Commit Graph

3630 Commits

Author SHA1 Message Date
AUTOMATIC1111
c1a31ec9f7 revert to applying mask before denoising for k-diffusion, like it was before 2023-08-14 08:59:15 +03:00
AUTOMATIC1111
cda2f0a162 make on_before_component/on_after_component possible earlier 2023-08-14 08:49:39 +03:00
AUTOMATIC1111
aeb76ef174 repair DDIM/PLMS/UniPC batches 2023-08-14 08:49:02 +03:00
AUTOMATIC1111
007ecfbb29 also use setup callback for the refiner instead of before_process 2023-08-13 21:01:13 +03:00
AUTOMATIC1111
9cd0475c08
Merge pull request #12526 from brkirch/mps-adjust-sub-quad
Fixes for `git checkout`, MPS/macOS fixes and optimizations
2023-08-13 20:28:49 +03:00
AUTOMATIC1111
8452708560
Merge pull request #12530 from eltociear/eltociear-patch-1
Fix typo in launch_utils.py
2023-08-13 20:27:17 +03:00
AUTOMATIC1111
16781ba09a fix 2 for git code botched by previous PRs 2023-08-13 20:15:20 +03:00
Ikko Eltociear Ashimine
09ff5b5416
Fix typo in launch_utils.py
existance -> existence
2023-08-14 01:03:49 +09:00
AUTOMATIC1111
f093c9d39d fix broken XYZ plot seeds
add new callback for scripts to be used before processing
2023-08-13 17:31:10 +03:00
brkirch
2035cbbd5d Fix DDIM and PLMS samplers on MPS 2023-08-13 10:07:52 -04:00
brkirch
5df535b7c2 Remove duplicate code for torchsde randn 2023-08-13 10:07:52 -04:00
brkirch
f4dbb0c820 Change the repositories origin URLs when necessary 2023-08-13 10:07:52 -04:00
brkirch
9058620cec git checkout with commit hash 2023-08-13 10:07:14 -04:00
brkirch
2489252099 torch.empty can create issues; use torch.zeros
For MPS, using a tensor created with `torch.empty()` can cause `torch.baddbmm()` to include NaNs in the tensor it returns, even though `beta=0`. However, with a tensor of shape [1,1,1], there should be a negligible performance difference between `torch.empty()` and `torch.zeros()` anyway, so it's better to just use `torch.zeros()` for this and avoid unnecessarily creating issues.
2023-08-13 10:06:25 -04:00
brkirch
87dd685224 Make sub-quadratic the default for MPS 2023-08-13 10:06:25 -04:00
brkirch
abfa4ad8bc Use fixed size for sub-quadratic chunking on MPS
Even if this causes chunks to be much smaller, performance isn't significantly impacted. This will usually reduce memory usage but should also help with poor performance when free memory is low.
2023-08-13 10:06:25 -04:00
AUTOMATIC1111
3163d1269a fix for the broken run_git calls 2023-08-13 16:51:21 +03:00
AUTOMATIC1111
1c6ca09992
Merge pull request #12510 from catboxanon/feat/extnet/hashes
Support search and display of hashes for all extra network items
2023-08-13 16:46:32 +03:00
AUTOMATIC1111
d73db17ee3
Merge pull request #12515 from catboxanon/fix/gc1
Clear sampler and garbage collect before decoding images to reduce VRAM
2023-08-13 16:45:38 +03:00
AUTOMATIC1111
127ab9114f
Merge pull request #12514 from catboxanon/feat/batch-encode
Encode batch items individually to significantly reduce VRAM
2023-08-13 16:41:07 +03:00
AUTOMATIC1111
d53f3b5596
Merge pull request #12520 from catboxanon/eta
Update description of eta setting
2023-08-13 16:40:17 +03:00
AUTOMATIC1111
d41a5bb97d
Merge pull request #12521 from catboxanon/feat/more-s-noise
Add `s_noise` param to more samplers
2023-08-13 16:39:25 +03:00
catboxanon
525b55b1e9 Restore extra_params that was lost in merge 2023-08-13 09:08:34 -04:00
catboxanon
ce0829d711 Merge branch 'feat/dpmpp3msde' into feat/more-s-noise 2023-08-13 08:46:58 -04:00
catboxanon
ac790fc49b Discard penultimate sigma for DPM-Solver++(3M) SDE 2023-08-13 08:46:07 -04:00
catboxanon
f4757032e7 Fix s_noise description 2023-08-13 08:24:28 -04:00
catboxanon
d1a70c3f05 Add s_noise param to more samplers 2023-08-13 08:22:24 -04:00
AUTOMATIC1111
d8419762c1 Lora: output warnings in UI rather than fail for unfitting loras; switch to logging for error output in console 2023-08-13 15:07:37 +03:00
catboxanon
60a7405165 Update description of eta setting 2023-08-13 08:06:40 -04:00
catboxanon
1ae9dacb4b Add DPM-Solver++(3M) SDE 2023-08-13 07:57:29 -04:00
catboxanon
69f49c8d39 Clear sampler before decoding images
More significant VRAM reduction.
