[Frontend] Upgrade XGrammar to 0.2.7 and Rust structural tags to 0.3.0 (#57272)

Bugen Zhao committed about 1 month ago

[Bugfix][Multimodal] Preserve DeepSeek V4 image block spacing (#56882)

adenzhou1350 committed about 1 month ago

[CI] Raise GSM8K startup max wait to 2400s for flaky 30B+ MoE eval configs (#57361)

Kevin H. Luu committed about 1 month ago

[CI] Add flaky rerun markers / timeout skips to sibling tests lacking them (#57359)

Kevin H. Luu committed about 1 month ago

[ROCm][Bugfix] Fix intermittent ROCR host segfault (#57328)

Matt committed about 1 month ago

[CI] Add residual timeout headroom after JIT rollback (#56649)

Kevin H. Luu committed about 1 month ago

[CI] Raise Elastic EP Scaling step timeout 30m -> 40m (#57287)

vllm-agent committed about 1 month ago

[CPU] Add CPU FP8 W8A8 linear/MoE support (#49942)

Yuwen Zhou committed about 1 month ago

[Quantization] Support native Quark W4A16 INT4/UINT4 exports in vLLM (#48606)

meng-amd committed about 1 month ago

[Bugfix][CPU] Fix macOS multimodal SHM cache initialization (#57142)

shaohuaxi committed about 1 month ago

[Bugfix][MLA] TritonMLA: fix illegal memory access on causal multi-token decode (#51065)

olka-amd committed about 1 month ago

C3x SM100 FP8 blockwise - Pad activation scales to multiple of 4 (#57377)

Thien Tran committed about 1 month ago

[Rust Frontend] Return sampling masks over gRPC (#56777)

Biswa Panda committed about 1 month ago

[Bugfix][DBO] Fix DeepEP low-latency profiling crash with DP+EP+DBO (#57502)

Micah Williamson committed about 1 month ago

[Perf] Fuse CohereASR relative attention score accumulation (#55190)

Sun committed about 1 month ago

[Frontend] Fix the parsing of missing `string=` in DeepSeek V4 (#56271)

lazymio committed about 1 month ago

[Frontend] Add `--tool-strict-level` for server-side control for structural tag activation (#56268)

lazymio committed about 1 month ago

[Bugfix][ROCm] Alias SparseAttnIndexerKpool.forward_cuda to forward_native (GLM-5.3-Flash boot crash) (#57425)

Mustafa YILDIRIM committed about 1 month ago

[Perf][MRV2] Share token-to-request mappings across KV cache groups (#57102)

Juntian Liu committed about 1 month ago

[Bugfix] Attach request-level tools to existing system message in DeepSeek V4 Python renderer (#51856)

Jason committed about 1 month ago