Commit Graph

3489 Commits

Author SHA1 Message Date
34ea49147c Update the check_models.py script during startup 2023-04-19 16:13:29 +05:30
c1e8637a9f Re-implement the code for downloading models in python. Save some eyeballs from bleeding 2023-04-19 16:11:16 +05:30
becbef4fac Include ROCm in the list of allowed versions 2023-04-18 17:36:52 +05:30
f22ecc454a Merge branch 'beta' of github.com:cmdr2/stable-diffusion-ui into beta 2023-04-18 17:14:44 +05:30
bf3df097b8 Don't use ROCm on Linux if an NVIDIA card is present 2023-04-18 17:14:24 +05:30
7fc2ed28b1 Merge pull request #1166 from JeLuF/not_yet
Don't save model_path if initial load fails
2023-04-18 16:34:12 +05:30
30a133bad9 Allow torch 1.11 to continue being installed 2023-04-18 16:10:46 +05:30
d8d44c579c Typo 2023-04-18 15:43:56 +05:30
80384e6ee1 Install PyTorch 2.0 by default, but allow existing PyTorch 1.13.1 installations to continue running; Unify and streamline the installation of dependencies 2023-04-18 15:42:33 +05:30
0898f98355 Merge branch 'beta' of github.com:cmdr2/stable-diffusion-ui into beta 2023-04-18 15:05:53 +05:30
e7dc41e271 Automatic AMD GPU detection on Linux (#1078)
* Automatic AMD GPU detection on Linux

Automatically detects AMD GPUs and installs the ROCm version of PyTorch instead of the cuda one

A later improvement may be to detect the GPU ROCm version and handle GPUs that dont work on upstream ROCm, ether because they're too old and need a special patched version, or too new and need `HSA_OVERRIDE_GFX_VERSION=10.3.0` added, possibly check through `rocminfo`?

* Address stdout suppression and download failure

* If any NVIDIA GPU is found, always use it

* Use /proc/bus/pci/devices to detect GPUs

* Fix comparisons

`-eq` and `-ne` only work for numbers

* Add back -q

---------

Co-authored-by: JeLuF <jf@mormo.org>
2023-04-18 15:02:39 +05:30
0f0f475241 Merge branch 'beta' of github.com:cmdr2/stable-diffusion-ui into beta 2023-04-18 14:44:23 +05:30
127ee68486 Merge pull request #1171 from JeLuF/p0417
Don't download 1.4 if other models are available
2023-04-18 14:43:45 +05:30
b204b02b05 Merge pull request #1173 from JeLuF/patch-20
Add "Start scanning..." to getModels()
2023-04-18 14:41:09 +05:30
893b6d985c Add "Start scanning..." to getModels()
Provide a hint to users what ED is currently using. 
Use case: User has built an infinite loop using symlinks, ED model scan will never finish.

https://discord.com/channels/1014774730907209781/1097764777553580092
2023-04-18 09:12:26 +02:00
44824fb5f9 Don't download 1.4 if other models are available 2023-04-17 23:22:44 +02:00
dc21cbe59d Typo 2023-04-17 16:25:49 +05:30
e16d9f4742 Merge branch 'torch2' into beta 2023-04-17 16:24:41 +05:30
f2b5843e6c merge beta 2023-04-17 15:50:51 +05:30
16229caa8e Merge branch 'beta' of github.com:cmdr2/stable-diffusion-ui into beta 2023-04-17 15:43:18 +05:30
0c0525e11b sdkit 1.0.72 - use the extra attn precision yaml code for diffusers, which doesn't auto-detect black images yet 2023-04-17 15:43:08 +05:30
11517b0969 Merge pull request #1168 from cmdr2/beta
Tiled upscaling for better VRAM usage, auto-detect black images and try fp32 attention precision, and full precision if that fails too
2023-04-17 15:11:41 +05:30
1ba3a139d9 Don't save model_path if initial load fails
Fixes #882

If the load of the model fails during the initialization, an attempt to render an
image using the same model fails because ED doesn't notice that the model has to
be loaded. This PR ensures that the model is being reloaded if the initial load
fails. If the second load attempt fails as well, the user will get a more helpful
error message than 'model not loaded yet'.
2023-04-16 21:59:29 +02:00
80bcfabc48 Upgrade to PyTorch 2.0; Doesn't use a special repo url for pytorch on Linux 2023-04-14 17:32:27 +05:30
4192f87d6b Don't scan safetensors files in load_default_models() (#1155)
* Don't scan safetensors when loading them

