ac02216e54
alternate implementation for unet forward replacement that does not depend on hijack being applied
2023-12-02 19:35:47 +03:00
a5f61aa8c5
potential fix for #14172
2023-12-02 18:03:34 +03:00
293f44e6c1
Fix bug where is_using_v_parameterization_for_sd2 fails because the sd_hijack is only partially undone
2023-12-01 22:56:08 -05:00
9c1c0da026
fix exception related to the pix2pix
2023-11-06 11:17:36 +03:00
6ad666e479
more changes for #13865 : fix formatting, rename the function, add comment and add a readme entry
2023-11-05 19:46:20 +03:00
80d639a440
linter
2023-11-05 19:32:21 +03:00
ff805d8d0e
Merge branch 'dev' into master
2023-11-05 19:30:57 +03:00
44c5097375
Use devices.torch_gc() instead of empty_cache()
2023-11-05 20:31:57 +05:30
44db35fb1a
Added memory clearance after deletion
2023-11-05 19:15:38 +05:30
ff1609f91e
Add SSD-1B as a supported model
2023-11-05 19:13:49 +05:30
902afa6b4c
Merge pull request #13364 from superhero-7/master
...
Add altdiffusion-m18 support
2023-10-14 07:29:01 +03:00
2d947175b9
fix linter issues
2023-10-01 12:25:19 +08:00
f8f4ff2bb8
support altdiffusion-m18
2023-09-23 17:55:19 +08:00
702a1e1cc7
support m18
2023-09-23 17:51:41 +08:00
59544321aa
initial work on sd_unet for SDXL
2023-09-11 21:17:40 +03:00
9d2299ed0b
implement undo hijack for SDXL
2023-08-19 10:16:27 +03:00
a8a256f9b5
REMOVE
2023-08-08 21:08:50 +03:00
22ecb78b51
Merge branch 'dev' into multiple_loaded_models
2023-08-05 07:52:29 +03:00
f0c1063a70
resolve some of circular import issues for kohaku
2023-08-04 09:13:46 +03:00
151b8ed3a6
repair PLMS
2023-08-01 00:38:34 +03:00
b235022c61
option to keep multiple models in memory
2023-08-01 00:24:48 +03:00
6f0abbb71a
textual inversion support for SDXL
2023-07-29 15:15:06 +03:00
5677296d1b
Merge pull request #11878 from Bourne-M/patch-1
...
【bug】reload altclip model error
2023-07-19 16:26:12 +03:00
cb75734896
【bug】reload altclip model error
...
When using BertSeriesModelWithTransformation as the cond_stage_model, the undo_hijack should be performed using the FrozenXLMREmbedderWithCustomWords type; otherwise, it will result in a failed model reload.
2023-07-19 17:53:28 +08:00
0198eaec45
Merge pull request #11757 from AUTOMATIC1111/sdxl
...
SD XL support
2023-07-16 12:04:53 +03:00
2b1bae0d75
add textual inversion hashes to infotext
2023-07-15 08:41:22 +03:00
6d8dcdefa0
initial SDXL refiner support
2023-07-14 09:16:01 +03:00
594c8e7b26
fix CLIP doing the unneeded normalization
...
revert SD2.1 back to use the original repo
add SDXL's force_zero_embeddings to negative prompt
2023-07-13 11:35:52 +03:00
da464a3fb3
SDXL support
2023-07-12 23:52:43 +03:00
af081211ee
getting SD2.1 to run on SDXL repo
2023-07-11 21:16:43 +03:00
36888092af
revert default cross attention optimization to Doggettx
...
make --disable-opt-split-attention command line option work again
2023-06-01 08:12:06 +03:00
339b531570
custom unet support
2023-05-27 15:47:33 +03:00
a6e653be26
possible fix for empty list of optimizations #10605
2023-05-23 18:49:15 +03:00
2140bd1c10
make it actually work after suggestions
2023-05-19 10:05:07 +03:00
8a3d232839
fix linter issues
2023-05-19 00:03:27 +03:00
2582a0fd3b
make it possible for scripts to add cross attention optimizations
...
add UI selection for cross attention optimization
2023-05-18 22:48:28 +03:00
1a43524018
fix model loading twice in some situations
2023-05-14 13:27:50 +03:00
49a55b410b
Autofix Ruff W (not W605) (mostly whitespace)
2023-05-11 20:29:11 +03:00
028d3f6425
ruff auto fixes
2023-05-10 11:05:02 +03:00
f741a98bac
imports cleanup for ruff
2023-05-10 08:43:42 +03:00
762265eab5
autofixes from ruff
2023-05-10 07:52:45 +03:00
8d7fa2f67c
sdp_attnblock_forward hijack
2023-03-10 22:48:41 +05:00
0981dea948
sdp refactoring
2023-03-10 12:58:10 +05:00
37acba2633
argument to disable memory efficient for sdp
2023-03-10 12:19:36 +05:00
fec0a89511
scaled dot product attention
2023-03-07 00:33:13 +05:00
dfb3b8f398
Merge branch 'master' into weighted-learning
2023-02-19 12:41:29 +03:00
c4bfd20f31
Hijack to add weighted_forward to model: return loss * weight map
2023-02-15 10:03:59 +01:00
2016733814
Apply hijacks in ddpm_edit for upcast sampling
...
To avoid import errors, ddpm_edit hijacks are done after an instruct pix2pix model is loaded.
2023-02-07 22:53:45 -05:00
fecb990deb
Merge pull request #7309 from brkirch/fix-embeddings
...
Fix embeddings, upscalers, and refactor `--upcast-sampling`
2023-01-28 18:44:36 +03:00
d04e3e921e
automatically detect v-parameterization for SD2 checkpoints
2023-01-28 15:24:41 +03:00
AUTOMATIC1111