Commit Graph

5364 Commits

Author SHA1 Message Date
Lincoln Stein
0a6dccd607 expose max_cache_size to invokeai-configure interface 2023-07-05 20:59:14 -04:00
Lincoln Stein
bf25818d76 rebuild front end; bump version 2023-07-05 20:33:28 -04:00
Lincoln Stein
83d3a043da merge latest changes from main 2023-07-05 19:15:53 -04:00
Lincoln Stein
52498cc0b9
Put tokenizer and text encoder in same clip-vit-large-patch14 (#3662)
This PR fixes the migrate script so that it uses the same directory for
both the tokenizer and text encoder CLIP models. This will fix a crash
that occurred during checkpoint->diffusers conversions

This PR also removes the check for an existing models directory in the
target root directory when `invokeai-migrate3` is run.
2023-07-05 16:29:33 -04:00
Lincoln Stein
cb947bcbf0
Merge branch 'main' into lstein/fix-migrate3-textencoder 2023-07-05 16:23:00 -04:00
Lincoln Stein
bbfb5bb1d4
Remove hardcoded cuda device in model manager init (#3624)
There was a line in model_manager.py in which the GPU device was
hardcoded to "cuda". This has now been removed.
2023-07-05 16:22:45 -04:00
Lincoln Stein
f8bbec8572 Recognize and load diffusers-style LoRAs (.bin)
Prevent double-reporting of autoimported models
- closes #3636

Allow autoimport of diffusers-style LoRA models
- closes #3637
2023-07-05 16:21:23 -04:00
Lincoln Stein
ad5d90aca8 prevent model install crash "torch needs to be restarted with spawn" 2023-07-05 15:38:07 -04:00
Lincoln Stein
9f9ce08e44
Merge branch 'main' into lstein/remove-hardcoded-cuda-device 2023-07-05 13:38:33 -04:00
Lincoln Stein
17c5568661
build: remove web ui dist from gitignore (#3650)
The web UI should manage its own .gitignore

I think would explain why certain files were not making it into the pypi
release
2023-07-05 13:36:16 -04:00
Lincoln Stein
94740e440d
Merge branch 'main' into build/gitignore 2023-07-05 13:35:54 -04:00
Lincoln Stein
5fe722900d allow clip-vit-large-patch14 text encoder to coexist with tokenizer in same directory 2023-07-05 13:15:08 -04:00
Lincoln Stein
cf173b522b allow clip-vit-large-patch14 text encoder to coexist with tokenizer in same directory 2023-07-05 13:14:41 -04:00
Mary Hipp Rogers
ea81ce9489
close modal when user clicks cancel (#3656)
* close modal when user clicks cancel

* close modal when delete image context cleared

---------

Co-authored-by: Mary Hipp <maryhipp@Marys-MacBook-Air.local>
2023-07-05 17:12:27 +00:00
blessedcoolant
8283b80b58
Fix ckpt scanning on conversion (#3653) 2023-07-06 05:09:13 +12:00
blessedcoolant
9e2d63ef97
Merge branch 'main' into fix/ckpt_convert_scan 2023-07-06 05:01:34 +12:00
blessedcoolant
dd946790ec
Fix loading diffusers ti (#3661) 2023-07-06 05:01:11 +12:00
Sergey Borisov
0ac9dca926 Fix loading diffusers ti 2023-07-05 19:46:00 +03:00
psychedelicious
acd3b1a512 build: remove web ui dist from gitignore
The web UI should manage its own .gitignore
2023-07-06 00:39:36 +10:00
Lincoln Stein
bd82c4ace0 model installer confirms deletion of models 2023-07-05 09:57:23 -04:00
Lincoln Stein
9edf78dd2e merge with main 2023-07-05 09:12:54 -04:00
Lincoln Stein
5d31703224 Merge branch 'release/invokeai-3-0-alpha' of github.com:invoke-ai/InvokeAI into release/invokeai-3-0-alpha 2023-07-05 09:05:59 -04:00
Lincoln Stein
44d5bef7e4 bump version number 2023-07-05 09:02:35 -04:00
blessedcoolant
818616a0c5
fix(ui): fix prompt resize & style resizer (#3652) 2023-07-05 23:42:23 +12:00
blessedcoolant
3b324a7d0a
Merge branch 'main' into fix/ui/fix-prompt-resize 2023-07-05 23:40:47 +12:00
blessedcoolant
c8cb43ff2d
Fix clip path in migrate script (#3651)
Update path for clip model according to path used in ckpt conversion and
invokeai-configure
2023-07-05 23:38:45 +12:00
Sergey Borisov
ee042ab76d Fix ckpt scanning on conversion 2023-07-05 14:18:30 +03:00
psychedelicious
596c791844 fix(ui): fix prompt resize & style resizer 2023-07-05 21:02:31 +10:00
blessedcoolant
780e77d2ae
Merge branch 'main' into fix/clip_path 2023-07-05 22:45:52 +12:00
Sergey Borisov
e3fc1b3816 Fix clip path in migrate script 2023-07-05 13:43:09 +03:00
Lincoln Stein
9ad9e91a06
Detect invalid model names when migrating 2.3->3.0 (#3623)
A user discovered that 2.3 models whose symbolic names contain the "/"
character are not imported properly by the `migrate-models-3` script.
This fixes the issue by changing "/" to underscore at import time.
2023-07-05 06:35:54 -04:00
Lincoln Stein
307a01d604 when migrating models, changes / to _ in model names to avoid breaking model name keys 2023-07-05 20:27:03 +10:00
blessedcoolant
0981a7d049
fix(ui): fix dnd on nodes (#3649)
I had broken this earlier today
2023-07-05 21:09:36 +12:00
psychedelicious
2a7dee17be fix(ui): fix dnd on nodes
I had broken this earlier today
2023-07-05 19:06:40 +10:00
blessedcoolant
6c6d600cea
fix(ui): deleting image selects first image (#3648)
@mickr777
2023-07-05 21:00:01 +12:00
blessedcoolant
1c7166d2c6
Merge branch 'main' into fix/ui/delete-image-select 2023-07-05 20:57:34 +12:00
blessedcoolant
07d7959dc0
feat(ui): improve accordion ux (#3647)
- Accordions now may be opened or closed regardless of whether or not
their contents are enabled or active
- Accordions have a short text indicator alerting the user if their
contents are enabled, either a simple `Enabled` or, for accordions like
LoRA or ControlNet, `X Active` if any are active



