[ROCm] Add QuickReduce min-size override and codec threshold (#41675)

akii96 committed 6 months ago

[Misc] downgrade nvidia-cutlass-dsl to 4.5.0 (#43230)

Jiangyun Zhu committed 6 months ago

[Perf][gpt-oss] Downgrade triton_kernels to v3.5.1 (#43135)

Michael Goin committed 6 months ago

Add dllehr-amd to CODEOWNERS and committers list (#42772)

Douglas Lehr committed 6 months ago

[Bugfix][CI] Add missing import of pad_nvfp4_activation_for_cutlass in flashinfer (#43237)

Flora Feng committed 6 months ago

[Perf] Optimize `CutlassFP8ScaledMMLinearKernel` when padding needed by pre-weight processing, 13.5% TTFT improvement (#42651)

Wentao Ye committed 6 months ago

Integrate flashinfer b12x MoE and FP4 GEMM kernels for SM120/121 (#40082)

meena-at-work committed 6 months ago

[CI] Lower granite-4.0-h-tiny gsm8k threshold for Hybrid SSM NixlConnector PD accuracy tests (4 GPUs) (#43186)

haosdent committed 6 months ago

[R3] Add routed experts to openai entrypoint (#38939)

Aaron Hao committed 6 months ago

[MISC] Fix symm_mem cap-equal gate; log AR backend selection (#42993)

Vadim Gimpelson committed 6 months ago

[Bugfix] Use shared coerce_to_schema_type in DeepSeekV32 tool parser (#43019)

Flora Feng committed 6 months ago

[Bugfix] Use enable_sm120_family for per-tensor FP8 CUTLASS kernels on SM12.1 (#41215)

Joel Smith committed 6 months ago

[Bug][Structured Outputs] Fix bug that leads to unconstrained generations with structural tags (#42452)

rishitdholakia13 committed 6 months ago

Remove additional dead code as a follow-up to #42889 (#43144)

Dipika Sikka committed 6 months ago

[Spec Decode] Support non-MTP speculation for NemotronH (#43130)

Benjamin Chislett committed 6 months ago

[Feature] Support manually enabling the cumem allocator (#33648)

Kebe committed 6 months ago

[Doc] Sync CLI guide with actual help modes and launch subcommand (#40326)

Ray Wang committed 6 months ago

[Frontend] Forward X-data-parallel-rank header on /inference/v1/generate (#42330)

hallerite committed 6 months ago

[XPU][CI] Add 2 server model test files in Intel GPU CI (#42499)

xiangdong committed 6 months ago

[GDN] Enable FI Blackwell GDN prefill kernel (#40717)

Artem Perevedentsev committed 6 months ago