[Bugfix] NemotronHMTP: add hf_to_vllm_mapper so quant exclusions reach the MTP draft (#53790)

Juhi Mittal committed about 2 months ago

[Bugfix] Reject empty bad-word tokenizations (#53433)

Andreas Karatzas committed about 2 months ago

[Rust Frontend][CI] Remove TCP port races from mock-engine tests (#54481)

Andreas Karatzas committed about 2 months ago

[Bugfix][Test] Fix off-by-one error in sampled token rank causing flaky logprobs test (#53976)

Qiming Zhang committed about 2 months ago

Add Laguna-XS-2.1-INT4 to nightly CI (#52961)

Joe Rowell committed about 2 months ago

[ROCm][DSpark][DCP] Serve prefix cache hits under DCP for Kimi-K3 (#53598)

YukioZzz committed about 2 months ago

[ROCm][MLA][DCP] Support causal multi-token verification (#51705)

YukioZzz committed about 2 months ago

[Structured Output] Let terminal grammars stop under min_tokens (II) (#54218)

Artem Perevedentsev committed about 2 months ago

[CI] Exclude kv_transfer changes from broad spec-decode/kernels/multimodal triggers (#54365)

Nicolò Lucchesi committed about 2 months ago

[ROCm] Add TheRock preview docker updates, Keep Python 3.12 and Ubuntu 22.04 (#49925)

rasmith committed about 2 months ago

[Tests][XPU] Limit Qwen2-VL generation length to avoid flaky numerical divergence (#54172)

Fanli Lin committed about 2 months ago

[AutoRound] Support AutoRound Format Block-Wise FP8 in vLLM (#47434)

Zhenzhong Xu committed about 2 months ago

[Core] Add `max_num_queued_reqs` and `max_num_queued_tokens` for queue size management (#49445)

Nicolò Lucchesi committed about 2 months ago

[Bugfix][ROCm] Preserve AITER unified-attention metadata during graph replay (#53821)

andyluo7 committed about 2 months ago

[KV Connector][Offloading] Certify attention-only hybrids in the canonical portability gate (#51689)

Itay Etelis committed about 2 months ago

[Bugfix][KV Offload] Unlink /dev/shm region after all workers map it (barrier variant of #51317) (#52596)

Itay Etelis committed about 2 months ago

[Bugfix][KV Offload][P2P] Preserve aborted loads until abort completion (#52571)

Uttam Kumar committed about 2 months ago

[kernel] Fused embedding kernel (#53677)

Jee Jee Li committed about 2 months ago

[KV offload] Order CPU->GPU loads against the compute stream (#50696)

Itay Etelis committed about 2 months ago

[Frontend][Performance] Resolve async media across modalities concurrently (#54537)

waizuichougou committed about 2 months ago