[Transformers backend] Find attention with a fuser and attach vLLM's layer to it (#54941)

Thomas Ortner committed about 2 months ago

[ROCm][CI] Bump ROCk release image build timeout to 3h (#55354)

Rohan Potdar committed about 2 months ago

[feat] add torchcodec as audio loader and implement selective audio backend (#51826)

Jared Wen committed about 2 months ago

[Frontend] Warn when removed guided-decoding fields are present in a request (#54285)

erdholion committed about 2 months ago

[CI] Only run GitHub Actions on the main repo (#55349)

Harry Mellor committed about 2 months ago

[Bugfix][MoE] Allow TRTLLM FP8 block-scale MoE with SwiGLU clamp (#55069)

aoshen02 committed about 2 months ago

[Bugfix] Fix double BOS in LLM.chat() for multimodal models (#55288)

summer committed about 2 months ago

Fast Start (#54921)

siyu committed about 2 months ago

[Test] Split test_sampling_mask_preserves_top_k_boundary_ties to remove Triton-kernel-specific assumption Description (#55315)

Qiming Zhang committed about 2 months ago

[Qwen3.8-Flash-Next] Improve QSA sparse GQA for prefill and short-ctx decode (#54873)

Thien Tran committed about 2 months ago

[CI/Build] Fix flaky failures in CPU CI image building (#55317)

Li, Jiang committed about 2 months ago

[Security] Cap GLMGA video sampling to prevent request-driven resource exhaustion (#54935)

Juan Pérez de Algaba committed about 2 months ago

[Mypy] Fix mypy typing for L models (#54177)

Taneem Ibrahim committed about 2 months ago

[AMD][kimik3][ROCm][Perf] Fuse MLA q/kv RMSNorm in AMD Kimi-K3 MLA wrapper (#52494)

Rita Brugarolas committed about 2 months ago

[Kernel] Reuse Qwen4Exp HC combine-norm for MTP input (#54687)

Canlin Guo committed about 2 months ago

[CI/Test] Add expert parallelism coverage to external LB tests (#53497)

Tony Lin committed about 2 months ago

[XPU] Fix device assignment for DP external LB (#53037)

Tony Lin committed about 2 months ago

[SpecDecode]Fix spec decode warmup device selection (#55245)

Kunshang Ji committed about 2 months ago

[Qwen3.8-Flash-Next] Compact indexer logits workspace to improve prefill efficiency (#54915)

Thien Tran committed about 2 months ago

[MRV2][Metrics] Support `CUDAGraphStat` in MRV2 (#52358)

Yizhou committed about 2 months ago