[KV Offload] Support packed HMA KV cache layout (#46205)

Lucas Wilkinson committed 4 months ago

[Bugfix] Move extract_layer_index back inside is_v32 guard (#46199)

Tyler Michael Smith committed 4 months ago

[Frontend] Report cache usage in Anthropic /v1/messages API (#40912)

shuoming zhang committed 4 months ago

[ROCm] [Bugfix] Bugfix ROCm Sparse Indexer (#46222)

TJian committed 4 months ago

Stop setting CUDA_VISIBLE_DEVICES internally in vLLM, add device_ids arg (#45026)

Tyler Michael Smith committed 4 months ago

[Bugfix] Guard model_config access in _log_compilation_config (#46198)

Tyler Michael Smith committed 4 months ago

[Perf] Optimize Qwen3-VL multi-video prompt processing (#46026)

L丶 committed 4 months ago

Revert "Fix Stale Encoder Cache After Weight Update" (#46125)

Sumanth R Hegde committed 4 months ago

[Hardware][AMD][CI] Fix e2e core test group (#46024)

Matt committed 4 months ago

[Bugfix] Fix gridDim.y overflow for large row counts (#45255)

JasonLi314 committed 4 months ago

[ROCm] Fix VRAM not freed in test_phi3v (#46046)

djramic committed 4 months ago

[ROCm][CI] Skip Qwen3.5-35B-A3B-MXFP4-AITER-TP2 for non gfx950 (#46109)

Charlie Fu committed 4 months ago

[Bugfix][Parser] Fix U+FFFD leak at reasoning-to-content transition in engine parsers (#46159)

Ben Browning committed 4 months ago

[ROCm][CI] Pin `test_rocm_compressed_tensors_w8a8` to TRITON_ATTN (#46180)

Micah Williamson committed 4 months ago

Move CI failure diagnosis docs into ci-fails-buildkite skill (#45975)

Vadim Gimpelson committed 4 months ago

[ROCm] Use vLLM's fp8 quant max in AITER hipBLASLt accuracy test (#46176)

djramic committed 4 months ago

Fix dead link in docs (#46181)

Harry Mellor committed 4 months ago

[Test] Migrate test_openai_schema.py to schemathesis 4.x (#46173)

Ben Browning committed 4 months ago

[DSv4] Pack KV caches into contiguous per-block allocations for DeepSeek V4 (#44577)

Tyler Michael Smith committed 4 months ago

[12/n] final _C library kernel migration (#45415)

Chris Leonard committed 4 months ago