0e35079f06
attempt to load incomplete model
...
Signed-off-by: Vladimir Mandic <mandic00@live.com >
2026-06-29 13:39:46 +02:00
e7e317191a
automated pipeline registrations and tests
...
Signed-off-by: Vladimir Mandic <mandic00@live.com >
2026-06-04 10:18:09 +02:00
c319f8f92d
refactor shared text encoder
...
Signed-off-by: Vladimir Mandic <mandic00@live.com >
2026-06-01 10:03:10 +02:00
77913c7c0a
docs: cleanup
...
- generic.load_transformer: shorter dispatch description, no internal taxonomy
- native_transformer.load: trim kwargs forwarding wording
- sd_unet revert: one-line comment
- test: shorten kwargs-forwarding docstrings
2026-05-31 02:04:56 +01:00
25fa8e0f15
refactor(ui): shorten unet-sync handlers, align load logs
...
- ui_settings: rename to sync_checkpoint_unet / sync_checkpoint_unet_forced
- sd_models: cross-arch unet reset log uses key=value house style
- generic: native transformer dispatch logs loader=native
2026-05-31 00:24:33 +01:00
18656d1d79
fix(native_transformer): forward load kwargs to from_config
...
load_transformer threads **kwargs into its from_pretrained and from_single_file branches. The native dispatch branch bypasses both loaders and builds via cls.from_config, so those kwargs were dropped there. Thread them through load, build_component, and build_component_quantized into from_config so callers passing extra args alongside native_spec are honored instead of silently dropped. Siblings do not receive them.
2026-05-31 00:24:33 +01:00
7ae45c59d4
refactor(native_transformer): accept precomputed quant_args from caller
...
generic.load_transformer already computes quant_args and quant_type at the top of the function, but native_transformer.load was redoing the same work and emitting a duplicate 'Quantization: module=Model ...' log line milliseconds later. Pass them through instead; fall back to computing them when called directly (model_anima still does so).
2026-05-31 00:24:33 +01:00
98fba8ddc3
feat(pipelines): wire native_spec dispatch in generic.load_transformer
...
New native_spec=None kwarg. When set and the UNET dropdown points at a
.safetensors, dispatches to native_transformer.load (threading
allow_quant/dtype/modules_to_not_convert/modules_dtype_dict). Pipelines
without a spec stay on cls.from_single_file unchanged.
2026-05-31 00:24:33 +01:00
113d4efae8
add lens img2img and inpaint
...
Signed-off-by: Vladimir Mandic <mandic00@live.com >
2026-05-25 10:40:20 +02:00
b6b5246c53
add vibe-sana
...
Co-authored-by: Copilot <copilot@github.com >
Signed-off-by: Vladimir Mandic <mandic00@live.com >
2026-05-05 18:46:20 +02:00
40330c8d76
patch rich traceback
...
Co-authored-by: Copilot <copilot@github.com >
Signed-off-by: Vladimir Mandic <mandic00@live.com >
2026-05-05 17:02:17 +02:00
ee5978f72b
custom vae loader
...
Co-authored-by: Copilot <copilot@github.com >
Signed-off-by: Vladimir Mandic <mandic00@live.com >
2026-05-02 10:22:43 +02:00
e5c494f999
cleanup logger
2026-02-19 11:09:13 +01:00
a3074baf8b
unified logger
2026-02-19 09:46:42 +01:00
a2ee885e28
refactor: update nunchaku repo URLs and version handling
...
- Rename HuggingFace org from nunchaku-tech to nunchaku-ai across all
nunchaku model repos (flux, sdxl, sana, z-image, qwen, t5)
- Add per-torch-version nunchaku version mapping instead of single global
version, with robust torch version parsing
2026-02-07 22:27:05 +00:00
dd9001b20d
fix(model_flux2_klein): update generic.py and pipeline
2026-01-17 21:36:44 +00:00
b3d65f4559
logging cleanup
...
Signed-off-by: vladmandic <mandic00@live.com >
2026-01-16 11:32:09 +01:00
fe99d3fe5d
feat: add FLUX.2 Klein model support
...
Add support for FLUX.2 Klein distilled models (4B and 9B variants):
- Add pipeline loader for Flux2KleinPipeline
- Add model detection for 'flux.2' + 'klein' patterns
- Add pipeline mapping in shared_items
- Add shared Qwen3ForCausalLM text encoder handling:
- 4B variants use Z-Image-Turbo's Qwen3-8B
- 9B variants use FLUX.2-klein-9B's Qwen3-14B
- Add reference entries for distilled (4B, 9B) and base models
- Update diffusers commit for Flux2KleinPipeline support
2026-01-16 01:35:20 +00:00
a319a98e59
error handling of meta embeds
...
