NumberWan 23d8c83dab [Bugfix][Qwen-Image] Keep AutoRound W4A16 on CPU when offloading; Fixes #7555 (#7579) ...

Signed-off-by: NumberWan <wantszkin2003@gmail.com>

2026-09-19 16:34:11 +00:00

..

__init__.py

enhancement: extend to dmd2 to image generation + add flux, qwen image pipelines (#2974)

2026-05-08 21:37:57 +08:00

autoencoder_kl_qwenimage.py

[New model] Support model qwen image layered (#381)

2025-12-20 14:51:50 +00:00

cfg_parallel.py

[UX] Add progress bar for diffusion models (#1652)

2026-03-05 23:56:14 +08:00

pipeline_qwen_image_edit_plus.py

[Refactor][Misc] Remove unused code in diffusion hooks/model_loader (#5270)

2026-07-22 09:45:17 +08:00

pipeline_qwen_image_edit.py

fix(qwen-image): drop the shadowed dead _get_qwen_prompt_embeds duplicate (#7461)

2026-09-14 13:48:51 +00:00

pipeline_qwen_image_layered.py

[Refactor][Misc] Remove unused code in diffusion hooks/model_loader (#5270)

2026-07-22 09:45:17 +08:00

pipeline_qwen_image.py

[Bugfix][Qwen-Image] Keep AutoRound W4A16 on CPU when offloading; Fixes #7555 (#7579)

2026-09-19 16:34:11 +00:00

qwen_image_transformer.py

[Bugfix][Qwen-Image] Restore RotaryEmbedding CUDA RoPE for Diffusers e2e (#7513)

2026-09-16 08:54:28 -07:00

rope_utils.py

[BugFix][Qwen-Image] align txt_seq_lens RoPE width with padded embeds (#4474)

2026-06-16 19:09:25 +08:00