[Bugfix][NIXL] Fix multiple handles xfer race (#56104)

Nicolò Lucchesi committed about 1 month ago

[CI] Pass the scale-out endpoint flag in EC E2E (#56819)

Kevin H. Luu committed about 1 month ago

[CI] Drop root privileges for the OTel GPU sampler (#56941)

Kevin H. Luu committed about 1 month ago

[CI/Build] Retry CPU image stream interruptions (#56597)

Kevin H. Luu committed about 1 month ago

[Kernel][MoE] DeepEP v2: async finalize to overlap shared experts with combine (#52781)

404mario committed about 1 month ago

[Bugfix] Avoid nested rerank cancellation handlers (#55602)

Miguel committed about 1 month ago

[Bugfix][NIXL] Report a full prefix cache hit as finished (#56640)

Yizheng Jiao committed about 1 month ago

[Model][DSv4.1] Store the whole KV in MXFP8 (FlashMLA V4.1 record) (#56893)

Yongye Zhu committed about 1 month ago

[Bugfix] Handle bare and malformed tool call openers in Gemma4 parser (#53444)

wuhangxian committed about 1 month ago

[Bugfix][KV Offload] Ignore pending chunks in invalid sliding windows (#56486)

Liqian Lin committed about 1 month ago

[Bugfix] Export weight-cache IPC tensors separately for each client (#56472)

shaohuaxi committed about 1 month ago

[KVConnector][P2P] Configurable unbound-store timeout and one-RTT rejection of a late fetch (#53453)

liranschour committed about 1 month ago

[XPU][CI] update gpt-oss package version (#56783)

liuzhenwei committed about 1 month ago

[ROCm][CI] Prepare TheRock image for CI (#56921)

Matt committed about 1 month ago

[Minor] Invert MRV1 specdec method fallback logic (#56899)

Nick Hill committed about 1 month ago

[Bugfix] Validate routed-expert prompt offsets before engine submission (#56844)

aoshen02 committed about 1 month ago

[ROCm][DSv4.1][Perf] Dequantize the MXFP8 weight once when dot_scaled cannot be used (#56560)

JohnQinAMD committed about 1 month ago

[Bugfix][Rust] Fix HF multi-turn dataset integration (#56915)

Kevin H. Luu committed about 1 month ago

[ROCm][Bugfix] Initialize Ray NIXL agents for sharded RDT (#56687)

Andreas Karatzas committed about 1 month ago

[Bugfix] Carry over queued work when materializing the dedicated stream (#56382)

zhaoguochun1995 committed about 1 month ago