vllm-project/vllm

Pull Requests Commits

[Core] [Bugfix] Add Input Embeddings (#15428)

qthequartermasterman committed 251 days ago

Verified cc2a77d7

[Bugifx] Remove TritonPlaceholder from sys.modules (#17317)

Isotr0py committed 251 days ago

Verified 9e2de9b9

Add `pt_load_map_location` to allow loading to cuda (#16869)

jerryzh168 committed 251 days ago

Verified 109e15a3

Fix PixtralHF missing spatial_merge_size (#17571)

mgoin committed 251 days ago

Verified f192ca90

[Misc] Continue refactoring model tests (#17573)

DarkLight1337 committed 251 days ago

Verified f89d0e11

Check if bitblas is installed during support check (#17572)

mgoin committed 251 days ago

Verified b4003d11

[CI] Actually run tests/kv_transfer/test_disagg.py in CI (#17555)

mgoin committed 251 days ago

Verified 292fc59d

[Attention] MLA move o_proj q_proj into cuda-graph region (#17484)

LucasWilkinson committed 251 days ago

Verified afcb3f88

[Doc] note that not all unit tests pass on CPU platforms (#17554)

davidxia committed 251 days ago

Verified afb12e42

Implement Jump-Forward (Fast-Forwrd) Decoding

WoosukKwon committed 251 days ago

eeb5761c

[Bugfix] Disable gptq_bitblas for <SM80 to fix GPTQ on V100/T4 (#17541)

mgoin committed 251 days ago

Verified 24aebae1

[V1][Spec Decode] Apply torch.compile & cudagraph to EAGLE3 (#17504)

zixi-qi committed 251 days ago

Verified 39c0813a

[Misc][Tools][Benchmark] Publish script to auto tune server parameters (#17207)

Chenyaaang committed 251 days ago

Verified 9b70e2b4

[Bug]change the position of cuda_graph_sizes in dataclasses (#17548)

CXIAAAAA committed 251 days ago

Verified 173daac1

Remove duplicate code from dbrx.py (#17550)

sstamenk committed 251 days ago

Verified 04f2cfc8

[ROCM] Add gfx950 to the custom attention archs (#16034)

jpvillam-amd committed 251 days ago

Verified 811a6c09

[Bugfix] Fix lint error (#17547)

DarkLight1337 committed 251 days ago

Verified 9b1769dd

[Misc]add configurable cuda graph size (#17201)

CXIAAAAA committed 251 days ago

Verified 61c299f8

[ROCm] update installation guide to include build aiter from source instructions (#17542)

hongxiayang committed 251 days ago

Verified 4acfa335

[Model] Refactor Ovis2 to support original tokenizer (#17537)

Isotr0py committed 251 days ago

Verified 88c83041

Move the last arguments in `arg_utils.py` to be in their final groups (#17531)

hmellor committed 251 days ago

Verified 6768ff4a

[CI/Build] Remove `awscli` dependency (#17532)

DarkLight1337 committed 251 days ago

Verified f2e7af9b

[Misc] refactor example - cpu_offload_lmcache (#17460)

reidliu41 committed 251 days ago

Verified 7423cf0a

[torch.compile] Add torch inductor pass for fusing silu_and_mul with subsequent scaled_fp8_quant operations (#10867)

SageMoore committed 251 days ago

Verified 460a2b11

[ROCm] remove unsupported archs from rocm triton flash-attention supported list (#17536)

hongxiayang committed 251 days ago

Verified 28566d73

[Feature][Frontend]: Deprecate --enable-reasoning (#17452)

chaunceyjiang committed 252 days ago

Verified 98060b00

[FEAT] [ROCm]: Add Qwen/Qwen3-235B-A22B-FP8 TP4 triton fused moe config (#17535)

tjtanaa committed 252 days ago

Verified f5a3c655

[doc] add streamlit integration (#17522)

reidliu41 committed 252 days ago

Verified 7169f87a

Fix more broken speculative decode tests (#17450)

huydhn committed 252 days ago

Verified b74d888c

[FEAT] [ROCm]: Add Qwen/Qwen3-30B-A3B-FP8 fused moe config for MI300X (#17530)

tjtanaa committed 252 days ago

Verified 2007d4d5

Newer Older