[ROCm][Compile] Fuse AITER static FP8 attention output (#58099)

Andreas Karatzas committed 29 days ago

[Refactor][Quantization] Make FP8 and MLA weight transforms reusable pure functions (#57732)

aoshen02 committed 29 days ago

[Mooncake] Address review nits from #56855 (#57174)

Nicolò Lucchesi committed 29 days ago

[ROCm][CI] Add MI355 dense NVFP4 and MoRI kernel mirrors (#58281)

Andreas Karatzas committed 29 days ago

[ROCm][CI] Mirror the three TurboQuant evaluation groups on MI355 (#58282)

Andreas Karatzas committed 29 days ago

[Quantization] Select per-token NVFP4 MoE backends explicitly (#57176)

Matej Sirovatka committed 29 days ago

[Tests] Select V2 for diffusion scheduler unit tests (#58272)

Andreas Karatzas committed 29 days ago

[ROCm][Perf] Use wvSplitK for single-output GEMMs (#53283)

00CC committed 29 days ago

[Bugfix][Frontend] Keep length finish_reason for max_tokens-truncated streaming tool calls (#46303)

Ting SUN committed 29 days ago

[MoE] Use GateLinear for all MoE models (#58234)

Jee Jee Li committed 29 days ago

[XPU][CI] Deselect tests/v1/spec_decode/test_mtp.py::test_glm_mtp_defers_lm_head (#58237)

xiangdong committed 29 days ago

[Spec decode] Support variable-length decode for Kimi-K3 adaptive ver (#52988)

Albert Cheng committed 29 days ago

[Compilation] Fix QuTLASS compilation with PyTorch 2.13 (#58173)

Wentao Ye committed 29 days ago

[XPU] Wire up SYCL apply_rotary_emb kernel in ApplyRotaryEmb (#55721)

Michał Ganczarenko committed 29 days ago

[Bugfix] Set worker runtime threads before profiling and compilation (#55891)

Andreas Karatzas committed 29 days ago

[ROCm][CI] Stage G gating (#50922)

Andreas Karatzas committed 29 days ago

[CI][Bugfix] Update IPC test caller for #57312's _apply_entries signature (#58107)

vllm-agent committed 29 days ago

[ROCm][Compile] Support BF16 AsyncTP fusion (#58098)

Andreas Karatzas committed 29 days ago

[ROCm][DSv4][Perf] Fuse the inverse RoPE into the sparse decode reduce (#57451)

Fangzhou Ai committed 29 days ago

[EPD] Support metadata-only audio inputs (#57887)

Tianyu Guo committed 29 days ago