d1d0ec46aa
Maintain patching related
...
1. fix several problems related to layerdiffuse not unloaded
2. fix several problems related to Fooocus inpaint
3. Slightly speed up on-the-fly LoRAs by precomputing them to computation dtype
2024-08-30 15:18:21 -07:00
df598c4df5
add "Diffusion in Low Bits" to info text
2024-08-22 23:23:14 -07:00
08f7487590
revise "send distilled_cfg_scale when generating hires conds"
2024-08-22 06:35:13 -07:00
bfd7fb1d9f
send distilled_cfg_scale when generating hires conds ( #1403 )
2024-08-22 06:33:48 -07:00
14ac95f908
fix
2024-08-20 01:37:01 -07:00
d38e560e42
Implement some rethinking about LoRA system
...
1. Add an option to allow users to use UNet in fp8/gguf but lora in fp16.
2. All FP16 loras do not need patch. Others will only patch again when lora weight change.
3. FP8 unet + fp16 lora are available (somewhat only available) in Forge now. This also solves some “LoRA too subtle” problems.
4. Significantly speed up all gguf models (in Async mode) by using independent thread (CUDA stream) to compute and dequant at the same time, even when low-bit weights are already on GPU.
5. View “online lora” as a module similar to ControlLoRA so that it is moved to GPU together with model when sampling, achieving significant speedup and perfect low VRAM management simultaneously.
2024-08-19 04:31:59 -07:00
93bfd7f85b
invalidate cond cache if distilled CFG changed ( #1240 )
...
* Update processing.py
add distilled_cfg_scale to params that invalidate cond cache
* Update ui.py
distilled CFG and CFG step size 0.1 (from 0.5)
2024-08-17 19:34:11 -07:00
0f266c48bd
make clear_prompt_cache a function
2024-08-17 19:00:13 -07:00
61f83dd610
support all flux models
2024-08-13 05:42:17 -07:00
19b41b9438
Add option to experiment with results from other impl
...
Setting -> Compatibility -> Try to reproduce the results from external software
2024-08-11 17:02:50 -07:00
cfa5242a75
forge 2.0.0
...
see also discussions
2024-08-10 19:24:19 -07:00
593455c4de
global unload after env var change
2024-08-08 22:34:55 -07:00
20e1ba4a82
fix
2024-08-08 15:08:20 -07:00
6921420b3f
Load Model only when click Generate
...
#964
2024-08-08 14:51:13 -07:00
a189f3e53e
distilled cfg scale info text
2024-08-08 01:22:27 -07:00
572a3c3d8b
distilled cfg scale
2024-08-08 00:57:40 -07:00
653be0a3ad
comments
2024-08-08 00:20:15 -07:00
396d9f378d
we do not need to waste 10 seconds on T5 when CFG=1
2024-08-08 00:19:16 -07:00
a6baf4a4b5
revise kernel
...
and add unused files
2024-08-07 16:51:24 -07:00
71eaa5ca12
rework UI so that the toolbar is managed by Forge
2024-08-06 13:54:06 -07:00
ae1d995d0d
Finally removed model_hijack
...
finally
2024-08-05 21:05:25 -07:00
27641043ee
remove dirty cond_stage_key
2024-08-05 11:39:20 -07:00
d77582aa5a
fix img2img inpaint model
2024-08-05 11:36:08 -07:00
bccf9fb23a
Free WebUI from its Prison
...
Congratulations WebUI. Say Hello to freedom.
2024-08-05 04:21:35 -07:00
0863765173
rework sd1.5 and sdxl from scratch
2024-08-05 03:08:17 -07:00
9679232d29
fix TI embedding info text
...
original webui is also broken but forge is fixed now
2024-08-04 18:49:35 -07:00
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
8c087f920e
rename files
2024-08-03 15:54:39 -07:00
f6a0c69f0f
fix ADetailer
2024-07-26 15:11:12 -07:00
e26abf87ec
Gradio 4 + WebUI 1.10
2024-07-26 12:02:46 -07:00
10b5ca2541
vae already sliced in inner loop
2024-03-08 00:40:33 -08:00
b9705c58f6
fix alphas cumprod ( #478 )
...
* fix alphas cumprod
* indentation
2024-03-04 01:03:32 -06:00
95bcea72b1
Revert "fix alphas cumprod ( #475 )"
...
This reverts commit 72139b000c
2024-03-03 22:40:56 -08:00
72139b000c
fix alphas cumprod ( #475 )
2024-03-03 20:09:04 -06:00
5166a723c2
bring back tiling
...
close #215
2024-02-25 21:29:35 -08:00
16caff3d14
Revert "try solve #381 "
...
This reverts commit 9d4c88912d
2024-02-23 16:13:40 -08:00
9d4c88912d
try solve #381
...
caused by some potential historical webui problems
2024-02-23 16:01:40 -08:00
bde779a526
apply_token_merging
2024-02-23 15:43:27 -08:00
95ddac3117
Merge upstream
...
upstream
2024-02-21 23:56:45 -08:00
33c8fe1221
avoid doble upscaling in inpaint
2024-02-19 16:57:49 +04:00
30c8d742b3
Merge branch 'main' into upt
2024-02-11 16:42:36 -08:00
e2b19900ec
add infotext entry for emphasis; put emphasis into a separate file, add an option to parse but still ignore emphasis
2024-02-11 09:39:51 +03:00
388ca351f4
Revert "Fix ruff linter ( #137 )" ( #143 )
...
This reverts commit 6b3ad64388
2024-02-08 21:24:04 -05:00
6b3ad64388
Fix ruff linter ( #137 )
...
* Fix ruff linter
* Remove unused imports
* Remove unused imports
2024-02-08 20:35:20 -05:00
760f727eb9
use better context manager to fix potential problems
2024-02-08 01:51:18 -08:00
c11490d560
add_alphas_cumprod_modifier
2024-02-03 22:33:18 -08:00
905a027fc1
Update processing.py
2024-02-02 23:01:03 -08:00
1edd626f7c
Update processing.py
2024-02-02 22:30:55 -08:00
66e563fe03
inpaint done!
2024-01-30 15:13:26 -08:00
d04765cf3d
Update processing.py
2024-01-27 22:47:04 -08:00
layerdiffusion