397631fb4c
update constraints, patch loader and create ext exclude list
...
Signed-off-by: vladmandic <mandic00@live.com >
2026-03-26 07:21:50 +01:00
48e8b3a513
Merge pull request #4711 from awsr/pep484-plus
...
Final PEP 484 updates
2026-03-26 07:10:22 +01:00
1ac3359e20
Update sd_checkpoint.py
2026-03-25 16:56:51 -07:00
2675e68813
Revert minor change
...
Will re-implement in a separate PR
2026-03-25 13:55:56 -07:00
375591ecdd
Revert non-critical changes
...
The following errors are re-introduced and will need follow-up:
- models.py, line 92
- sd_unet.py, line 41
- sd_models.py, line 1440
2026-03-25 13:47:27 -07:00
a255e5bdcf
SDNQ set sdnq_dequantize_fp32 to True by default
2026-03-25 21:46:17 +03:00
b90558430e
fix grid
...
Signed-off-by: vladmandic <mandic00@live.com >
2026-03-25 09:11:31 +01:00
400d284711
fix metadata parser
...
Signed-off-by: vladmandic <mandic00@live.com >
2026-03-25 07:32:37 +01:00
18568db41c
Add native LoRA loading for Flux2/Klein models
...
Load Flux2/Klein LoRAs as native NetworkModuleLora objects, bypassing
diffusers PEFT. Handles kohya (lora_unet_), AI toolkit (diffusion_model.),
diffusers PEFT (transformer.), and bare BFL key formats with automatic
QKV splitting for double block fused attention weights.
Includes shape validation to reject architecture-mismatched LoRAs early.
Respects lora_force_diffusers setting to fall back to PEFT when needed.
2026-03-25 04:24:49 +00:00
10942032a3
Add SDNQ_COMPILE_KWARGS env var
2026-03-25 02:10:58 +03:00
11f3921832
SDNQ make dyn loss threshold auto selected by default
2026-03-25 01:47:59 +03:00
ff247d8fd2
Merge branch 'dev' into RUF013
2026-03-24 07:12:51 -07:00
d429e11d55
Update OpenVINO to 2026 and Torch to 2.11
2026-03-24 17:06:57 +03:00
53839e464c
remove legacy quant loaders
...
Signed-off-by: vladmandic <mandic00@live.com >
2026-03-24 15:01:07 +01:00
acf475ee45
update locales
...
Signed-off-by: vladmandic <mandic00@live.com >
2026-03-24 15:01:07 +01:00
14c3cf9172
IPEX fix transformers and disable torch.tensor hijack
2026-03-24 16:58:53 +03:00
fd1646cc37
Revert extra return statement
2026-03-24 06:28:39 -07:00
876e3b9897
RUF013 updates
2026-03-24 06:15:26 -07:00
1003926646
RUF013 updates and import updates
2026-03-24 06:08:15 -07:00
09ab19c438
RUF013 updates and formatting
2026-03-24 06:07:34 -07:00
c4ebef29a9
RUF013 updates
2026-03-24 05:48:19 -07:00
3f830589d1
RUF013 updates and typing update
2026-03-24 05:19:51 -07:00
f0bb0a921a
RUF013 updates and value handling fix
2026-03-24 05:13:11 -07:00
62d2229520
RUIF013 updates and formatting
2026-03-24 05:12:18 -07:00
4f0fb7cc29
More RUF013 updates for processing_class
2026-03-24 05:00:41 -07:00
fb853af4b0
RUF013 updates
2026-03-24 04:58:19 -07:00
cff7f8d4a9
IPEX remove the use of lambdas
2026-03-24 14:36:58 +03:00
641321d7d2
RUF013 updates
2026-03-24 04:34:38 -07:00
92960de8d6
RUF013 updates
2026-03-24 04:27:30 -07:00
6ab9b7bc62
RUF013 updates
2026-03-24 04:19:53 -07:00
4f8260e670
Update IPEX to Torch 2.11
2026-03-24 14:17:24 +03:00
2d81bcdc69
RUF013 updates + type fixes
2026-03-24 04:12:14 -07:00
abe25e7b07
RUF013 update
2026-03-24 03:46:44 -07:00
d3f925e8e5
RUF013 update
2026-03-24 03:46:26 -07:00
481d974b91
RUF013 update
2026-03-24 03:40:15 -07:00
b75a1f971f
RUF013 update + minor typing fixes
2026-03-24 03:39:02 -07:00
a5977f09f5
Fix JPEG XL
2026-03-24 13:36:52 +03:00
53598d40ab
RUF013 update
2026-03-23 22:59:32 -07:00
3e228afa78
RUF013 update
2026-03-23 22:58:08 -07:00
de86927c1b
Remove unneeded type and default
2026-03-23 22:54:40 -07:00
d474f28cb3
RUF013 update
2026-03-23 22:53:34 -07:00
bc22e0f0dc
monkey patch transformers version checks
...
Signed-off-by: vladmandic <mandic00@live.com >
2026-03-23 18:30:28 +01:00
dbd2293586
remove hf token from load args
...
Signed-off-by: vladmandic <mandic00@live.com >
2026-03-23 17:53:47 +01:00
09b9ae32c1
add color grading to processing
...
Signed-off-by: vladmandic <mandic00@live.com >
2026-03-23 08:44:07 +01:00
57990fd8c0
add FireRed-Image-Edit-1.1
...
Signed-off-by: vladmandic <mandic00@live.com >
2026-03-23 06:41:46 +01:00
091f31d4bf
add Flux2/Klein LoRA support
...
- detect f2 model type for LoRAs via metadata, architecture, and filename/folder
- preprocess bare BFL-format keys with diffusion_model prefix for Flux2LoraLoaderMixin
- handle LoKR format via native NetworkModuleLokr with on-the-fly kron(w1, w2)
- add NetworkModuleLokrChunk for fused QKV split into separate Q/K/V modules
- activate native modules loaded via diffusers path
- improve error message for Flux1/Flux2 architecture mismatch
2026-03-23 02:14:07 +00:00
0d248c45e7
fix(upscaler): convert non-RGB images before ESRGAN forward pass
2026-03-23 02:12:59 +00:00
e87fcd1bc5
latent correction hints
...
Signed-off-by: vladmandic <mandic00@live.com >
2026-03-22 11:27:36 +01:00
b2b4e13bab
lint
...
Signed-off-by: vladmandic <mandic00@live.com >
2026-03-22 11:17:41 +01:00
afe3786f5f
add unet/dir to networks
...
Signed-off-by: vladmandic <mandic00@live.com >
2026-03-22 11:09:45 +01:00
vladmandic
awsr
Disty0
CalamitousFelicitousness