https://github.com/invoke-ai/InvokeAI/assets/4822129/43db63bd-7ef3-43f2-8dad-59fc7200af2e
2023-07-05 20:57:23 +12:00
psychedelicious
9ebab013c1 fix(ui): deleting image selects first image 2023-07-05 18:21:46 +10:00
psychedelicious
e41e8606b5 feat(ui): improve accordion ux
- Accordions now may be opened or closed regardless of whether or not their contents are enabled or active
- Accordions have a short text indicator alerting the user if their contents are enabled, either a simple `Enabled` or, for accordions like LoRA or ControlNet, `X Active` if any are active
2023-07-05 17:33:03 +10:00
blessedcoolant
6ce867feb4
Fix model detection (#3646) 2023-07-05 19:00:31 +12:00
blessedcoolant
bc8cfc2baa
Merge branch 'main' into fix/model_detect 2023-07-05 18:52:11 +12:00
Eugene Brodsky
7170e82f73 expose max_cache_size in config 2023-07-05 02:44:15 -04:00
Sergey Borisov
2beb8f049e Fix model detection 2023-07-05 09:43:46 +03:00
blessedcoolant
66c10cc2f7
fix: Change Lora weight bounds to -1 to 2 (#3645) 2023-07-05 18:23:06 +12:00
blessedcoolant
1fb317243d fix: Change Lora weight bounds to -1 to 2 2023-07-05 18:12:45 +12:00
blessedcoolant
71310a180d
feat: Add Lora to Canvas (#3643)
- Add Loras to Canvas
- Revert inference_mode to no_grad coz inference tensors fail with
latent to latent.
2023-07-05 17:15:28 +12:00
blessedcoolant
1a29a3fe39 feat: Add Lora to Canvas 2023-07-05 16:39:28 +12:00
blessedcoolant
639d88afd6 revert: inference_mode to no_grad 2023-07-05 16:39:15 +12:00
psychedelicious
f155887b7d fix(ui): change multi image drop to not have selection as payload
This caused a lot of re-rendering whenever the selection changed, which caused a huge performance hit. It also made changing the current image lag a bit.

Instead of providing an array of image names as a multi-select dnd payload, there is now no multi-select dnd payload at all - instead, the payload types are used by the `imageDropped` listener to pull the selection out of redux.

Now, the only big re-renders are when the selectionCount changes. In the future I'll figure out a good way to do image names as payload without incurring re-renders.
2023-07-05 13:25:07 +10:00
psychedelicious
1358c5eb7d fix(ui): fix selector memoization
Every `GalleryImage` was rerendering any time the app rerendered bc the selector function itself was not memoized. This resulted in the memoization cache inside the selector constantly being reset.

Same for `BatchImage`.

Also updated memoization for a few other selectors.
2023-07-05 13:25:07 +10:00