from prettier config
* fix(workflow_records): fix SQLite workflow insertion to ignore duplicates
* feat(backend): update workflows handling
Update workflows handling for Workflow Library.
**Updated Workflow Storage**
"Embedded Workflows" are workflows associated with images, and are now only stored in the image files. "Library Workflows" are not associated with images, and are stored only in DB.
This works out nicely. We have always saved workflows to files, but recently began saving them to the DB in addition to in image files. When that happened, we stopped reading workflows from files, so all the workflows that only existed in images were inaccessible. With this change, access to those workflows is restored, and no workflows are lost.
**Updated Workflow Handling in Nodes**
Prior to this change, workflows were embedded in images by passing the whole workflow JSON to a special workflow field on a node. In the node's `invoke()` function, the node was able to access this workflow and save it with the image. This (inaccurately) models workflows as a property of an image and is rather awkward technically.
A workflow is now a property of a batch/session queue item. It is available in the InvocationContext and therefore available to all nodes during `invoke()`.
**Database Migrations**
Added a `SQLiteMigrator` class to handle database migrations. Migrations were needed to accomodate the DB-related changes in this PR. See the code for details.
The `images`, `workflows` and `session_queue` tables required migrations for this PR, and are using the new migrator. Other tables/services are still creating tables themselves. A followup PR will adapt them to use the migrator.
**Other/Support Changes**
- Add a `has_workflow` column to `images` table to indicate that the image has an embedded workflow.
- Add handling for retrieving the workflow from an image in python. The image file must be fetched, the workflow extracted, and then sent to client, avoiding needing the browser to parse the image file. With the `has_workflow` column, the UI knows if there is a workflow to be fetched, and only fetches when the user requests to load the workflow.
- Add route to get the workflow from an image
- Add CRUD service/routes for the library workflows
- `workflow_images` table and services removed (no longer needed now that embedded workflows are not in the DB)
* feat(ui): updated workflow handling (WIP)
Clientside updates for the backend workflow changes.
Includes roughed-out workflow library UI.
* feat: revert SQLiteMigrator class
Will pursue this in a separate PR.
* feat(nodes): do not overwrite custom node module names
Use a different, simpler method to detect if a node is custom.
* feat(nodes): restore WithWorkflow as no-op class
This class is deprecated and no longer needed. Set its workflow attr value to None (meaning it is now a no-op), and issue a warning when an invocation subclasses it.
* fix(nodes): fix get_workflow from queue item dict func
* feat(backend): add WorkflowRecordListItemDTO
This is the id, name, description, created at and updated at workflow columns/attrs. Used to display lists of workflowsl
* chore(ui): typegen
* feat(ui): add workflow loading, deleting to workflow library UI
* feat(ui): workflow library pagination button styles
* wip
* feat: workflow library WIP
- Save to library
- Duplicate
- Filter/sort
- UI/queries
* feat: workflow library - system graphs - wip
* feat(backend): sync system workflows to db
* fix: merge conflicts
* feat: simplify default workflows
- Rename "system" -> "default"
- Simplify syncing logic
- Update UI to match
* feat(workflows): update default workflows
- Update TextToImage_SD15
- Add TextToImage_SDXL
- Add README
* feat(ui): refine workflow list UI
* fix(workflow_records): typo
* fix(tests): fix tests
* feat(ui): clean up workflow library hooks
* fix(db): fix mis-ordered db cleanup step
It was happening before pruning queue items - should happen afterwards, else you have to restart the app again to free disk space made available by the pruning.
* feat(ui): tweak reset workflow editor translations
* feat(ui): split out workflow redux state
The `nodes` slice is a rather complicated slice. Removing `workflow` makes it a bit more reasonable.
Also helps to flatten state out a bit.
* docs: update default workflows README
* fix: tidy up unused files, unrelated changes
* fix(backend): revert unrelated service organisational changes
* feat(backend): workflow_records.get_many arg "filter_text" -> "query"
* feat(ui): use custom hook in current image buttons
Already in use elsewhere, forgot to use it here.
