v1.1.9
What's Changed
- feat: uaa avoid extra memory IO access by @triple-mu in #551
- chore: simplify quantize flags in example utils by @DefTruth in #553
- chore: fix quantize flags in example by @DefTruth in #554
- chore: fix quantize & TP conflicts for wan by @DefTruth in #556
- feat: support serving text2video by @BBuf in #555
- chore: Update SERVING Doc and FAQ Doc by @BBuf in #557
- chore: qwen edit lightning cp/tp examples by @DefTruth in #559
- feat: support ovis-image context parallel by @DefTruth in #560
- feat: serving support image2video by @BBuf in #558
- chore: add collect_env script by @DefTruth in #562
- Add pre-commit and GitHub Actions CI by @DefTruth in #564
- chore: refactor parallelism for better reusability by @DefTruth in #565
- chore: Update vLLM-Omni integration by @SamitHuang in #566
- feat: add pipe quant config for serving by @nono-Sang in #563
- News: 🔥vLLM-Omni x Cache-DiT ready! by @DefTruth in #567
- feat: enable custom attn backend for TP by @DefTruth in #568
- feat: support TP for many text encoder by @DefTruth in #569
- fix qwen-edit-lightning examples by @DefTruth in #571
- fix get_text_encoder_from_pipe by @DefTruth in #572
- fix: handle general compile options in example utils by @DefTruth in #573
- chore: reduce un-popular examples by @DefTruth in #574
- feat: add text_encoder tp for serving by @nono-Sang in #570
- chore: simplify example by @DefTruth in #575
- chore: make unified examples by @DefTruth in #576
- chore: fix vllm-omni docs link by @DefTruth in #577
- chore: optimize examples default path mapping by @DefTruth in #579
- chore: fix vllm-omni docs link by @DefTruth in #580
- feat: support Ovis-Image tensor parallel by @DefTruth in #582
- chore: fix typo in User_Guide.md by @DefTruth in #583
- chore: fail fast TP validation for attn heads by @CPFLAME in #581
- fix patch functor for multi transformers by @DefTruth in #586
- chore: add qwen image controlnet example by @DefTruth in #588
- chore: update docs by @DefTruth in #590
- feat: register fa3 backend for context parallel by @nono-Sang in #589
- chore: support separate quant-type for text encoder by @DefTruth in #591
- hotfix for fa3 backend import error by @DefTruth in #593
- chore: fix typo in README.md by @DefTruth in #594
- chore: set save_ctx to False for inference by @nono-Sang in #596
- fix flux examples model path mismatch by @DefTruth in #597
New Contributors
- @SamitHuang made their first contribution in #566
- @nono-Sang made their first contribution in #563
- @CPFLAME made their first contribution in #581
Full Changelog: v1.1.8...v1.1.9