[Quantization]: Support compressed-tensors mixed-precision model loading (#22468)

Dipika Sikka committed over 1 year ago

[Kernel] Add nvfp4 gemm flashinfer backends (#22346)

nvjullin committed over 1 year ago

[CI] Temporarily disable flaky test (#22930)

Lucas Wilkinson committed over 1 year ago

docs: update fastsafetensors usage instructions (#22891)

Nir committed over 1 year ago

[Kernel] [Quantization] Add MXFP4 and bias support for marlin kernel (#22428)

Jinzhen Lin committed over 1 year ago

[CI] [Hybrid] Bump min transformers version for Bamba and Jamba (#22908)

Thomas Parnell committed over 1 year ago

[Bugfix] Fix parsing of `--disable-mm-preprocessor-cache` (#22909)

Cyrus Leung committed over 1 year ago

[Bugfix] Replace custom Encoding class with BatchEncoding in MistralTokenizer (#22786)

Jiangyun Zhu committed over 1 year ago

[Model] Modify the gate implementation of glm4_moe (#22832)

Jee Jee Li committed over 1 year ago

[Perf] Dont create unnecessary pooling params (#22876)

Lucas Wilkinson committed over 1 year ago

[CI] Re-enable transcriptions `test_long_audio_request` (#22890)

Nicolò Lucchesi committed over 1 year ago

[Doc] fix dead link (#22898)

Daniele committed over 1 year ago

[Bugfix] Add reset prefix cache for online serving (#22726)

iAmir97 committed over 1 year ago

Remove Phi 4 Flash configuration workaround (#22723)

Harry Mellor committed over 1 year ago

[BugFix] Threadsafe close async zmq sockets (#22877)

Nick Hill committed over 1 year ago

[Bugfix] Fix `PixtralHFImagePixelInputs` dynamic shape check (#22827)

Isotr0py committed over 1 year ago

vLLM Benchmark suite improvement (#22119)

Louie Tsai committed over 1 year ago

[CI] remove flaky v0 test (#22864)

Robert Shaw committed over 1 year ago

[CI] Fix `tests/distributed/test_ca_buffer_sharing.py` (#22849)

Ilya Markov committed over 1 year ago

[CI][Entrypoints]: add filter to generation to filter out invalid tool calls (#22826)

Will Eaton committed over 1 year ago