8553 Commits

Author SHA1 Message Date
layerdiffusion
c5853ef9f8 dep ver 2024-10-28 23:04:13 -07:00
lllyasviel
2d5b6cacef Merge pull request #2183 from graemeniedermayer/sd35_integration
sd3.5 integration (naive)
2024-10-28 21:54:10 -07:00
layerdiffusion
623778a3a3 dep ver 2024-10-28 21:50:29 -07:00
licyk
b592142f3b fix multidiffusion integrated with controlnet (#2199) 2024-10-28 12:15:37 +00:00
psydok
878c056576 Any exception is translated in http response (#2186)
* fix: get error in http request
2024-10-28 00:27:38 +00:00
Symbiomatrix
41a21f66fd Restore embedding filepath. (#2030) 2024-10-26 23:53:48 +01:00
DenOfEquity
70f3499d0f avoid potential unnecessary model load with hiresfix quickbutton (#2189) 2024-10-26 21:48:36 +01:00
altoiddealer
145a46907e Improve option handling (sd_model_checkpoint / forge_additional_modules) (#2181)
* Sort modules when checking for changes
* Compare consistent checkpoint values
2024-10-26 21:10:05 +01:00
grae
1363999fb1 sd3.5 integration 2024-10-25 18:39:45 -06:00
DenOfEquity
d4d8ad406e save (Hires) Distilled CFG to infotext (#2172)
changing hires model could lead to not writing
2024-10-24 20:37:34 +01:00
catboxanon
edeb2b883f Fix loading diffusers format VAEs (#2171) 2024-10-24 14:27:57 -04:00
DenOfEquity
9a698e26d6 Properly load/save infotext for empty selection of hrfix additional modules (#2156)
empty selection for *Hires VAE / TE* means 'use modules built-in to the model file', same as top-row VAE / TE selection. But this must be written to infotext as something, to distinguish from the non-HiRes case.
2024-10-23 11:08:45 +01:00
DenOfEquity
0c97a5347a properly load/save 'Use same choices' infotext (#2147)
* save 'Use same choices' to infotext, and load
* reload models as needed for batch iterations
2024-10-22 16:34:40 +01:00
Kazancev Danil
518e3cdbd5 fix txt2img pydantic error (#2138)
Co-authored-by: kazanplova <kazanplova@mt-dev-gpu-hotel-01.novalocal>
2024-10-22 15:31:58 +01:00
DenOfEquity
e4ad1140c8 handle before_process scripts earlier (#2132)
this allows the built-in extension 'extra options section' to override correctly
2024-10-21 11:07:48 +01:00
DenOfEquity
aaa2fe761b Hires additional modules (#2116)
adds selection of none/same/different modules for hiresfix
('Use same choices' default option has priority over other selections made at same time.)
includes saving/loading from infotext
2024-10-21 11:03:12 +01:00
catboxanon
edc46380cc Automatically enable ztSNR when applicable (#2122) 2024-10-19 20:33:34 -04:00
catboxanon
a27a0c6936 Merge pull request #2119 from lllyasviel/fix/ztsnr 2024-10-19 12:29:21 -04:00
catboxanon
aba35cde5f Fix Zero Terminal SNR option 2024-10-19 12:28:14 -04:00
catboxanon
534405e597 Merge pull request #2117 from lllyasviel/fix/prediction-detect 2024-10-19 07:48:45 -04:00
catboxanon
f620f55e56 Fallback to already detected prediction type if none applicable found 2024-10-19 07:47:56 -04:00
catboxanon
0511318eff Merge pull request #2115 from lllyasviel/fix/prediction-type-mk2 2024-10-19 07:42:19 -04:00
catboxanon
5ec47a6b93 Fix model prediction detection
Closes #1109
2024-10-19 06:18:02 -04:00
DenOfEquity
1fae20d94f Clipskip override from infotext (#2099)
* support more infotext overrides
clip skip, model, and vae/te, and old VAE too
2024-10-18 16:47:49 +01:00
DenOfEquity
1ef5436833 add .avif to extensions for img2img batch (#2102) 2024-10-17 20:23:42 +01:00
DenOfEquity
a06cfc9634 Hiresfix checkpoint (#2100)
restores checkpoint selection for highres fix
2024-10-17 13:39:42 +01:00
altoiddealer
6dc71b7e1d prevent undoing refresh model load params (#2092)
Ensures `should_refresh_model_loading_params()` is called when needed. Improved code clarity.
2024-10-16 14:52:35 +01:00
DenOfEquity
9efa4eabfd prevent (Managed by Forge) settings from reverting to default (#2090)
* prevent (Managed by Forge) settings reverting to default
* save CLIP skip to config when changed, inconsistent / unpredictable saving behaviour otherwise
2024-10-16 14:10:26 +01:00
altoiddealer
2c543719e3 Improve options management (#2078)
- `/sdapi/v1/options` GET now calls `get_config()` from **sysinfo** module, instead of from its own version of the function.

