[Bugfix] Fix more multimodal tests for transformers V5 (#34334)

Raushan Turganbay committed 9 months ago

[Bugfix] send None sentinel on final commit so server properly sends transcription.done (#33963)

Junseo Park committed 9 months ago

[Benchmarks] Reduce ready checker log verbosity (#34349)

Tomas Ruiz committed 9 months ago

[GPT-OSS] Remove unnecessary contiguous (#34337)

elvischenv committed 9 months ago

[ROCm] [CI] fix test_unrecognized_env (#34350)

TJian committed 9 months ago

[Bugfix] Enable attn quantization of Llama-4 by correctly permuting scales for rope (int8, fp8) (#34243)

Eldar Kurtić committed 9 months ago

[Bugfix] fix default is_neox_style is True for deepseek (#34353)

Xinyu Dong committed 9 months ago

[Multimodal] Expose `mm_processor_kwargs` for `DummyInputsBuilder` (#34330)

Isotr0py committed 9 months ago

[Model Runner V2] Init cuda graph pool when necessary (#33217)

Xinyu Chen committed 9 months ago

[CI][BugFix] Fix silent failure in shellcheck hook and baseline exist… (#32458)

junuxyz committed 9 months ago

[Docs] Fix typo ("defult") and double spacing (#34348)

SorenDreano committed 9 months ago

[ROCm] [aiter] Split KV cache update for AiterFlashAttention (#33681)

kliuae committed 9 months ago

[Bugfix]: Fix ROCm fusion attn test; use AttentionBackend utils to create kv cache (#33948)

Rohan Potdar committed 9 months ago

Don't try and run GLM-ASR with remote code (#34352)

Harry Mellor committed 9 months ago

Reapply [Attention][FA3] Update FA3 to include new swizzle optimization (#34043)

Lucas Wilkinson committed 9 months ago

Responses harmony system message structured (#34268)

Adam Binford committed 9 months ago

[NVIDIA][test] Tests for flashinfer TRTLLM BF16 MoE (#33715)

Linda committed 9 months ago

Make JAIS compatible with Transformers v5 (#34264)

Harry Mellor committed 9 months ago

Make Qwen3VL compatible with Transformers v5 (#34262)

Harry Mellor committed 9 months ago

[Bugfix][CPU] Fix llama4 inference on CPU (#34321)

Li, Jiang committed 9 months ago