[Bugfix] Fix DeepGEMM warmup when using `FlashInferFp8DeepGEMMDynamicBlockScaledKernel` (#49467)

Michael Goin committed 3 months ago

[Bugfix][CI] Fix `topk_softplus_sqrt` no-op on non-XPU platforms (#49452)

stefankoncarevic committed 3 months ago

[Bugfix] Retry config read to survive concurrent HF cache refresh (#49001)

peizhang56 committed 3 months ago

[Bugfix] Restore `gather_and_maybe_dequant_cache` OOB guard (#49427)

Nick Hill committed 3 months ago

[CI] Fix stale/fragile untethered kernels-root tests (#49423)

Nick Hill committed 3 months ago

Bump Flashinfer version to 0.6.15 (#48914)

Wei Zhao committed 3 months ago

[Bugfix][Parser] Fix special tokens (EOS/BOS) leaking into reasoning content (#48748)

Ben Browning committed 3 months ago

[CI] stabilize GDN prefill CuTeDSL test (#49388)

Thien Tran committed 3 months ago

[CI] Increase timeout of pytorch-compilation-unit-tests (#49450)

Nick Hill committed 3 months ago

[ROCm][CI] skip moe weight padding for eplb (#49350)

Divakar Verma committed 3 months ago

[DSv4 Perf] Skip empty c128 kernel launch, around 2x kernel performance improvement. (#48957)

Wentao Ye committed 3 months ago

[Bugfix][Renderer] Rebuild vision chunk UUIDs in async render path (#49400)

Guan-Ming Chiu committed 3 months ago

[Build] Bump vllm-flash-attn to C++20-compatible commit for torch-nightly (#49326)

Andrey Talman committed 3 months ago

Upgrade tpu-inference to v0.25.0 (#49431)

Teresa Chen committed 3 months ago

[Rust Frontend] Extract request preparation from the inference path (#49045)

Sage committed 3 months ago

[Rust Frontend][gRPC] Add abort control RPC (#49255)

Connor Carpenter committed 3 months ago

[Frontend] Parallelize preprocessing within the same request for pooling models online serving. (#49153)

wang.yuqi committed 3 months ago

[XPU] WA of topk_softplus_sqrt arg mismatch on XPU (#49408)

Liangqiusong committed 3 months ago

[XPU] WA of topk_softmax arg mismatch on XPU (#49395)

liuzhenwei committed 3 months ago

[CI] Increase timeouts for jobs exceeding current limits (#49374)

Kevin H. Luu committed 3 months ago