vllm/vllm at a15f86ecfae5bbf9130991c7543f8a2034052f42 - vllm

assets

[Frontend] Remove librosa from audio dependency (#37058 )

2026-03-21 11:36:15 +08:00

benchmarks

[Misc] Reorganize inputs (#35182 )

2026-03-25 10:22:54 -07:00

compilation

[BugFix] fix VLLM_USE_STANDALONE_COMPILE=0 (#38015 )

2026-03-24 19:08:26 +00:00

config

[Misc] Rename think_start_str/think_end_str to reasoning_start_str/reasoning_end_str (#38242 )

2026-04-01 12:10:53 -07:00

device_allocator

[CI] Fix mypy for vllm/device allocator (#35518 )

2026-03-02 15:53:18 +00:00

distributed

[Feat][v1] Simple yet General CPU KV Cache Offloading (#37160 )

2026-04-01 01:03:14 -07:00

engine

Add short flag -sc for --speculative-config argument (#38380 )

2026-03-27 12:04:22 -07:00

entrypoints

[BugFix][Frontend] apply task instruction as system prompt in cohere v2/embed (#38362 )

2026-03-28 18:30:54 +00:00

inputs

[Misc] Reorganize inputs (#35182 )

2026-03-25 10:22:54 -07:00

kernels

[Kernel][Helion] [16/N] Refactor register_kernel API to be more Dynamo-friendly (#36705 )

2026-03-18 01:23:35 +00:00

logging_utils

[Logging] add --disable-access-log-for-endpoints CLI option (#30011 )

2026-01-26 21:49:03 +00:00

lora

[Bugfix][ROCm][MoE] Fix mxfp4 oracle regressions from #37128 (#37787 )

2026-03-25 08:17:33 +08:00

model_executor

feat(models): implement Google Gemma 4 architecture support (MoE, Multimodal, Reasoning, Tool-Use) (#38826 )

2026-04-02 11:49:53 -07:00

multimodal

(security) Enforce frame limit in VideoMediaIO (#38636 )

2026-04-01 12:10:40 -07:00

parser

[Misc] Use logger.info_once for auto tool choice log message (#37661 )

2026-03-20 10:40:36 +00:00

platforms

[ROCm][CI] Fix cross-attention dispatch for encoder-decoder models (#38450 )

2026-03-28 22:08:03 -07:00

plugins

[Misc] Reorganize inputs (#35182 )

2026-03-25 10:22:54 -07:00

profiler

Add PyTorch profiler schedule support with warmup/active iterations (#35240 )

2026-03-04 12:53:38 -08:00

ray

[Ray] Propagate third-party env vars to Ray workers via prefix matching (#34383 )

2026-02-17 01:08:42 -08:00

reasoning

feat(models): implement Google Gemma 4 architecture support (MoE, Multimodal, Reasoning, Tool-Use) (#38826 )

2026-04-02 11:49:53 -07:00

renderers

[Bugfix] Offload blocking tokenizer ops to shared thread pool to unblock event loop (#34789 )

2026-03-26 22:17:00 -07:00

third_party

[FlashMLA] Update FlashMLA to expose new arguments (#32810 )

2026-01-21 22:02:39 -07:00

tokenizers

[Bugfix] Standardize custom HF Processor init (#37289 )

2026-03-17 15:38:55 +00:00

tool_parsers

[Bugfix]: Fix Gemma4ToolParser.__init__() missing tools parameter (#38847 )

2026-04-02 16:45:38 -07:00

tracing

[Feature] OTEL tracing during loading (#31162 )

2026-02-05 16:59:28 -08:00

transformers_utils

feat(models): implement Google Gemma 4 architecture support (MoE, Multimodal, Reasoning, Tool-Use) (#38826 )

2026-04-02 11:49:53 -07:00

triton_utils

Add XPU MLA Sparse backend for DeepSeek v3.2 (#33230 )

2026-03-11 19:19:15 +08:00

usage

[XPU] Support Intel XPU hardware information collection in usage stats (#37964 )

2026-03-24 10:29:17 -07:00

utils

Revert "[Bugfix] Restrict TRTLLM attention to SM100, fixing GB300 (SM103) hang (#38730 )"

2026-04-01 15:14:58 -07:00

v1

[Misc] Rename think_start_str/think_end_str to reasoning_start_str/reasoning_end_str (#38242 )

2026-04-01 12:10:53 -07:00

vllm_flash_attn

[Attention] FA4 integration (#32974 )

2026-03-01 23:44:57 +00:00

__init__.py

Remove bc-lint (#35274 )

2026-02-26 03:01:01 -08:00

_aiter_ops.py

[ROCm] Utilize persistent MLA kernel from AITER (#36574 )

2026-03-26 03:00:42 +08:00

_custom_ops.py

[Sparse24] [Deprecation] Remove Sparse24 CT integration and kernels (#36799 )

2026-03-23 16:03:29 -04:00

_oink_ops.py

[Perf] Add opt-in SM100 Oink RMSNorm custom-op path (#31828 )

2026-02-24 23:01:53 -08:00

_xpu_ops.py

[XPU] support MLA model on Intel GPU (#37143 )

2026-03-25 17:43:42 +08:00

beam_search.py

[Misc] Reorganize inputs (#35182 )

2026-03-25 10:22:54 -07:00

collect_env.py

[Refactor] Remove unused dead code (#33718 )

2026-02-03 21:25:11 -08:00

connections.py

[CI] Add retry with 4x backoff to HTTP fetches for transient failures (#37218 )

2026-03-19 19:00:20 +00:00

env_override.py

[Performance] Enable Triton autotuning disk cache by default (#37188 )

2026-03-19 17:36:28 -04:00

envs.py

[Feat][v1] Simple yet General CPU KV Cache Offloading (#37160 )

2026-04-01 01:03:14 -07:00

exceptions.py

[openapi server] log exception in exception handler(2/N) (#36201 )

2026-03-10 20:16:30 -07:00

forward_context.py

[V0 Deprecation] Deprecate virtual engine (#37195 )

2026-03-18 14:30:14 -07:00

logger.py

[Log] Log once in local node by default (#37568 )

2026-03-19 12:04:59 -07:00

logits_process.py

[Misc] Refactor tokenizer interface (#29693 )

2025-11-29 04:02:21 -08:00

logprobs.py

Make mypy opt-out instead of opt-in (#33205 )

2026-01-29 09:12:26 +00:00

model_inspection.py

[UX] Add vLLM model inspection view (#29450 )

2026-01-09 10:12:35 -07:00

outputs.py

[V0 Deprecation] Remove unused MM placeholders in request output (#34944 )

2026-02-20 06:19:23 -08:00

pooling_params.py

[Model] Deprecate the score task (this will not affect users). (#37537 )

2026-03-20 08:07:56 +00:00

py.typed

…

sampling_params.py

fix(security): Add VLLM_MAX_N_SEQUENCES environment variable and enforce limit (#37952 )

2026-03-27 09:02:10 -04:00

scalar_type.py

Update Optional[x] -> x | None and Union[x, y] to x | y (#26633 )

2025-10-12 09:51:31 -07:00

scripts.py

Convert formatting to use ruff instead of yapf + isort (#26247 )

2025-10-05 07:06:22 -07:00

sequence.py

[Model Runner V2][Minor] Simplify PP logic (#38031 )

2026-03-24 13:57:17 -07:00

tasks.py

[Model] Deprecate the score task (this will not affect users). (#37537 )

2026-03-20 08:07:56 +00:00

version.py

Convert formatting to use ruff instead of yapf + isort (#26247 )

2025-10-05 07:06:22 -07:00