* fix(ui): remove commented out property
* fix(ui): fix workflow loading
- Different handling for loading from library vs external
- Fix bug where only nodes and edges loaded
* fix(ui): fix save/save-as workflow naming
* fix(ui): fix circular dependency
* fix(db): fix bug with releasing without lock in db.clean()
* fix(db): remove extraneous lock
* chore: bump ruff
* fix(workflow_records): default `category` to `WorkflowCategory.User`
This allows old workflows to validate when reading them from the db or image files.
* hide workflow library buttons if feature is disabled
---------
Co-authored-by: Mary Hipp <maryhipp@Marys-MacBook-Air.local>
2023-12-08 22:48:38 +00:00
"problemLoading" : "Problem Loading Workflows" ,
"loading" : "Loading Workflows" ,
"noDescription" : "No description" ,
"searchWorkflows" : "Search Workflows" ,
"clearWorkflowSearchFilter" : "Clear Workflow Search Filter" ,
"workflowName" : "Workflow Name" ,
2023-12-14 08:55:47 +00:00
"newWorkflowCreated" : "New Workflow Created" ,
2024-01-31 13:32:31 +00:00
"workflowCleared" : "Workflow Cleared" ,
2024-04-09 00:17:03 +00:00
"workflowEditorMenu" : "Workflow Editor Menu" ,
"loadFromGraph" : "Load Workflow from Graph" ,
"convertGraph" : "Convert Graph" ,
"loadWorkflow" : "$t(common.load) Workflow" ,
"autoLayout" : "Auto Layout"
feat: workflow library (#5148)
* chore: bump pydantic to 2.5.2
This release fixes pydantic/pydantic#8175 and allows us to use `JsonValue`
* fix(ui): exclude public/en.json from prettier config
* fix(workflow_records): fix SQLite workflow insertion to ignore duplicates
* feat(backend): update workflows handling
Update workflows handling for Workflow Library.
**Updated Workflow Storage**
"Embedded Workflows" are workflows associated with images, and are now only stored in the image files. "Library Workflows" are not associated with images, and are stored only in DB.
This works out nicely. We have always saved workflows to files, but recently began saving them to the DB in addition to in image files. When that happened, we stopped reading workflows from files, so all the workflows that only existed in images were inaccessible. With this change, access to those workflows is restored, and no workflows are lost.
**Updated Workflow Handling in Nodes**
Prior to this change, workflows were embedded in images by passing the whole workflow JSON to a special workflow field on a node. In the node's `invoke()` function, the node was able to access this workflow and save it with the image. This (inaccurately) models workflows as a property of an image and is rather awkward technically.
A workflow is now a property of a batch/session queue item. It is available in the InvocationContext and therefore available to all nodes during `invoke()`.
**Database Migrations**
Added a `SQLiteMigrator` class to handle database migrations. Migrations were needed to accomodate the DB-related changes in this PR. See the code for details.
The `images`, `workflows` and `session_queue` tables required migrations for this PR, and are using the new migrator. Other tables/services are still creating tables themselves. A followup PR will adapt them to use the migrator.
**Other/Support Changes**
- Add a `has_workflow` column to `images` table to indicate that the image has an embedded workflow.
- Add handling for retrieving the workflow from an image in python. The image file must be fetched, the workflow extracted, and then sent to client, avoiding needing the browser to parse the image file. With the `has_workflow` column, the UI knows if there is a workflow to be fetched, and only fetches when the user requests to load the workflow.
- Add route to get the workflow from an image
- Add CRUD service/routes for the library workflows
- `workflow_images` table and services removed (no longer needed now that embedded workflows are not in the DB)
* feat(ui): updated workflow handling (WIP)
Clientside updates for the backend workflow changes.
Includes roughed-out workflow library UI.
* feat: revert SQLiteMigrator class
Will pursue this in a separate PR.
* feat(nodes): do not overwrite custom node module names
Use a different, simpler method to detect if a node is custom.
* feat(nodes): restore WithWorkflow as no-op class
This class is deprecated and no longer needed. Set its workflow attr value to None (meaning it is now a no-op), and issue a warning when an invocation subclasses it.
