psychedelicious
9b48029bc9
tidy(mm): ModelImages service
2024-03-06 21:57:41 -05:00
Jennifer Player
4af5a09a68
cleanup
2024-03-06 21:57:41 -05:00
Jennifer Player
8df02623f2
cleanup
2024-03-06 21:57:41 -05:00
Jennifer Player
aa88fadc30
use webp images
2024-03-06 21:57:41 -05:00
Jennifer Player
8411029d93
get model image url from model config, added thumbnail formatting for images
2024-03-06 21:57:41 -05:00
Jennifer Player
239b1e8cc7
moved upload image field and added delete image functionality
2024-03-06 21:57:41 -05:00
Jennifer Player
2f6964bfa5
fetching model image, still not working
2024-03-06 21:57:41 -05:00
Jennifer Player
c1cdfd132b
moved model image to edit page, added model_images service
2024-03-06 21:57:41 -05:00
psychedelicious
528ac5dd25
refactor(nodes): model identifiers
...
- All models are identified by a key and optionally a submodel type via new model `ModelField`. Previously, a few model types had their own class, but not all of them. This inconsistency just added complexity without any benefit.
- Update all invocation to use the new format.
- In the node API, models are loaded by key or an instance of `ModelField` as a convenience.
- Add an enriched model schema for metadata. It includes key, hash, name, base and type.
2024-03-07 10:56:59 +11:00
psychedelicious
afd9ae7712
tidy(mm): remove convenience methods from high level model manager service
...
These were added as a hold-me-over for the nodes API changes, no longer needed. A followup commit will fix the nodes API to not rely on these.
2024-03-07 10:56:59 +11:00
maryhipp
cdc0d0c182
add config_path to ModelRecordChanges
2024-03-07 10:29:29 +11:00
psychedelicious
b0615bdfd4
fix(nodes): correctly serialize outputs
...
In order for delete by match to work, we need the whole invocation output to be stringified.
For some reason, the serialization of the output was set to only include the `type` field. It should instead include the whole output.
I don't understand how this ever worked unless pydantic had different serialization behaviour in v1 (though it appears to have been the same).
Closes #5805
2024-03-06 08:14:12 -05:00
psychedelicious
bab20467fb
fix(nodes): fix invocation cache clear method args
2024-03-06 08:14:12 -05:00
psychedelicious
e24624109e
fix(nodes): fix invocation cache ABC typing
2024-03-06 08:14:12 -05:00
psychedelicious
bcf58cac59
feat(mm): add config to skip model hash
...
This is useful for when you are using a memory DB and do not want to wait for all models to be hashed on startup.
2024-03-05 23:50:19 +11:00
psychedelicious
e8797787cf
fix(mm): fix incorrect calls to update_model
2024-03-05 23:50:19 +11:00
psychedelicious
0082ecb22b
feat(mm): add path to ModelRecordChanges
2024-03-05 23:50:19 +11:00
psychedelicious
656839fcd1
fix(mm): fix typing on heuristic_import
2024-03-05 23:50:19 +11:00
psychedelicious
78895b3e80
fix(mm): add missing inplace parameter to model install abc
2024-03-05 23:50:19 +11:00
psychedelicious
58fa9c2fac
fix(mm): do not allow extra fields on ModelRecordChanges
2024-03-05 23:50:19 +11:00
psychedelicious
5551cf8ac4
feat(mm): revise update_model to use ModelRecordChanges
2024-03-05 23:50:19 +11:00
psychedelicious
37b969d339
tidy(mm): add default_settings to model config
2024-03-05 23:50:19 +11:00
psychedelicious
c953e61294
tidy(mm): "trigger_words" -> "trigger_phrases"
2024-03-05 23:50:19 +11:00
psychedelicious
44c40d7d1a
refactor(mm): remove unused metadata logic, fix tests
...
- Metadata is merged with the config. We can simplify the MM substantially and remove the handling for metadata.
- Per discussion, we don't have an ETA for frontend implementation of tags, and with the realization that the tags from CivitAI are largely useless, there's no reason to keep tags in the MM right now. When we are ready to implement tags on the frontend, we can refer back to the implementation here and use it if it supports the design.
- Fix all tests.