- Defined a new, flexible and more robust `set_config()` function in **sysinfo** module, which:
  - obsoletes redundant code
  - skips updating values that are unchanged
  - has flexible args for both API and UI use

- `/sdapi/v1/options` POST and `override_settings` now use the new `set_config()` function.  `set_config()` could possibly obsolete additional functions, but I'm not going to get into that just yet.

- Options for `forge_additional_modules` can now be provided either as the file path, or just the module name.

- Most importantly, `refresh_model_loading_parameters()` is now only called ONCE per request, and **only** if necessary.

- It is now much easier to call `shared.opts.save()` as needed
2024-10-16 11:21:54 +01:00
altoiddealer
cce30d3340 Include Extra Images via API (#2066) 2024-10-15 13:52:08 +01:00
DenOfEquity
b4d2266dc6 Checkpoint vae+te metadata should be in new attribute (#2073)
* new metadata in new attribute
maintain compatibility with old webui by not reusing the vae metadata attribute
2024-10-15 13:40:15 +01:00
altoiddealer
77464215c3 Make SD Models work with Override Settings (#2054)
* Do refresh load params for modules
* Adjust call order for model mgmt/prompt cache
* new function `manage_model_and_prompt_cache()` to improve code clarity
2024-10-14 11:47:39 +01:00
altoiddealer
862c7a589e API Improvements: Modules Change AND Restore override_settings (#2027)
* Improve API modules change
* Restore override_settings and make it work
* Simplify some memory management
2024-10-13 12:29:02 +01:00
DenOfEquity
ae8187bf2d Unload all models control (#2038)
* make unload_model_weights do that
* rename Settings > Actions > unload checkpoint button to 'Unload all models'
* remove (comment out) reload button, as it does nothing and is unlikely to ever do anything since models are loaded on demand
2024-10-12 12:51:26 +01:00
DenOfEquity
82eb756617 checks for GPU weights from Settings (#2019)
for the case when GPU has changed and saved settings are not possible anymore
2024-10-09 22:10:06 +01:00
DenOfEquity
7eb824a0da fix for control images (#2011) 2024-10-08 13:37:46 +01:00
DenOfEquity
4b695514de preserve infotext with Extras > Single Image (#2006)
credit to @PumpkinHat
https://github.com/lllyasviel/stable-diffusion-webui-forge/issues/1779#issuecomment-2364974588
2024-10-08 12:18:39 +01:00
Won-Kyu Park
c0a7438fd8 partial revert to fix compatible issue (#1892)
* partial revert to fix A1111 compatible issue
* fix for submit_extras()
2024-10-08 11:13:32 +01:00
layerdiffusion
e9d1736b9f [bot] dep ver 2024-10-07 20:52:55 -07:00
DenOfEquity
cc378589a4 (T5) pad chunks to length of largest chunk (#1990)
When the prompt is chunked using the BREAK keyword, chunks will be padded to the minimum size of 256 tokens - but chunks can be longer. torch.stack then fails if all chunks are not the same size, so find the largest and pad all to match.
#1988 (doesn't quite ID the real issue, prompts longer than 255 tokens work fine)
2024-10-07 11:37:06 +01:00
layerdiffusion
1d3f73b78d [bot] dep ver 2024-10-06 18:27:37 -07:00
DenOfEquity
2467c88c50 fix for XPU (#1997)
use float32 for XPU, same as previous fix for MPS
2024-10-06 14:33:47 +01:00
layerdiffusion
4f7f815b9f [bot] sync 2024-10-05 20:04:57 -07:00
layerdiffusion
953e3f3983 [bot] sync 2024-10-04 21:45:43 -07:00
layerdiffusion
7d88741f57 fix 2024-10-04 13:00:22 -07:00
layerdiffusion
4e16689ea9 update dep ver 2024-10-04 12:20:33 -07:00
layerdiffusion
4baf8ffb9c less prints 2024-10-03 21:23:56 -07:00
layerdiffusion
dc4f5e4119 remove some assets and dependences from repo 2024-10-03 21:17:51 -07:00
DenOfEquity
34609fde67 fix Extra networks tree show / edit (#1973)
Fix 'Show internal metadata' and 'Edit metadata' buttons when used from treeview. Issue seems like an oversight from an upstream change to handle filenames with an apostrophe.
#1244
2024-10-03 18:44:35 +01:00
altoiddealer
f4d5e8cac1 Apply memory changes via API (#1954)
Actually apply memory related changes posted to `/sdapi/v1/options`:
- 'forge_inference_memory'
- 'forge_async_loading'
- 'forge_pin_shared_memory'
2024-10-01 21:34:11 +01:00