* fix(nodes): fix get_workflow from queue item dict func
* feat(backend): add WorkflowRecordListItemDTO
This is the id, name, description, created at and updated at workflow columns/attrs. Used to display lists of workflowsl
* chore(ui): typegen
* feat(ui): add workflow loading, deleting to workflow library UI
* feat(ui): workflow library pagination button styles
* wip
* feat: workflow library WIP
- Save to library
- Duplicate
- Filter/sort
- UI/queries
* feat: workflow library - system graphs - wip
* feat(backend): sync system workflows to db
* fix: merge conflicts
* feat: simplify default workflows
- Rename "system" -> "default"
- Simplify syncing logic
- Update UI to match
* feat(workflows): update default workflows
- Update TextToImage_SD15
- Add TextToImage_SDXL
- Add README
* feat(ui): refine workflow list UI
* fix(workflow_records): typo
* fix(tests): fix tests
* feat(ui): clean up workflow library hooks
* fix(db): fix mis-ordered db cleanup step
It was happening before pruning queue items - should happen afterwards, else you have to restart the app again to free disk space made available by the pruning.
* feat(ui): tweak reset workflow editor translations
* feat(ui): split out workflow redux state
The `nodes` slice is a rather complicated slice. Removing `workflow` makes it a bit more reasonable.
Also helps to flatten state out a bit.
* docs: update default workflows README
* fix: tidy up unused files, unrelated changes
* fix(backend): revert unrelated service organisational changes
* feat(backend): workflow_records.get_many arg "filter_text" -> "query"
* feat(ui): use custom hook in current image buttons
Already in use elsewhere, forgot to use it here.
* fix(ui): remove commented out property
* fix(ui): fix workflow loading
- Different handling for loading from library vs external
- Fix bug where only nodes and edges loaded
* fix(ui): fix save/save-as workflow naming
* fix(ui): fix circular dependency
* fix(db): fix bug with releasing without lock in db.clean()
* fix(db): remove extraneous lock
* chore: bump ruff
* fix(workflow_records): default `category` to `WorkflowCategory.User`
This allows old workflows to validate when reading them from the db or image files.
* hide workflow library buttons if feature is disabled
---------
Co-authored-by: Mary Hipp <maryhipp@Marys-MacBook-Air.local>
2023-12-08 22:48:38 +00:00
} ,
"app" : {
"storeNotInitialized" : "Store is not initialized"
2024-04-09 09:09:34 +00:00
} ,
2024-04-30 03:28:03 +00:00
"controlLayers" : {
2024-04-19 12:30:20 +00:00
"deleteAll" : "Delete All" ,
2024-04-09 09:09:34 +00:00
"addLayer" : "Add Layer" ,
"moveToFront" : "Move to Front" ,
"moveToBack" : "Move to Back" ,
"moveForward" : "Move Forward" ,
"moveBackward" : "Move Backward" ,
2024-04-19 12:30:20 +00:00
"brushSize" : "Brush Size" ,
2024-05-03 22:37:59 +00:00
"controlLayers" : "Control Layers" ,
2024-04-22 12:12:29 +00:00
"globalMaskOpacity" : "Global Mask Opacity" ,
2024-04-19 12:59:03 +00:00
"autoNegative" : "Auto Negative" ,
2024-04-22 12:12:29 +00:00
"deletePrompt" : "Delete Prompt" ,
2024-04-19 13:11:06 +00:00
"resetRegion" : "Reset Region" ,
2024-04-20 13:21:32 +00:00
"debugLayers" : "Debug Layers" ,
2024-04-22 12:12:29 +00:00
"rectangle" : "Rectangle" ,
"maskPreviewColor" : "Mask Preview Color" ,
"addPositivePrompt" : "Add $t(common.positivePrompt)" ,
