[Bugfix] fix use-ep bug to enable ep by dp/tp size > 1 (#16161)

zxfan-cpu committed over 1 year ago

[V1] Scatter and gather placeholders in the model runner (#16076)

Roger Wang committed over 1 year ago

Update BASE_IMAGE to 2.22 release of Neuron (#16218)

Satyajith Chilappagari committed over 1 year ago

[Misc] Update compressed-tensors to version 0.9.3 (#16196)

Miles Williams committed over 1 year ago

[core] do not send error across process (#16174)

youkaichao committed over 1 year ago

[ROCm][Bugfix][FP8] Make fp8 quant respect fused modules mapping (#16031)

Michael Goin committed over 1 year ago

Torchao (#14231)

Driss Guessous committed over 1 year ago

[Bugfix][V0] XGrammar structured output supports Enum (#15878)

leon-seidel committed over 1 year ago

[Misc] add description attribute in CLI (#15921)

Reid committed over 1 year ago

[V1][BugFix] Exit properly if engine core fails during startup (#16137)

Nick Hill committed over 1 year ago

[Bugfix] Fix guidance backend for Qwen models (#16210)

Benjamin Chislett committed over 1 year ago

[Misc] Move Llama 4 projector call into encoder execution (#16201)

Roger Wang committed over 1 year ago

[Misc] Human-readable `max-model-len` cli arg (#16181)

Nicolò Lucchesi committed over 1 year ago

Print the warning only once (#16193)

Gregory Shtrasberg committed over 1 year ago

[V1] Revert the default `max_num_seqs` to V0 values for most hardware (#16158)

Cyrus Leung committed over 1 year ago

[Bugfix] Re-enable support for `ChatGLMForConditionalGeneration` (#16187)

Cyrus Leung committed over 1 year ago

Upstream Llama4 Support to Main (#16113)

Lu Fang committed over 1 year ago

[Doc]Update image to latest version (#16186)

Robin committed over 1 year ago

[Misc] improve example mlpspeculator and llm_engine_example (#16175)

Reid committed over 1 year ago

[Model] Add Qwen3 and Qwen3MoE (#15289)

YamPengLi committed over 1 year ago