[Model] Remove unnecessary CUDA sync of GLM-4.1V image and video preprocess (#24332)

Chatcharin Sangbutsarakum committed about 1 year ago

[P/D] Add a shutdown method to the Connector API (#22699)

Chauncey committed about 1 year ago

[Doc] Fix UTF-8 encoding issues in documentation generation on Windows (#24361)

Al-Ekram Elahee Hridoy committed about 1 year ago

[CI/Build] Fix local image inputs in test_pixtral.py (#24401)

Chenheli Hua committed about 1 year ago

[CI/Build] Disable flaky test_structured_output tests (#24404)

22quinn committed about 1 year ago

[Sampler] Support returning all prompt logprobs (#23868)

Xingyu Liu committed about 1 year ago

[xpu] upgrade ipex/python3.12 for xpu (#23830)

Yan Ma committed about 1 year ago

[Kernel] Support decode context parallelism on Blackwell with CUTLASS MLA (#24385)

Ming Yang committed about 1 year ago

Add @luccafong to codeowner for spec decode (#24397)

Lucia Fang committed about 1 year ago

Skip MM Encoder for non-first PP ranks (#24387)

Woosuk Kwon committed about 1 year ago

Add renderer-based prompt processing for embedding and classification endpoints (#24356)

Flora Feng committed about 1 year ago

[TPU] Remove TopKTopPSampler dependency for TPU sampler (#24391)

Woosuk Kwon committed about 1 year ago

[Bugfix] Fix Qwen3-coder moe tuned config (#24072)

Jee Jee Li committed about 1 year ago

[Misc] collect flashinfer version in collect_env.py (#24378)

Ye (Charlotte) Qi committed about 1 year ago

QWEN3 Thinking Fused MoE kernels Optimization configs (#24330)

Saman A. Pour committed about 1 year ago

[Benchmark] add benchmark for custom activation op (#23908)

Jiangyun Zhu committed about 1 year ago

[CI][Fix] deterministic seed for flaky CI runs on structured outputs (#24380)

Aaron Pham committed about 1 year ago

Migrate Qwen2 inputs to TensorSchema (#23475)

Benji Beck committed about 1 year ago

[attention][DCP] use AttentionImpl.need_to_return_lse_for_decode (#24372)

youkaichao committed about 1 year ago

[V0 deprecation] Deprecate V0 Neuron backend (#21159)

Woosuk Kwon committed about 1 year ago