* Don't scan safetensors files

* Update model_manager.py

---------

Co-authored-by: cmdr2 <shashank.shekhar.global@gmail.com>
2023-04-14 17:11:35 +05:30
03c8a0fca5 sdkit 1.0.70 - use plms for warming up the model, avoiding any non-deterministic effects from the default ancestral sampler 2023-04-12 15:31:10 +05:30
a3d2c71ed6 sdkit 1.0.69 - allow loading models without vae weights in diffusers 2023-04-11 15:34:44 +05:30
8ba0b34853 Log the stack trace when the model fails to load 2023-04-11 15:13:41 +05:30
424ec40fa5 sdkit 1.0.68 - fix brainfade 2023-04-11 12:00:03 +05:30
210429a259 sdkit 1.0.67 - detect black-images on model load and use fp32 attention precision or full precision if needed 2023-04-10 17:59:14 +05:30
df806d5dfa changelog 2023-04-10 16:31:58 +05:30
b07046f6a2 gc between rendering images and applying filters 2023-04-10 16:31:24 +05:30
4cdb8a7d2a sdkit 1.0.66 - lower VRAM usage for realesrgan upscaling 2023-04-10 16:29:30 +05:30
2f83f2bd48 Merge branch 'beta' of github.com:cmdr2/stable-diffusion-ui into beta 2023-04-10 15:56:07 +05:30
8a6cf3cfae Remove redundant hotfix 2023-04-10 15:55:55 +05:30
514e40569e Merge pull request #1138 from patriceac/patch-65
Properly reset LoRA selection from Use Settings
2023-04-10 10:54:12 +05:30
f30f98abd8 Merge pull request #1143 from patriceac/patch-67
Fix the tooltip display for long modifiers
2023-04-10 10:53:46 +05:30
c611d26306 Merge pull request #1144 from patriceac/patch-68
Fix the styling of disabled image modifiers
2023-04-10 10:53:06 +05:30
9ee38d0b70 Fix the styling of disabled image modifiers
Long custom modifiers in a disabled state (e.g. right-click on the image tag) are properly restored as disabled but are incorrectly shown as "active" in the UI. I somehow missed that in https://github.com/cmdr2/stable-diffusion-ui/pull/1062, so here is the fix for that. This change only applies to plugins that try to restore image modifiers, no change to the regular UI.
2023-04-09 18:38:36 -07:00
5e45f37232 Fix the tooltip display for long modifiers 2023-04-09 15:33:05 -07:00
4f2df2d188 Properly reset LoRA selection from Use Settings 2023-04-08 18:19:03 -07:00
ae470e35c8 Merge pull request #1137 from cmdr2/beta
Beta
2023-04-08 20:16:31 +05:30
0f4b62cb97 Hotfix - apply the config overrides to the Settings UI *after* the default config-apply function, not before it 2023-04-08 20:13:44 +05:30
c086098af1 Update README.md 2023-04-08 11:11:25 +05:30
4c7b4c7592 Merge pull request #1135 from JeLuF/ctrlv
Don't paste empty prompts
2023-04-08 09:01:07 +05:30
9e244f758c Don't paste empty prompts
When pasting e.g. an image, window.clipboardData).getData('text') returns an empty string, which would delete the prompt.
https://discord.com/channels/1014774730907209781/1093186485563424838
2023-04-07 22:24:35 +02:00
e7c0b9bd76 Merge branch 'beta' of github.com:cmdr2/stable-diffusion-ui into beta 2023-04-07 15:52:02 +05:30
d7317e8252 sdkit 1.0.65 - upgrade compel to 1.0.5 2023-04-07 15:51:44 +05:30
36a187d3c5 Merge pull request #1133 from cmdr2/beta
Beta
2023-04-07 15:12:53 +05:30
a6ec401440 Merge pull request #1130 from JeLuF/callback
fix filename_format for 'show only upscaled'==false
2023-04-07 15:12:33 +05:30