2024-03-05 23:50:19 +11:00
psychedelicious
c3aa985c93
refactor(mm): get metadata working
2024-03-05 23:50:19 +11:00
psychedelicious
7cb0da1f66
refactor(mm): wip schema changes
2024-03-05 23:50:19 +11:00
psychedelicious
f13f5984c0
fix(mm): update db schema & migration
2024-03-05 23:50:19 +11:00
psychedelicious
9378e47a06
feat(mm): add source_type to model configs
2024-03-05 23:50:19 +11:00
psychedelicious
2c835fd550
refactor(mm): WIP db schema
2024-03-05 23:50:19 +11:00
psychedelicious
a8cd3dfc99
refactor(mm): add models table (schema WIP), rename "original_hash" -> "hash"
2024-03-05 23:50:19 +11:00
psychedelicious
9b40c28144
tidy(mm): rename ckpy "config" -> "config_path"
2024-03-05 23:50:19 +11:00
psychedelicious
76cbc745e1
refactor(mm): add CheckpointConfigBase for all ckpt models
2024-03-05 23:50:19 +11:00
psychedelicious
b880a31039
refactor(mm): remove ztsnr_training field on _MainConfig
...
This is used to determine the CFG Rescale Multiplier setting. We'll handle this in the UI as a default setting.
2024-03-05 23:50:19 +11:00
Mary Hipp Rogers
8b34f5298c
Default model settings ( #5850 )
...
* UI in MM to create trigger phrases
* add scheduler and vaePrecision to config
* UI for configuring default settings for models'
* hook MM default model settings up to API
* add button to set default settings in parameters
* pull out trigger phrases
* back-end for default settings
* lint
* remove log;
gi
* ruff
* ruff format
---------
Co-authored-by: Mary Hipp <maryhipp@Marys-MacBook-Air.local>
2024-03-04 09:39:03 -05:00
Brandon Rising
893bcd16fc
Next: Allow in place local installs of models
2024-03-04 23:11:41 +11:00
Ryan Dick
f6028a4c61
Log a stack trace for invocation errors.
2024-03-04 23:01:56 +11:00
psychedelicious
ae99428883
fix(mm): use UUIDv4 for key
...
This changes the functionality of this PR to only use the updated hashing for model hashes with a UUID for the key.
2024-03-03 14:32:14 +11:00
psychedelicious
86982f3059
feat(mm): make ModelHash instantiatable, taking an algorithm as arg
2024-03-03 14:32:14 +11:00
psychedelicious
982076d7d7
feat(mm): add hashing algos to ModelHash
...
- Some algos are slow, so it is now just called ModelHash
- Added all hashlib algos, plus BLAKE3 and the fast (but incorrect) SHA1 algo
2024-03-03 14:32:14 +11:00
Lincoln Stein
a72056e0df
make model key assignment deterministic
...
- When installing, model keys are now calculated from the model contents.
- .safetensors, .ckpt and other single file models are hashed with sha1
- The contents of diffusers directories are hashed using imohash (faster)
fixup yaml->sql db migration script to assign deterministic key
- this commit also detects and assigns the correct image encoder for
ip adapter models.
2024-03-03 14:32:14 +11:00
Lincoln Stein
0832e1818e
Fix problem of all installed models being assigned "<NOKEY>"
...
- Also fix redundant scanning of models directory at startup.
2024-02-29 22:31:05 -05:00
Lincoln Stein
813a086cfe
fix race condition between downloading last file and starting install
2024-03-01 10:42:33 +11:00
Lincoln Stein
e18533e3b5
add debugging statements and a timeout to download test
2024-03-01 10:42:33 +11:00
psychedelicious
dd9daf8efb
chore: ruff
2024-03-01 10:42:33 +11:00
psychedelicious
753919c6d7
docs(nodes): update all docstrings for public nodes API
2024-03-01 10:42:33 +11:00
psychedelicious
0b0128647b
feat(nodes): revise model load API args
2024-03-01 10:42:33 +11:00
psychedelicious
12e859835b
feat(mm): add log stmt for download complete event
2024-03-01 10:42:33 +11:00
psychedelicious
ea7b7bcf40
chore: ruff
2024-03-01 10:42:33 +11:00
psychedelicious
e22c4987bf
chore: ruff
2024-03-01 10:42:33 +11:00
maryhipp
26a209a00d
add error_reason to ModelInstallJob
2024-03-01 10:42:33 +11:00
Jennifer Player
ea364bdf82
delete model imports and prune all finished, update state with socket messages
2024-03-01 10:42:33 +11:00
Brandon Rising
c778ab8db4
Allow passing in key on register
2024-03-01 10:42:33 +11:00
Brandon Rising
65b91356d0
Remove passing keys in on register
2024-03-01 10:42:33 +11:00
Brandon Rising
de9287a3e4
Run ruff
2024-03-01 10:42:33 +11:00
Brandon Rising
008716040b
Allow users to run model manager without cuda
2024-03-01 10:42:33 +11:00
Lincoln Stein
cc41e8912c
several small model install enhancements
...