"addNegativePrompt" : "Add $t(common.negativePrompt)" ,
2024-04-29 13:31:33 +00:00
"addIPAdapter" : "Add $t(common.ipAdapter)" ,
2024-04-30 03:41:29 +00:00
"regionalGuidance" : "Regional Guidance" ,
"regionalGuidanceLayer" : "$t(controlLayers.regionalGuidance) $t(unifiedCanvas.layer)" ,
2024-04-30 03:51:07 +00:00
"opacity" : "Opacity" ,
"globalControlAdapter" : "Global $t(controlnet.controlAdapter_one)" ,
"globalControlAdapterLayer" : "Global $t(controlnet.controlAdapter_one) $t(unifiedCanvas.layer)" ,
"globalIPAdapter" : "Global $t(common.ipAdapter)" ,
2024-04-30 08:50:29 +00:00
"globalIPAdapterLayer" : "Global $t(common.ipAdapter) $t(unifiedCanvas.layer)" ,
2024-05-02 11:12:58 +00:00
"globalInitialImage" : "Global Initial Image" ,
"globalInitialImageLayer" : "$t(controlLayers.globalInitialImage) $t(unifiedCanvas.layer)" ,
2024-05-01 22:46:21 +00:00
"opacityFilter" : "Opacity Filter" ,
"clearProcessor" : "Clear Processor" ,
2024-05-03 01:34:01 +00:00
"resetProcessor" : "Reset Processor to Defaults" ,
2024-05-07 07:47:09 +00:00
"noLayersAdded" : "No Layers Added" ,
"layers_one" : "Layer" ,
"layers_other" : "Layers"
2024-05-02 20:27:25 +00:00
} ,
2024-07-19 00:26:17 +00:00
"upscaling" : {
2024-08-18 22:44:08 +00:00
"upscale" : "Upscale" ,
2024-07-19 00:26:17 +00:00
"creativity" : "Creativity" ,
2024-08-05 23:19:39 +00:00
"exceedsMaxSize" : "Upscale settings exceed max size limit" ,
"exceedsMaxSizeDetails" : "Max upscale limit is {{maxUpscaleDimension}}x{{maxUpscaleDimension}} pixels. Please try a smaller image or decrease your scale selection." ,
2024-07-19 00:26:17 +00:00
"structure" : "Structure" ,
2024-07-23 03:07:45 +00:00
"upscaleModel" : "Upscale Model" ,
2024-07-23 21:32:33 +00:00
"postProcessingModel" : "Post-Processing Model" ,
2024-07-23 00:01:07 +00:00
"scale" : "Scale" ,
2024-07-23 21:35:52 +00:00
"postProcessingMissingModelWarning" : "Visit the <LinkComponent>Model Manager</LinkComponent> to install a post-processing (image to image) model." ,
2024-07-23 04:22:35 +00:00
"missingModelsWarning" : "Visit the <LinkComponent>Model Manager</LinkComponent> to install the required models:" ,
"mainModelDesc" : "Main model (SD1.5 or SDXL architecture)" ,
"tileControlNetModelDesc" : "Tile ControlNet model for the chosen main model architecture" ,
"upscaleModelDesc" : "Upscale (image to image) model" ,
"missingUpscaleInitialImage" : "Missing initial image for upscaling" ,
"missingUpscaleModel" : "Missing upscale model" ,
"missingTileControlNetModel" : "No valid tile ControlNet models installed"
2024-07-19 00:26:17 +00:00
} ,
2024-08-08 17:56:37 +00:00
"stylePresets" : {
"active" : "Active" ,
"choosePromptTemplate" : "Choose Prompt Template" ,
2024-08-08 18:18:05 +00:00
"clearTemplateSelection" : "Clear Template Selection" ,
2024-08-12 16:32:31 +00:00
"copyTemplate" : "Copy Template" ,
2024-08-08 17:56:37 +00:00
"createPromptTemplate" : "Create Prompt Template" ,
"defaultTemplates" : "Default Templates" ,
"deleteImage" : "Delete Image" ,
"deleteTemplate" : "Delete Template" ,
"deleteTemplate2" : "Are you sure you want to delete this template? This cannot be undone." ,
2024-08-15 23:58:19 +00:00
"exportPromptTemplates" : "Export My Prompt Templates (CSV)" ,
2024-08-08 17:56:37 +00:00
"editTemplate" : "Edit Template" ,
2024-08-15 16:39:55 +00:00
