vllm-project/vllm

Pull Requests Commits

[BugFix] Allow use_cudagraph to work with dynamic VLLM_USE_V1 (#19390)

zou3519 committed 207 days ago

Verified 77f0d465

Add H20-3e fused MoE kernel tuning configs for Qwen3-235B-A22B-FP8 (#19401)

Xu-Wenqing committed 207 days ago

Verified 22c3c0aa

[Model] use AutoWeightsLoader for commandr (#19399)

py-andy-c committed 207 days ago

Verified 33f8dba7

[ROCm][V1] Adding ROCm to the list of plaforms using V1 by default (#19440)

gshtras committed 207 days ago

Verified 5241ca50

[Docs] Note that alternative structured output backends are supported (#19426)

russellb committed 207 days ago

Verified da9b523c

[Misc] Slight improvement of the BNB (#19418)

jeejeelee committed 208 days ago

Verified b6553be1

Simplify ep kernels installation (#19412)

youkaichao committed 208 days ago

Verified 64a9af5a

[BugFix][CPU] Fix CPU CI by ignore collecting test_pixtral (#19411)

bigPYJ1151 committed 208 days ago

Verified e4248849

[BugFix][FlashInfer] Fix attention backend interface mismatch with unexpected keyword `use_irope` (#19134)

YUNQIUGUO committed 208 days ago

Verified 467bef18

Revert "[v1] Add fp32 support to v1 engine through flex attn" (#19404)

Isotr0py committed 208 days ago

Verified 5f1ac1e1

Automatically bind CPU OMP Threads of a rank to CPU ids of a NUMA node. (#17930)

louie-tsai committed 208 days ago

Verified 9368cc90

Add clear documentation around the impact of debugging flag (#19369)

annapendleton committed 208 days ago

Verified 32b3946b

[Misc] refactor neuron_multimodal and profiling (#19397)

reidliu41 committed 208 days ago

Verified 6b1391ca

Add security warning to bug report template (#19365)

russellb committed 208 days ago

Verified a3f66e75

[Core] Batch multi modal input using pinned memory (#19169)

lgeiger committed 208 days ago

Verified 319cb1e3

[Bugfix] Fix modelscope token passed in (#19389)

Potabk committed 208 days ago

Verified 1efef716

[Core] Use tuple for kv cache group block ids (#19175)

njhill committed 208 days ago

Verified 646d62f6

[Frontend] Add tqdm_leave_pbar to control progress bar visibility (#19357)

reidliu41 committed 208 days ago

Verified 6cd4ae8a

Fix docs/mkdocs/hooks/remove_announcement.py (#19382)

hmellor committed 208 days ago

Verified c016047e

Use xla flag to improve the quantized model performance (#19303)

vanbasten23 committed 208 days ago

Verified 9af6d22e

[Bugfix] Fix benchmark_moe.py (#19016)

gty111 committed 208 days ago

Verified 4589b940

[V1] Reuse V0's memory_profiling util for gpu worker memory profiling (#19312)

yeqcharlotte committed 208 days ago

Verified cc867be1

[Misc] Fix a config typo in disable_hybrid_kv_cache_manager configuration (#19383)

lsy323 committed 208 days ago

Verified 3a7cd627

[HOT-FIX] Add `kv_sharing_target_layer_name` argument to cutlass_mla backend (#19374)

pavanimajety committed 208 days ago

Verified 8058c911

[TPU]Fix KV cache sharing tests (#19371)

lsy323 committed 208 days ago

Verified 7d44c469

[Frontend] Make TIMEOUT_KEEP_ALIVE configurable through env var (#18472)

liusiqian-tal committed 208 days ago

Verified 31f58be9

[Quantization] Bump compressed-tensors version (#19295)

kylesayrs committed 208 days ago

Verified ebb2f383

[Bugfix][Core] Prevent token lengths exceeding `max_model_len` in V0 (#19348)

22quinn committed 209 days ago

Verified c1c7dbbe

[Misc] Fixes and Optimizations for DeepEP + DeepGEMM combination. (#19298)

varun-sundar-rabindranath committed 209 days ago

Verified 5cf2daea

[v1] Add fp32 support to v1 engine through flex attn (#19319)

Isotr0py committed 209 days ago

Verified b8089195

Newer Older