[ROCm][CI] Fix ModernBERT token classification test numerical accuracy on ROCm (#31820)

Andreas Karatzas committed 10 months ago

[Spec Decode][UX] Add acceptance stats to `vllm bench serve` report (#31739)

Matthew Bonanni committed 10 months ago

Report error log after vllm bench serve (#31808)

Elvir Crnčević committed 10 months ago

Fix RecursionError in MediaWithBytes unpickling (#31191)

Nikhil G committed 10 months ago

[Quantization][Refactor] Move CPU GPTQ kernel into MP linear (#31801)

Li, Jiang committed 10 months ago

[ROCm][CI] Fix tests/compile unit tests (#28895)

Charlie Fu committed 10 months ago

[Perf] Async Scheduling + Speculative Decoding + Structured Outputs (#29821)

Benjamin Chislett committed 10 months ago

[Bugfix] Fix GLM-4 MoE router logits dtype for data parallel chunking (#31055)

Yakine Tahtah committed 10 months ago

make 500: InternalServerError more informative (#20610)

Masataro Asai committed 10 months ago

[Log] add log about gpu worker init snapshot and requested memory (#29493)

Ning Xie committed 10 months ago

[PERF] Speed-up of GDN attention decode part (Qwen3-Next) (#31722)

Vadim Gimpelson committed 10 months ago

[Attention][2/n] Remove usage of deprecated `seq_lens_cpu` and `num_computed_tokens_cpu` CommonAttentionMetadata properties (#31774)

Lucas Wilkinson committed 10 months ago

[Quantization][MoE] remove unused ep logic from moe marlin (#31571)

Jinzhen Lin committed 10 months ago

[NemotronH] Use ReplicatedLinear for fc1_latent_proj (#31807)

roikoren755 committed 10 months ago

[MoE Refactor][14/N] Clean Up FI Quant Config Smuggling (#31593)

Robert Shaw committed 10 months ago

[MoE Refactor] Add Temporary Integration Tests - H100/B200 (#31759)

Robert Shaw committed 10 months ago

[Bugfix]: Fix cross attention backend selection for Turing GPU (#31806)

Isotr0py committed 10 months ago

[LoRA]Disable linear LoRA kernel PDL (#31777)

Jee Jee Li committed 10 months ago

[Model] rename use_pad_token to use_sep_token (#31784)

wang.yuqi committed 10 months ago

[Frontend] Support GLM-4.5 / GLM-4.7 with enable_thinking: false (#31788)

Chauncey committed 10 months ago