CalamitousFelicitousness 028e892104 feat(caption): warn when qwen3.5 linear attention kernels are missing ...

Qwen3.5 runs most of its layers as gated delta linear attention. Without
flash-linear-attention, transformers falls back to a per-token torch loop that
runs sequentially over prefill and decode, so a caption takes minutes with no
indication of why.

2026-07-15 01:36:12 +01:00

..

__init__.py

linting

2026-07-03 22:05:33 +02:00

attention.py

captioning improvements and cleanup

2026-05-29 21:10:56 +02:00

caption.py

cleanup logger

2026-02-19 11:09:13 +01:00

deepbooru_model.py

modernize typing

2026-02-19 09:15:37 +01:00

deepbooru.py

regen all localizations

2026-04-01 10:29:08 +02:00

deepseek.py

captioning improvements and cleanup

2026-05-29 21:10:56 +02:00

gemini.py

regen all localizations

2026-04-01 10:29:08 +02:00

helpers.py

captioning improvements and cleanup

2026-05-29 21:10:56 +02:00

joycaption.py

captioning improvements and cleanup

2026-05-29 21:10:56 +02:00

joytag.py

captioning improvements and cleanup

2026-05-29 21:10:56 +02:00

logits.py

prompt-enhance-v2 with steering

2026-05-22 16:58:23 +02:00

models_def.py

feat(caption): add toriigate 0.5

2026-07-15 01:36:12 +01:00

moondream3.py

fix ui/misc: None desc crash, ratio/tile/lora split guards, parse_metadtaa typo, moondream type guard

2026-06-12 13:00:50 -04:00

openclip.py

caption: add image analyze

2026-05-14 13:16:26 +02:00

tagger.py

RUF013 update

2026-03-23 22:58:08 -07:00

toriigate.py

feat(caption): add toriigate 0.5

2026-07-15 01:36:12 +01:00

vqa_detection.py

complete refactor javascript to typescript and reorg frontend files and folders

2026-05-20 18:08:10 +02:00

vqa.py

feat(caption): warn when qwen3.5 linear attention kernels are missing

2026-07-15 01:36:12 +01:00

waifudiffusion.py

regen all localizations

2026-04-01 10:29:08 +02:00