Commits vllm-project/vllm

[MISC] typo fix (#19672)

andyxning committed 198 days ago

Verified 26bc46ef

[TPU] support attention head dim smaller than 128 (#19620)

yaochengji committed 198 days ago

Verified a77aea59

[Misc] Fix skipped max-model-len validation when deriving max model length from tokenizer config (#19660)

yeqcharlotte committed 198 days ago

Verified b692e9cd

[Misc][Frontend] passthrough `bad_words` (#19564)

f14-bertolotti committed 198 days ago

Verified 367871a4

[Bugfix][Core] Prefix caching causes incorrect outputs due to outdated ComputedBlocksTracker (#18957)

quanliu1991 committed 198 days ago

Verified 92183b41

[MISC] Remove unused variableds in C++ (#19609)

houseroad committed 198 days ago

Verified c6703d1e

[Misc] Remove duplicate multiproc method setting for CPU platform (#19649)

Isotr0py committed 198 days ago

Verified a5e7242d

[CI/Build] Fix torch nightly CI dependencies part 2 (#19589)

zou3519 committed 198 days ago

Verified 91b2c17a

Enable prefix caching with full cuda graphs (#19617)

WoosukKwon committed 199 days ago

Verified 055915e6

[Benchmark] Refactor benchmark script for fp8 & int8 (#19627)

yewentao256 committed 199 days ago

Verified 3d330c4c

[Kernel] Raise verbose error and consolidate `num_heads/num_kv_heads` divisibility check (#19339)

22quinn committed 199 days ago

Verified 0b73736a

[Bugfix][2/n] Fix speculative decoding CI - Fix test_ngram_e2e_greedy_correctness (#19644)

houseroad committed 199 days ago

Verified ee1531bc

[Perf] Further tunings for SM100 FP8 CUTLASS kernel (#19566)

ilmarkov committed 199 days ago

Verified e13945f9

[Fix] Convert kv_transfer_config from dict to KVTransferConfig (#19262)

maobaolong committed 199 days ago

Verified 08500011

[Bugfix] Don't attempt to use triton if no driver is active (#19561)

kzawora-intel committed 199 days ago

Verified 861a0a0a

Only build CUTLASS MoE kernels on Hopper (#19648)

huydhn committed 199 days ago

Verified bc956b38

[Hardware][NVIDIA][kernel] Fp4 MOE quant kernel optimization (#19500)

jiahanc committed 199 days ago

Verified 294fc1e2

[Bugfix] Fix auto dtype casting for BatchFeature (#19316)

Isotr0py committed 199 days ago

Verified 2db9044a

[Misc] Modularize CLI Argument Parsing in Benchmark Scripts (#19593)

reidliu41 committed 200 days ago

Verified 6fa718a4

[Bugfix] Fix the speculative decoding test by setting the target dtype (#19633)

houseroad committed 200 days ago

Verified 06be8588

[V1][Metrics] Deprecate metrics with gpu_ prefix for non GPU specific metrics. (#18354)

sahelib25 committed 200 days ago

Verified d1e34cc9

[BugFix] Fix DP Coordinator incorrect debug log message (#19624)

njhill committed 200 days ago

Verified bd517eb9

Adding "AMD: Multi-step Tests" to amdproduction. (#19508)

Concurrensee committed 200 days ago

Verified d65668b4

[torch.compile] Use custom ops when use_inductor=False (#19618)

WoosukKwon committed 200 days ago

Verified aafbbd98

[Doc] Add troubleshooting section to k8s deployment (#19377)

annapendleton committed 200 days ago

Verified 0f087451

[CUDA] Enable full cudagraph for FlashMLA (#18581)

ProExpertProg committed 200 days ago

Verified 3597b06a

[doc][mkdocs] fix the duplicate Supported features sections in GPU docs (#19606)

reidliu41 committed 200 days ago

Verified 1015296b

[Refactor] Remove unused variables in `moe_permute_unpermute_kernel.inl` (#19573)

yewentao256 committed 200 days ago

Verified ce9dc02c

[Model] Fix minimax model cache & lm_head precision (#19592)

qscqesze committed 200 days ago

Verified a24cb916

[BugFix] Honor `enable_caching` in connector-delayed kvcache load case (#19435)

njhill committed 200 days ago

Verified 7e8d97dd