[ROCm][CI] Fix TestSiluMulGroupFp8QuantModel after W8A8 block linear refactor (#39799)

Andreas Karatzas committed 6 months ago

[Test] Increase qwen2_vl num_logprobs to fix torch 2.12 update (#40818)

Angela Yi committed 6 months ago

[Refactor] Remove unused dead code (#40640)

Wentao Ye committed 6 months ago

[BE][Bugfix] Respect TORCH_COMPILE_DISABLE env var at the vLLM config level for torch 2.12 (#40715)

Lucas Kabela committed 6 months ago

[Bugfix] add seq_lens_cpu_upper_bound to CommonAttentionMetadata in mla_runner.py (#40844)

Ignacio Sica committed 6 months ago

[EPLB] Fix replica selection bias in fused_moe router (#40810)

Artem Perevedentsev committed 6 months ago

[Bugfix][CI] Fix wrong residual shape in TestFusedAddRMSNorm.example_inputs that causes flaky test (#40629)

Zhang Jian committed 6 months ago

[Bug] Fix GLM-5.1 running error on ROCm platform (#40763)

qli88 committed 6 months ago

[Build] Add Python 3.14 to supported version list. (#34770)

Neil Schemenauer committed 6 months ago

[Frontend] Delegate to vLLM Omni When `--omni` Passed (#40744)

Alex Brooks committed 6 months ago

[Docs] Add docs for context extension using the yarn method (#37430)

labAxiaoming committed 6 months ago

[Refactor] Unify 2D/3D kernels in triton_unified_attention (#40631)

JartX committed 6 months ago

[Quantization] add humming quantization kernel (#34556)

Jinzhen Lin committed 6 months ago

fused_moe: treat NIXL EP as batched experts (#40412)

Itay Alroy committed 6 months ago

[kv_offload+HMA][9/N]: Support lookup with multiple KV groups (#39401)

Or Ozeri committed 6 months ago

[Bugfix][MoE] Unpad routed output before shared expert add [Fixes #35949] (#40794)

Netanel Haber committed 6 months ago

[Attention] use diff kv backend for mimo v2 flash (#40045)

Jiangyun Zhu committed 6 months ago

[Bugfix][Parser] Fix Mistral tool parser for HF tokenizers (#39294)

Thomas committed 6 months ago

[Build] Bump CUDA to 13.0.2 to match PyTorch 2.11.0 (#40669)

Dmitry Tokarev committed 6 months ago

[CI/Build] Add e2e test for ViT CUDA graph (#40780)

Shanshan Shen committed 6 months ago