Almog TavorandOr Ozeri dc02934a07 [Bugfix][KV Offload] Stop offloading the final sampled token's KV slot (#54288) ...

Signed-off-by: almogtavor <almogtavor@gmail.com>
Signed-off-by: Almog Tavor <70065337+almogtavor@users.noreply.github.com>
Co-authored-by: Or Ozeri <or@ozery.com>

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

..

basic_correctness

[CI][Test] Deflake test_mem.py sleep-mode asserts via allocator bookeeping (#54312)

2026-08-29 12:05:17 -07:00

benchmarks

[Bugfix] Restore multimodal support on the plain "vllm" throughput backend (#52168)

2026-08-28 10:28:25 +00:00

compile

[Fusion] Manual ActivationQuantFusionPass initial application (#51415)

2026-09-03 13:37:31 -04:00

config

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

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

cuda

[Bugfix][Platform] Stop re-initializing NVML on every device-capability check (fixes #50381) (#50393)

2026-08-07 08:30:00 +00:00

detokenizer

[Bugfix][Spec Decode] Select earliest-completing stop string in check_stop_strings (#49391)

2026-07-23 18:14:06 +08:00

distributed

[Core][KV Events] Echo session_id on GPU BlockStored events (#51381)

2026-09-04 02:22:50 +00:00

engine

[Bugfix] Account for PCP in multi-node world size validation (#55111)

2026-09-03 10:52:42 +00:00

entrypoints

Validate scale-out multimodal data before engine handoff (#51898)

2026-09-06 04:58:04 +00:00

evals

[CI][ROCm] Restore Wikitext coverage for Qwen OCP-MX (#55410)

2026-09-05 10:50:54 +08:00

fusion

[Fusion] Manual ActivationQuantFusionPass initial application (#51415)

2026-09-03 13:37:31 -04:00

ir

[Perf] Set IR Op Priority Once at Worker Init (#42631)

2026-05-15 15:56:13 +00:00

jit_monitor

[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

kernels

[CPU] [Feat] Add native AMX-FP8 attention impl for Diamond Rapids (#49410)

2026-09-06 11:06:56 +08:00

lora

[Misc] Use VLLMValidationError in offline inference input validation (#53432)

2026-08-23 08:43:51 -07:00

model_executor

[Core][MRV2] Support eagle3 spec decode with pipeline parallel (#50514)

2026-09-05 11:45:34 +08:00

models

[Kimi K3] Support internal prefix checkpoints with partial prefix caching and spec-decoding (#53614)

2026-09-06 13:18:15 +08:00

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

[CPU] Support FP16/BF16 persisted GDN state on AMX (#52191)

2026-08-31 16:32:20 +08:00

plugins

[Frontend] Move engine/protocol.py out openai folder (#54492)

2026-08-31 10:13:50 +00:00

plugins_tests

[Frontend] Move cli_args.py and dp_supervisor.py out openai folder (#53659)

2026-08-25 07:02:32 +00:00

prompts

…

quantization

[CI][ROCm] Restore Wikitext coverage for Qwen OCP-MX (#55410)

2026-09-05 10:50:54 +08:00

reasoning

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

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

renderers

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

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

rocm

[Bugfix][MoE] Preserve unquantized weight storage on ROCm (#46009)

2026-09-03 11:06:09 +08:00

samplers

[ROCm] Remove stale SDPA and skinny GEMM workarounds (#50907)

2026-08-11 14:19:16 -04:00

spec_decode

[MyPy][1/N] Fix mypy errors in some tests/ directories and enforce follow-imports=silent (#49570)

2026-07-30 05:02:51 -07:00

standalone_tests

[ROCm][CI] Use the same-build wheel in Python-only CI (#49514)

2026-08-16 22:52:24 -05:00

system_messages

…

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

tool_use

[Frontend] Move engine/protocol.py out openai folder (#54492)

2026-08-31 10:13:50 +00:00

tools

[CI][ROCm] Avoid redundant image pulls during smoke validation (#54408)

2026-08-31 21:48:25 -05:00

tracing

[CI][Tracing] Fix race condition by adding server readiness check (#34364)

2026-02-16 12:57:39 +00:00

transformers_utils

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

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

utils_

[Frontend] Add site-packages support for reasoning/tool parser plugins (#45241)

2026-09-02 07:41:13 +00:00

v1

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

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

vllm_test_utils

…

weight_loading

[CI][ROCm] Prefetch safetensors weights in AMD CI (#54898)

2026-09-02 03:28:39 -05:00

__init__.py

…

ci_envs.py

…

conftest.py

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

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

test_access_log_filter.py

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

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

test_attention_backend_registry.py

[6/N][KV-Cache Layout Refactor] Standardize KV cache layout (#51718)

2026-08-21 17:47:49 -07:00

test_audio_media_size_precheck.py

fix(security): enforce VLLM_MAX_AUDIO_CLIP_FILESIZE_MB on all audio paths (#53561)

2026-08-24 21:37:02 +00:00

test_cmake_utils.py

[CI/Build] Add warning for unsupported global PTX architecture requests in... (#51901)

2026-08-15 07:01:52 -07:00

test_config.py

[Perf][Model Runner V2] Compact sampling masks on GPU instead of unpacking the full-vocab bitmask on CPU (#54901)

2026-09-03 19:03:36 -07:00

test_connections.py

Reject oversized media before fully downloading it (#51896)

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

test_embedded_commit.py

…

test_envs.py

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

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

test_force_first_config.py

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

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

test_fxgraphcache_pickle_patch.py

[ZenCPU] Make PT Backport Patch Accessible to vLLM (#38205)

2026-04-10 08:29:35 +00:00

test_logger.py

[Frontend] Consolidate online serving utils. (#44479)

2026-06-04 06:49:31 +00:00

test_logprobs.py

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

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

test_outputs.py

…

test_pooling_params.py

[Frontend][Core] Standardize request error handling with VLLMError hierarchy (#49665)

2026-07-29 04:20:40 +00:00

test_quantization_revision_pin.py

Preserve revision pins in secondary artifact loaders (#51446)

2026-08-11 04:33:10 +00:00

test_ray_env_utils.py

[Bugfix] Set breakable graph env before Ray actor import (#53293)

2026-08-28 23:11:50 +00:00

test_ray_env.py

fix nixl side-channel host selection (#41806)

2026-05-11 07:40:37 +00:00

test_regression.py

[CI/Build] Use modelscope's international site for regression test (#42176)

2026-05-09 19:47:09 -07:00

test_request_input_bounds.py

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

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

test_sampling_params.py

[Frontend][Core] Standardize request error handling with VLLMError hierarchy (#49665)

2026-07-29 04:20:40 +00:00

test_scalartype.py

…

test_sequence.py

…

test_tensorizer_revision_pin.py

Preserve revision pins in secondary artifact loaders (#51446)

2026-08-11 04:33:10 +00:00

test_triton_utils.py

Exclude the cpu backend from vLLM's active-Triton-driver count (#53530) (#53530)

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

test_version.py

…

test_vllm_port.py

…

test_zen_cpu_platform_detection.py

[Bugfix] Log platform plugin detection failures (#52285)

2026-09-01 19:11:52 +00:00

utils.py

[CI] Avoid logging test server environment values (#54379)

2026-09-03 17:00:29 -04:00