"exportDownloaded" : "Export Downloaded" ,
"exportFailed" : "Unable to generate and download CSV" ,
2024-08-08 18:18:05 +00:00
"flatten" : "Flatten selected template into current prompt" ,
2024-08-15 23:58:19 +00:00
"importTemplates" : "Import Prompt Templates (CSV/JSON)" ,
"acceptedColumnsKeys" : "Accepted columns/keys:" ,
"nameColumn" : "'name'" ,
"positivePromptColumn" : "'prompt' or 'positive_prompt'" ,
"negativePromptColumn" : "'negative_prompt'" ,
2024-08-08 18:18:05 +00:00
"insertPlaceholder" : "Insert placeholder" ,
2024-08-08 17:56:37 +00:00
"myTemplates" : "My Templates" ,
"name" : "Name" ,
"negativePrompt" : "Negative Prompt" ,
2024-08-16 00:11:43 +00:00
"noTemplates" : "No templates" ,
2024-08-08 17:56:37 +00:00
"noMatchingTemplates" : "No matching templates" ,
2024-08-12 17:37:04 +00:00
"promptTemplatesDesc1" : "Prompt templates add text to the prompts you write in the prompt box." ,
"promptTemplatesDesc2" : "Use the placeholder string <Pre>{{placeholder}}</Pre> to specify where your prompt should be included in the template." ,
"promptTemplatesDesc3" : "If you omit the placeholder, the template will be appended to the end of your prompt." ,
2024-08-08 17:56:37 +00:00
"positivePrompt" : "Positive Prompt" ,
2024-08-12 17:33:39 +00:00
"preview" : "Preview" ,
2024-08-13 18:08:54 +00:00
"private" : "Private" ,
2024-08-20 19:04:03 +00:00
"promptTemplateCleared" : "Prompt Template Cleared" ,
2024-08-08 18:18:05 +00:00
"searchByName" : "Search by name" ,
2024-08-13 18:08:54 +00:00
"shared" : "Shared" ,
"sharedTemplates" : "Shared Templates" ,
2024-08-15 16:39:55 +00:00
"templateActions" : "Template Actions" ,
2024-08-08 17:56:37 +00:00
"templateDeleted" : "Prompt template deleted" ,
2024-08-08 18:18:05 +00:00
"toggleViewMode" : "Toggle View Mode" ,
2024-08-13 18:08:54 +00:00
"type" : "Type" ,
2024-08-08 17:56:37 +00:00
"unableToDeleteTemplate" : "Unable to delete prompt template" ,
"updatePromptTemplate" : "Update Prompt Template" ,
"uploadImage" : "Upload Image" ,
2024-08-08 18:18:05 +00:00
"useForTemplate" : "Use For Prompt Template" ,
2024-08-09 20:00:13 +00:00
"viewList" : "View Template List" ,
2024-08-08 18:18:05 +00:00
"viewModeTooltip" : "This is how your prompt will look with your currently selected template. To edit your prompt, click anywhere in the text box."
2024-08-08 17:56:37 +00:00
} ,
2024-07-23 14:34:11 +00:00
"upsell" : {
"inviteTeammates" : "Invite Teammates" ,
"professional" : "Professional" ,
"professionalUpsell" : "Available in Invoke’ ,
"shareAccess" : "Share Access"
} ,
2024-05-02 20:27:25 +00:00
"ui" : {
"tabs" : {
"generation" : "Generation" ,
2024-05-02 21:04:14 +00:00
"generationTab" : "$t(ui.tabs.generation) $t(common.tab)" ,
2024-05-02 20:27:25 +00:00
"canvas" : "Canvas" ,
2024-05-02 21:04:14 +00:00
"canvasTab" : "$t(ui.tabs.canvas) $t(common.tab)" ,
2024-05-02 20:27:25 +00:00
"workflows" : "Workflows" ,
2024-05-02 21:04:14 +00:00
"workflowsTab" : "$t(ui.tabs.workflows) $t(common.tab)" ,
2024-05-02 20:27:25 +00:00
"models" : "Models" ,
2024-05-02 21:04:14 +00:00
"modelsTab" : "$t(ui.tabs.models) $t(common.tab)" ,
"queue" : "Queue" ,
2024-07-22 22:44:26 +00:00
"queueTab" : "$t(ui.tabs.queue) $t(common.tab)" ,
"upscaling" : "Upscaling" ,
"upscalingTab" : "$t(ui.tabs.upscaling) $t(common.tab)"
2024-05-02 20:27:25 +00:00
}
2023-02-18 04:23:24 +00:00
}
}
