layerdiffusion
4189045323
old backend removal - part 1
2024-08-05 03:15:44 -07:00
layerdiffusion
0863765173
rework sd1.5 and sdxl from scratch
2024-08-05 03:08:17 -07:00
layerdiffusion
0091e82f53
remove old stuff
2024-08-04 18:45:31 -07:00
layerdiffusion
a72154405e
Text Processing Engine is Finished
...
100% reproduce all previous results, including TI embeddings, LoRAs in CLIP, emphasize settings, BREAK, timestep swap scheduling, AB mixture, advanced uncond, etc
Backend is 85% finished
2024-08-04 18:42:51 -07:00
layerdiffusion
430482d1a0
fix cast
2024-08-03 16:08:22 -07:00
layerdiffusion
8c087f920e
rename files
2024-08-03 15:54:39 -07:00
layerdiffusion
a5265df7f4
rename file
2024-08-03 15:51:04 -07:00
layerdiffusion
3ecdcee5a8
add sampler
2024-08-03 15:46:27 -07:00
layerdiffusion
b8b29a2372
remove forge_sampler
2024-08-03 15:38:07 -07:00
layerdiffusion
4add428e25
move to new backend - part 2
2024-08-03 15:10:37 -07:00
layerdiffusion
bb5083f3c2
rework sample function
2024-08-03 13:27:23 -07:00
layerdiffusion
0871870342
Update supported_controlnet.py
2024-08-03 12:52:32 -07:00
layerdiffusion
e722991752
control rework
2024-08-02 22:17:27 -07:00
layerdiffusion
6ed95d66e5
rework several component patcher
...
backend is 65% finished
2024-08-02 14:56:57 -07:00
layerdiffusion
d1b8a2676d
rework lora and patching system
...
and dora etc - backend rework is 60% finished
And I also removed the webui’s extremely annoying lora filter from model versions.
2024-08-02 13:45:26 -07:00
layerdiffusion
f367b07282
fix gradio url timeout
...
#909
2024-08-02 11:46:36 -07:00
layerdiffusion
e5860a4999
move stream file
2024-08-02 03:37:22 -07:00
layerdiffusion
3dd084d55b
more attention types
2024-08-02 03:37:21 -07:00
layerdiffusion
f6981339b0
predictor is a better name
2024-08-01 21:29:34 -07:00
layerdiffusion
bc9977a305
UNet from Scratch
...
Now backend rewrite is about 50% finished.
Estimated finish is in 72 hours.
After that, many newer features will land.
2024-08-01 21:19:41 -07:00
continue revolution
5192e912ab
GroupNorm patcher ( #593 )
...
* add gn wrapper and corresponding patcher
* add gn wrapper and corresponding patcher
2024-08-01 12:53:27 -07:00
layerdiffusion
4d1be42975
Intergrate CLIP
2024-08-01 12:27:20 -07:00
layerdiffusion
0d079a846d
Intergrate Native AutoEncoderKL
2024-07-31 21:10:19 -07:00
layerdiffusion
c8156fcf41
rework model loader and configs
2024-07-30 13:27:26 -06:00
layerdiffusion
40dd61ba6c
improve vae key mapping
2024-07-30 09:23:58 -06:00
layerdiffusion
dc62b0d2d7
begin to use new vae impl
2024-07-30 08:43:14 -06:00
layerdiffusion
37386592ad
version
2024-07-28 08:08:28 -06:00
layerdiffusion
918c6af162
upgrade js canvas for #881
2024-07-28 08:07:37 -06:00
layerdiffusion
2f87f123b2
less aggressive clip skip
...
to make CivitAI pony image meta works better
2024-07-27 13:23:39 -07:00
layerdiffusion
6eb23c6d09
fix #861
2024-07-27 07:52:20 -07:00
layerdiffusion
7e31721803
fix #861
2024-07-27 07:50:59 -07:00
layerdiffusion
3ba632b64b
version
2024-07-27 06:54:05 -07:00
layerdiffusion
0ba9f7f399
Clip Skip to All Models
...
Technically Correct Implementation of CLIP skip for all models.
Automatic1111 uses a weird logic to ignore SDXL clip skip but added an option to only set CLIP-L Skip with an offset 1, which means if one set skip as 3, then the SDXL CLIP G does not skip but SDXL CLIP L uses 3-1=2 as skip (not 3?!)
Forge now uses technically correct and consistent clip skip for all models. Forge now outweigh technical correctness over reproducing Auto results.
But one can get same results by just do not set clip skip for SDXL.
2024-07-27 05:46:53 -07:00
layerdiffusion
1673a5ac2d
Perturbed Attention Guidance Integrated
2024-07-26 15:00:31 -07:00
layerdiffusion
c2bfa6f6a3
fix gradio
2024-07-26 13:08:15 -07:00
layerdiffusion
e1082b8799
fix loading order
2024-07-26 12:57:14 -07:00
layerdiffusion
e26abf87ec
Gradio 4 + WebUI 1.10
2024-07-26 12:02:46 -07:00
lllyasviel
ef35383b4a
VAE patcher and more types of unet patches
2024-02-29 22:37:34 -08:00
lllyasviel
5166a723c2
bring back tiling
...
close #215
2024-02-25 21:29:35 -08:00
lllyasviel
6287c73d98
rework memory management for extras
...
now face post-processing uses gpu
close #312
2024-02-25 20:37:14 -08:00
lllyasviel
437c348926
Add build-in extension "NeverOOM"
...
see also discussions
2024-02-24 19:09:06 -08:00
lllyasviel
434ca2169f
Add optimization --cuda-stream
...
See also the readme for more details
2024-02-24 14:00:48 -08:00
lllyasviel
a844834193
safer stream initialization
2024-02-23 20:28:27 -08:00
lllyasviel
88f395091b
add two optimizations
...
--pin-shared-memory and --cuda-malloc
See also the updates in Readme for more details
2024-02-23 18:39:32 -08:00
lllyasviel
539bc5035d
safe cleanup to avoid potential problems
2024-02-22 01:28:38 -08:00
lllyasviel
638ee43bf1
Merge upstream PR 14855
2024-02-21 23:59:40 -08:00
lllyasviel
4ce60f2176
support SDXL-Lightning
...
added samplers: Euler SGMUniform, Euler A SGMUniform, DPM++ 2M SGMUniform, DPM++ 2M SDE SGMUniform
ByteDance official recommended one is Euler SGMUniform
Remember to use CFG=1.0
2024-02-21 15:42:38 -08:00
lllyasviel
43c9e3b5ce
begin to use double versioning
...
now you will see things like f0.0.14v1.8.0rc. It means forge 0.0.14 on webui 1.8.0rc
2024-02-17 08:32:15 -08:00
lllyasviel
07659efd88
fix import order
2024-02-17 08:03:17 -08:00
lllyasviel
d81e353d89
fix outpaint in inpaint_only+lama
...
the problem in inpaint_only+lama+"Resize and Fill" is fixed now.
SDXL+inpaint_only+lama+"Resize and Fill"+inpaint_fooocus_v26+"stop at 0.5 or 0.6" can be used as one SOTA outpaint method now
2024-02-14 02:45:23 -08:00