- Support extended HF repoid syntax in TUI. This allows
installation of subfolders and safetensors files, as in
`XpucT/Deliberate::Deliberate_v5.safetensors`
- Add `error` and `error_traceback` properties to the install
job objects.
- Rename the `heuristic_import` route to `heuristic_install`.
- Fix the example `config` input in the `heuristic_install` route.
2024-03-01 10:42:33 +11:00
Lincoln Stein
65dd4f4abc
fix repo-id for the Deliberate v5 model
...
prevent lora and embedding file suffixes from being stripped during installation
apply psychedelicious patch to get compel to load proper TI embedding
2024-03-01 10:42:33 +11:00
Lincoln Stein
5bb3aeaccd
remove startup dependency on legacy models.yaml file
2024-03-01 10:42:33 +11:00
psychedelicious
a37b60db13
feat(bulk_download): update response model, messages
2024-03-01 10:42:33 +11:00
psychedelicious
cbb997e7d0
tidy(bulk_download): don't store events service separately
...
Using the invoker object directly leaves no ambiguity as to what `_events_bus` actually is.
2024-03-01 10:42:33 +11:00
psychedelicious
98441ad08d
tidy(bulk_download): do not rely on pagination API to get all images for board
...
We can get all images for the board as a list of image names, then pass that to `_image_handler` to get the DTOs, decoupling from the pagination API.
2024-03-01 10:42:33 +11:00
psychedelicious
80c67dd6e0
tidy(bulk_download): nit - use or as a coalescing operator
...
Just a bit cleaner.
2024-03-01 10:42:33 +11:00
psychedelicious
38af234108
tidy(bulk_download): use single underscore for private attrs
...
Double underscores are used in the app but it doesn't actually do or convey anything that single underscores don't already do. Considered unpythonic except for actual dunder/magic methods.
2024-03-01 10:42:33 +11:00
psychedelicious
2291122c2b
tidy(bulk_download): remove class-level attr annotations
...
These can be misleading as they shadow actual assigned class attributes. This pattern is in the rest of the app but it shouldn't be.
2024-03-01 10:42:33 +11:00
psychedelicious
bf3b10cb1c
tidy(bulk_download): remove extraneous abstract methods
...
`start`, `stop` and `__init__` are not required in implementations of an ABC or service.
2024-03-01 10:42:33 +11:00
psychedelicious
7f8f182a00
tidy(bulk_download): clean up comments
2024-03-01 10:42:33 +11:00
Stefan Tobler
e51867756a
adding bulk_download_item_name to socket events
2024-03-01 10:42:33 +11:00
Stefan Tobler
a8d7cf4e97
refactoring handlers to do null check
2024-03-01 10:42:33 +11:00
Stefan Tobler
037cac8154
removing dependency on an output folder, embrace python temp folder for bulk download
2024-03-01 10:42:33 +11:00
Stefan Tobler
b5a9ed351d
moving the responsibility of cleaning up board names to the service not the route
2024-03-01 10:42:33 +11:00
Stefan Tobler
f15aa562c2
using temp directory for downloads
2024-03-01 10:42:33 +11:00
Stefan Tobler
d0f3571e59
returning the bulk_download_item_name on response for possible polling
2024-03-01 10:42:33 +11:00
Stefan Tobler
b5ca1643a6
narrowing bulk_download stop service scope
2024-03-01 10:42:33 +11:00
Stefan Tobler
79eb871683
cleaning up bulk download zip after the response is complete
2024-03-01 10:42:33 +11:00
Stefan Tobler
7d91426d8f
refactoring bulk_download to be better managed
2024-03-01 10:42:33 +11:00
Stefan Tobler
795fbf0e81
refactoring bulkdownload to consider image category
2024-03-01 10:42:33 +11:00
Stefan Tobler
7114d64b86
fixing issue where default board did not return images
2024-03-01 10:42:33 +11:00
Stefan Tobler
c43ea9f25c
using the board name to download boards
2024-03-01 10:42:33 +11:00
Stefan Tobler
52b0deb179