Signed-off-by: vladmandic <mandic00@live.com >
2025-12-23 10:23:14 +01:00
55c089ae48
Update runai-model-streamer logging integration
...
- Remove stdout redirect monkeypatch (fixed in runai v0.15.1 via PR #97 )
- Add RUNAI_STREAMER_LOG_LEVEL controlled by SD_LOAD_DEBUG
- Add one-time runai config log when hijack is activated
- Add `loader=runai|default` to model loading logs
- Remove per-file logging clutter from sd_hijack_safetensors.py
2025-12-02 02:01:51 +00:00
c50c987a33
sdnq always-register
...
Signed-off-by: Vladimir Mandic <mandic00@live.com >
2025-11-16 08:14:01 -05:00
93f28f07ac
Make SDNQ not depended on quantization_config.json and fix invalid quantization_config getting attached to the model on load
2025-11-07 18:11:21 +03:00
d2caffa7b4
improve runai-streamer integration
...
Signed-off-by: Vladimir Mandic <mandic00@live.com >
2025-11-02 12:14:15 -05:00
3ae10dd0e1
add nvidia-chronoedit
...
Signed-off-by: Vladimir Mandic <mandic00@live.com >
2025-10-30 19:52:29 -04:00
b54839004a
Fix modules_to_not_convert not getting reset
2025-10-27 22:29:08 +03:00
9308a28c9d
split diffusers/transformers safetensors patch
...
Signed-off-by: Vladimir Mandic <mandic00@live.com >
2025-10-27 14:39:24 -04:00
7ada86032a
add krea, cleanup, lint
...
Signed-off-by: Vladimir Mandic <mandic00@live.com >
2025-10-25 13:25:09 -04:00
f903fe3473
cleanup
2025-10-25 20:09:24 +03:00
975361b404
import sdnq before load
2025-10-25 20:06:15 +03:00
01c4ce18b6
SDNQ fix T5 Loading
2025-10-25 20:03:12 +03:00
203e46af0c
update reference sizes
...
Signed-off-by: Vladimir Mandic <mandic00@live.com >
2025-10-25 12:18:52 -04:00
e4711bcb6c
add hunyuanimage-2.1
...
Signed-off-by: Vladimir Mandic <mandic00@live.com >
2025-10-24 10:29:58 -04:00
7fdc880a73
sdnq patches
...
Signed-off-by: Vladimir Mandic <mandic00@live.com >
2025-10-07 09:43:34 -04:00
a6108dd6df
add qwen pruning variants
...
Signed-off-by: Vladimir Mandic <mandic00@live.com >
2025-10-03 11:36:14 -04:00
d351fdb98f
add more job state updates and update history tab
...
Signed-off-by: Vladimir Mandic <mandic00@live.com >
2025-09-13 10:54:04 -04:00
0bb6603c2e
cleanup diag logging
...
Signed-off-by: Vladimir Mandic <mandic00@live.com >
2025-08-16 09:46:49 -04:00
65847bc022
conditional cache clear
...
Signed-off-by: Vladimir Mandic <mandic00@live.com >
2025-08-15 15:49:58 -04:00
41ae06bd90
fix loading custom t5
...
Signed-off-by: Vladimir Mandic <mandic00@live.com >
2025-08-13 16:06:25 -04:00
15cb8fe9f8
SDNQ add modules_dtype_dict and fix Qwen Image with quants less than 5 bits
2025-08-13 00:07:36 +03:00
362ec0d412
Fix Chroma quantization
2025-08-12 22:07:52 +03:00
8b9ba82982
Use generic model loader with lumina2
2025-08-11 23:57:40 +03:00
9daa30f371
Add torch_dtype to transformer.from_single_file
2025-08-11 23:37:07 +03:00
232a631be5
Use pre-mode quant on custom transformer files
2025-08-11 23:31:06 +03:00
87bd347116
cleanup flux loader
...
Signed-off-by: Vladimir Mandic <mandic00@live.com >
2025-08-11 16:05:39 -04:00
dc8a72947d
fix meissonic
...
Signed-off-by: Vladimir Mandic <mandic00@live.com >
2025-08-11 16:05:39 -04:00
e42a27a0e4
fix chroma
...
Signed-off-by: Vladimir Mandic <mandic00@live.com >
2025-08-11 16:05:39 -04:00
d7daefb8ac
update cogview links, fix auraflow and hidream loaders
...
Signed-off-by: Vladimir Mandic <mandic00@live.com >
2025-08-11 16:05:39 -04:00
2a85c05689
refactor pipeline loaders to generic methods and introduce te_shared_t5 option
...
Signed-off-by: Vladimir Mandic <mandic00@live.com >
2025-08-11 16:05:39 -04:00
Vladimir Mandic
CalamitousFelicitousness
Disty0