vLLM/vllm at ed29dfae6e76235e01fcc0eca21cf0dcbfc40af1 - vLLM

assets

[Bugfix] Align OpenCV video metadata timeline (#47099 )

2026-06-30 20:43:42 +00:00

benchmarks

[Bugfix] Reject 0 or non-positive max concurrency (#54887 )

2026-09-04 15:59:29 -04:00

compilation

[Kernel] Remove unused fake implementation (#55535 )

2026-09-06 17:54:32 -07:00

config

[Qwen3.8-Flash-Next] Remove torch.compile for NVIDIA implementation (#55272 )

2026-09-06 21:57:56 -07:00

cute_utils

[Kimi-K3] Add GEMM-RS for sequence parallelism (#52079 )

2026-08-13 09:02:28 -07:00

device_allocator

[Core] Release NCCL communicator memory in sleep mode (#51485 )

2026-09-01 14:31:16 +08:00

distributed

[Bugfix][KV Offload] Stop offloading the final sampled token's KV slot (#54288 )

2026-09-06 15:22:19 +03:00

engine

[Core] Sync DP state on the first step of a wave (#52957 )

2026-09-06 02:16:58 +00:00

entrypoints

[Frontend] Expose multimodal metadata for disaggregated prefill (#54659 )

2026-09-06 14:46:52 +00:00

inputs

[Feat][MM Hashing] include media_io_kwargs in multi-modal hashes (#54241 )

2026-09-01 13:17:01 +00:00

ir

[CI] Use explicit devices in IR tests (#49513 )

2026-07-23 20:08:14 -07:00

kernels

[Kernel] Remove unused fake implementation (#55535 )

2026-09-06 17:54:32 -07:00

logging_utils

…

lora

[Kernel] Remove unused fake implementation (#55535 )

2026-09-06 17:54:32 -07:00

model_executor

[XPU] Use fused_input_norm kernel in FusedInputNorm (#52945 )

2026-09-07 13:44:31 +08:00

models

[Qwen3.8-Flash-Next] Remove torch.compile for NVIDIA implementation (#55272 )

2026-09-06 21:57:56 -07:00

multimodal

[Bugfix][Multimodal] Bound renderer warmup to the prefill token budget (#55448 )

2026-09-05 16:43:28 +00:00

parser

[Bugfix] Implicitly close DeepSeek DSML parameters (#54838 )

2026-09-02 14:13:36 +08:00

platforms

[Model] add GLM-5.3-Flash support (#53906 )

2026-09-03 09:40:35 -07:00

plugins

[MISC] Cleanup deprecated parameters (#53559 )

2026-08-24 10:19:04 -07:00

profiler

[Profiler] Fix start_profile permanently no-op after max_iterations auto-stop (#51839 )

2026-08-25 08:13:29 -04:00

ray

Propagate Flash Attention cache configuration to Ray workers (#49177 )

2026-07-21 07:47:53 +00:00

reasoning

feat: Add support for reasoning_token_count to reasoning parser (#54982 )

2026-09-03 05:02:49 -07:00

renderers

[Bugfix] Gracefully handle unsupported reasoning_effort in chat templates (#54022 )

2026-09-06 21:11:35 -07:00

third_party

[Kernel] Warm up Qwen GDN gated RMSNorm (#54251 )

2026-09-03 18:19:48 +08:00

tilelang_utils

[ROCm] Defer tilelang import through its import from vllm.tilelang_utils import tilelang and relaxed has_tilelang (#51159 )

2026-08-13 13:12:40 -05:00

tokenizers

[New model][Multimodal] Add DeepSeek-V4-Flash-Vision-Exp support (#54566 )

2026-09-02 11:09:09 -07:00

tool_parsers

[Model] Add K2-Horizon model support (#55063 )

2026-09-03 01:17:08 -07:00

tracing

…

transformers_utils

[Model] add GLM-5.3-Flash support (#53906 )

2026-09-03 09:40:35 -07:00

triton_utils

[Bugfix][CPU] Fix several bugs (#54042 )

2026-08-31 15:33:06 +08:00

usage

[Bugfix] Close usage telemetry HTTP sessions (#51219 )

2026-08-08 00:29:36 +00:00

utils

[Kernel] Remove unused fake implementation (#55535 )

2026-09-06 17:54:32 -07:00

v1

[Bugfix] Fix cuda profiler missing bug (#55237 )

2026-09-07 03:15:57 +00:00

vllm_flash_attn

Revert "[Kernel] Gemma-4 FA4 FP8 Kernel" (#52987 )

2026-08-19 10:48:46 -07:00

__init__.py

…

_aiter_ops.py

[ROCm][Perf][DeepSeek V4] Fuse native FP8 shared expert with MXFP4 routed experts (#53161 )

2026-09-06 19:42:21 -07:00

_custom_ops.py

[Kernel] Remove unused fake implementation (#55535 )

2026-09-06 17:54:32 -07:00

_xpu_ops.py

[XPU] Use fused_input_norm kernel in FusedInputNorm (#52945 )

2026-09-07 13:44:31 +08:00

collect_env.py

[XPU] fix collecting oneccl version info (#47104 )

2026-08-04 11:53:28 +00:00

connections.py

Reject oversized media before fully downloading it (#51896 )

2026-08-24 01:15:09 +00:00

env_override.py

[Perf] Add VLLM_TRITON_FORCE_FIRST_CONFIG to skip Triton autotuning (#42425 )

2026-06-16 15:16:45 +02:00

envs.py

[Frontend] Gate scale-out endpoints behind opt-in flag (#54579 )

2026-09-02 14:31:14 +08:00

exceptions.py

[Core] Add max_num_queued_reqs and max_num_queued_tokens for queue size management (#49445 )

2026-08-31 14:26:49 +00:00

forward_context.py

[Feature] Support sequence parallel without the need for DP, 1.9%~5.0% E2E Throughput Improvement (#47070 )

2026-07-05 05:41:30 -07:00

logger.py

…

logits_process.py

…

logprobs.py

[Bugfix] Handle empty FlatLogprobs slices and delta output (#53704 )

2026-08-26 21:00:54 +08:00

model_inspection.py

…

outputs.py

[Frontend][Core][Spec Decode] Per-request acceptance stats in OpenAI API responses (#48915 )

2026-08-20 15:42:49 -07:00

pooling_params.py

[MISC] Cleanup deprecated parameters (#53559 )

2026-08-24 10:19:04 -07:00

py.typed

…

sampling_params.py

[Bugfix] Reject empty bad-word tokenizations (#53433 )

2026-08-31 12:07:17 -07:00

scalar_type.py

Humming support for 2/3/5/6/7-bit pack-quantized weight-only inference (#46389 )

2026-06-24 13:53:54 -04:00

scripts.py

…

sequence.py

…

tasks.py

[Frontend] Move api_server.py out openai folder (#52131 )

2026-08-19 09:43:46 +00:00

version.py

…