2023-08-13 04:40:34 -04:00
catboxanon
822597db49 Encode batches separately
Significantly reduces VRAM.
This makes encoding more inline with how decoding currently functions.
2023-08-13 04:16:48 -04:00
catboxanon
7fa5ee54b1 Support search and display of hashes for all extra network items 2023-08-13 02:32:54 -04:00
AUTOMATIC1111
599f61a1e0 use dataclass for StableDiffusionProcessing 2023-08-13 08:24:16 +03:00
AUTOMATIC1111
fa9370b741 add refiner to StableDiffusionProcessing class
write out correct model name in infotext, rather than the refiner model
2023-08-13 06:07:30 +03:00
AUTOMATIC1111
b2080756fc make "send to" buttons into small tool buttons 2023-08-12 19:03:33 +03:00
AUTOMATIC1111
6816ad5ed8 fix broken reuse seed 2023-08-12 18:36:30 +03:00
AUTOMATIC1111
4e8690906c update seed/subseed HTML widths 2023-08-12 18:00:30 +03:00
AUTOMATIC1111
f0b72b8121 move seed, variation seed and variation seed strength to a single row, dump resize seed from UI
add a way for scripts to register a callback for before/after just a single component's creation
2023-08-12 17:46:13 +03:00
AUTOMATIC1111
b293ed3061 make it possible to use hires fix together with refiner 2023-08-12 12:54:32 +03:00
AUTOMATIC1111
64311faa68 put refiner into main UI, into the new accordions section
add VAE from main model into infotext, not from refiner model
option to make scripts UI without gr.Group
fix inconsistencies with refiner when usings samplers that do more denoising than steps
2023-08-12 12:39:59 +03:00
AUTOMATIC1111
26c92f056a
Merge pull request #12480 from catboxanon/fix/cc
Fix color correction by converting image to RGB
2023-08-12 09:12:30 +03:00
AUTOMATIC1111
ebc1bafb03
Merge pull request #12479 from catboxanon/fix/extras-generator
Refactor postprocessing/extras tab to use generator to resolve OOM issues
2023-08-12 08:58:14 +03:00
w-e-w
f57bc1a21b disable extensions installer with arg 2023-08-12 12:06:31 +09:00
catboxanon
af27b716e5 Fix color correction by converting image to RGB 2023-08-11 12:22:11 -04:00
catboxanon
7c9c19b2a2 Refactor postprocessing to use generator to resolve OOM issues 2023-08-11 11:32:12 -04:00
AUTOMATIC1111
ae6b30907d
Merge pull request #12470 from Splendide-Imaginarius/mask-blur-property+kernel
Make `StableDiffusionProcessingImg2Img.mask_blur` a property, make more inline with PIL `GaussianBlur`
2023-08-11 15:03:18 +03:00
catboxanon
d456fb797a fix: Properly return None when VAE hash is None 2023-08-10 16:04:49 -04:00
AUTOMATIC1111
458eda1321
Merge pull request #12456 from AUTOMATIC1111/patch-#12453
Patch #12453
2023-08-10 17:55:31 +03:00
w-e-w
a75d756a6f use default value if value error 2023-08-10 23:47:28 +09:00