reworking some of the logic to use a default room, adding endpoint to download file on complete
2024-03-01 10:42:33 +11:00
Stefan Tobler
7ecc18938b
linted and styling
2024-03-01 10:42:33 +11:00
Stefan Tobler
56d2d220a8
implementation of bulkdownload background task
2024-03-01 10:42:33 +11:00
Stefan Tobler
f1967c3393
adding socket events for bulk download
2024-03-01 10:42:33 +11:00
Stefan Tobler
812e24cbd2
groundwork for the bulk_download_service
2024-03-01 10:42:33 +11:00
psychedelicious
89fa36a818
chore(nodes): update TODO comment
2024-03-01 10:42:33 +11:00
psychedelicious
e3f9da29ba
tidy(nodes): clean up profiler/stats in processor, better comments
2024-03-01 10:42:33 +11:00
psychedelicious
763debdeeb
fix(nodes): fix typing on stats service context manager
2024-03-01 10:42:33 +11:00
psychedelicious
8bf9fd34ad
fix(nodes): fix model load events
...
was accessing incorrect properties in event data
2024-03-01 10:42:33 +11:00
psychedelicious
0b0cb0ccc6
feat(nodes): making invocation class var in processor
2024-03-01 10:42:33 +11:00
psychedelicious
fa39523b11
feat(nodes): improved error messages in processor
2024-03-01 10:42:33 +11:00
psychedelicious
16676feea8
feat(nodes): make processor thread limit and polling interval configurable
2024-03-01 10:42:33 +11:00
psychedelicious
d53a2a2d4e
chore(nodes): better comments for invocation context
2024-03-01 10:42:33 +11:00
psychedelicious
ccfe6b6bef
chore(nodes): "context_data" -> "data"
...
Changed within InvocationContext, for brevity.
2024-03-01 10:42:33 +11:00
psychedelicious
fdac0c3c9b
refactor(nodes): move is_canceled to context.util
2024-03-01 10:42:33 +11:00
psychedelicious
18adcc1dd2
feat(nodes): add whole queue_item to InvocationContextData
...
No reason to not have the whole thing in there.
2024-03-01 10:42:33 +11:00
psychedelicious
86c50f2d5b
tidy(nodes): remove extraneous comments
2024-03-01 10:42:33 +11:00
psychedelicious
3cfac8b843
feat(nodes): better invocation error messages
2024-03-01 10:42:33 +11:00
psychedelicious
0788b6ecee
chore(nodes): add comments for cancel state
2024-03-01 10:42:33 +11:00
psychedelicious
317d076a1a
feat(nodes): promote is_canceled to public node API
2024-03-01 10:42:33 +11:00
psychedelicious
725c03cf87
refactor(nodes): merge processors
...
Consolidate graph processing logic into session processor.
With graphs as the unit of work, and the session queue distributing graphs, we no longer need the invocation queue or processor.
Instead, the session processor dequeues the next session and processes it in a simple loop, greatly simplifying the app.
- Remove `graph_execution_manager` service.
- Remove `queue` (invocation queue) service.
- Remove `processor` (invocation processor) service.
- Remove queue-related logic from `Invoker`. It now only starts and stops the services, providing them with access to other services.
- Remove unused `invocation_retrieval_error` and `session_retrieval_error` events, these are no longer needed.
- Clean up stats service now that it is less coupled to the rest of the app.
- Refactor cancellation logic - cancellations now originate from session queue (i.e. HTTP cancel endpoint) and are emitted as events. Processor gets the events and sets the canceled event. Access to this event is provided to the invocation context for e.g. the step callback.
- Remove `sessions` router; it provided access to `graph_executions` but that no longer exists.
2024-03-01 10:42:33 +11:00
psychedelicious
7e71effa17
tidy(nodes): remove no-op model_config
...
Because we now customize the JSON Schema creation for GraphExecutionState, the model_config did nothing.
2024-03-01 10:42:33 +11:00
