[chore] log process manager shutdown with more details (#49437)

Ning Xie committed 3 months ago

[chore] delete useless code (#49424)

Ning Xie committed 3 months ago

Bump Helion to 1.4.0 (#50307)

Shangdi Yu committed 3 months ago

[Bugfix][TurboQuant] Add KV quant mode for turboquant (#50533)

Soila Kavulya committed 3 months ago

[ROCm][Quark][7/N] Use MXFP4 linear kernel abstraction for `emulation` backend (#48949)

fxmarty-amd committed 3 months ago

[Compressed-Tensors] Support Kimi-K3 quantized models (#50500)

Kyle Sayers committed 3 months ago

[ROCm][CI] Restore Mistral tool-parser compatibility after unification (#50515)

Andreas Karatzas committed 3 months ago

Upgrade tpu-inference to v0.26.0 (#50522)

Dennis Yeh committed 3 months ago

Update torch version to 2.13.0+cpu (#50412)

ylangtsou committed 3 months ago

[Doc] Add BgeM3EmbeddingModel to embedding supported models (#50571)

LG committed 3 months ago

[Bugfix] Universally align block table width to 128 tokens (#50302)

Matthew Bonanni committed 3 months ago

[Bugfix] Don't transpose fused MoE quantization scales in `RoutedExperts.load_weights` (#50137)

Harry Mellor committed 3 months ago

[DSv4 Perf] Optimize workspace reuse for eager break, 3.9% E2E TTFT improvement. (#49236)

Wentao Ye committed 3 months ago

[ROCm][CI] Fall back to lossless Kimi K3 MXFP4 emulation on gfx942 (#50516)

Andreas Karatzas committed 3 months ago

[2/N][Attention] Enable masked MHA for sparse MLA prefills (#48770)

Matthew Bonanni committed 3 months ago

[Kernel][CI] `--jit-monitor-mode error` e2e tests for kernel warmup infra (#50109)

Nicolò Lucchesi committed 3 months ago

[CI] Remove default_torch_num_threads workaround from llava-onevision-transformers test (#50560)

Oğuzhan KIR committed 3 months ago

[Bugfix][Frontend] Raise VLLMValidationError for user-facing errors in chat_utils.py (#50491)

Recoordinate committed 3 months ago

[Renderer] Warm up the renderer properly. (#50408)

wang.yuqi committed 3 months ago

[KV Connector] Add per-layer canonical KV page mappings for parallelism-agnostic offload (#48408)

Itay Etelis committed 3 months ago