Commit Graph - vLLM - iHore Software Repository

iHore/vLLM

Fork 0

e746a2eebf [Model] Use AutoWeightsLoader for Voyage (#42972) Furkan F 2026-05-22 07:28:23 +02:00
1fe3303983 [CI] De-flake renderers/test_hf.py::test_resolve_content_format_fallbacks[Qwen/Qwen-VL-string] (#43064) haosdent 2026-05-22 12:15:22 +08:00
8c8b1825eb [XPU] Enable multiple key kernels for sparse attention (#37888) Xiaochang WuandKunshang Ji 2026-05-22 12:02:51 +08:00
18a27cc9a3 [Bugfix] Make CuMemAllocator free callback stream-aware (#43020) qizixiandClaude 2026-05-21 20:36:22 -07:00
0ddd7dd656 [Frontend] DP Supervisor (#40841)
Wentao Ye robertgshaw2-redhat Robert Shaw Nick Hill
2026-05-21 23:33:16 -04:00
60af5c16ee [Frontend] Add truncation side to OpenAI endpoints (#43260) ruizhangandRui Zhang 2026-05-21 20:32:31 -07:00
35d0141a0b [ROCm][CI] add warmup to mem_util test before measurement (#43236) Divakar Verma 2026-05-21 23:17:54 -04:00
86ccef7d44 [ROCm] Add XGMI backend for MoRI Connector (#41753) Simon Danielsson 2026-05-22 05:06:40 +02:00
2998a047aa [Bugfix] Fix DSV4 Base model swiglu limit issue in FP8 path (#42855) Chengze FanandLu Fang 2026-05-21 19:43:01 -07:00
ba369b7eb5 [CI] Fix dockerfile dependency graph failure for pre-commit (#43378) Isotr0py 2026-05-22 10:26:05 +08:00
39910f2b25 [Rust Frontend] Move code from vllm-frontend-rs (#43283)
Bugen Zhao Nick Hill Eric Curtin Dev-X25874 Will.hou Will.hou
2026-05-22 08:21:48 +08:00
39d5fa96a7 [Bugfix] Zero stale is_prefilling in padded CUDA graph rows for Mamba (#41873) Lanze Liu 2026-05-21 15:42:42 -07:00
565b745ec5 [BugFix] Use correct logprobs for logprob_token_ids (#43125) Nick Hill 2026-05-21 15:42:20 -07:00
e26e1f0928 [Feature] Add --cpu-distributed-timeout-seconds CLI Option for CPU Process Group Timeout (#42968)
fangyuchu zWaNg3mergify[bot] <37929162+mergify[bot]@users.noreply.github.com>
2026-05-22 06:42:07 +08:00
0f66623b0d [Frontend] Rework fastokens integration (#43168) Nick Hill 2026-05-21 15:36:58 -07:00
0b59fc45dd Disable build isolation to bypass CUDA related deps for vllm-tpu (#43038)
ylangtsou Ylang Tsou Michael Goin
2026-05-22 06:00:52 +08:00
17b69828a0 [Core] Add native ModelExpress load format (#43105)
Zheng Luo OpenAI Codex Robert Shaw
2026-05-21 13:05:01 -07:00
b29cbf0652 [Perf] zeros -> empty to remove additional fill (#42988) Wentao Ye 2026-05-21 16:00:29 -04:00
9b54e50e2c [Deprecation] Mark env vars covered by --moe-backend / --linear-backend (#43148) Michael Goin 2026-05-21 15:51:12 -04:00
1c78f76c29 [Bugfix] Add early validation to reject incompatible runner types for embedding models (#43079)
anish anish Claude Opus 4.6 Wentao Ye
2026-05-21 11:07:46 -04:00
9b9d5dbaab [CI] Fix CPU tests failing on tl.exp2 import (#43311) haosdent 2026-05-21 22:28:34 +08:00
b730c46352 [Perf] [Hybrid] Fused Triton kernel for GPU-side Mamba state postprocessing (#40172) Francesco Fuscoandmergify[bot] <37929162+mergify[bot]@users.noreply.github.com> 2026-05-21 13:50:54 +02:00
c68c55d43e [CPU][RISC-V] Add VLEN=256 support to RVV attention kernels (#42943)
velonica0gemini-code-assist[bot] <176961590+gemini-code-assist[bot]@users.noreply.github.com>Li, Jiang <jiang1.li@intel.com>
2026-05-21 19:50:49 +08:00
5ecd8e9c70 [XPU][CI]Fix Docker image pull-to-run race in Intel GPU CI (#43266) xiangdongandKunshang Ji 2026-05-21 18:41:38 +08:00
caf69823d6 [CI] Pin protoc binary in rust-build stages (#43292) haosdent 2026-05-21 18:38:07 +08:00
68e07d5916 [Bug] Fix ci issue assert output_size is not None AssertionError (#43261) Wentao YeandIsotr0py 2026-05-21 04:58:09 -04:00
ebbfb34e3e [Test] Replace zephyr-7b-beta (7B) with SmolLM2-135M in tokenization test (#43085) Kevin H. LuuandClaude Opus 4.6 2026-05-21 01:57:47 -07:00
edafea3555 Fix FlashInfer TRTLLM NvFP4 monolithic MoE routing (#43223) zhangxin81 2026-05-21 16:17:12 +08:00
b719b1635b Update KDA chunk prefill decay to use exp2 semantics (#43195) zexplorerhjandzexplorerhj 2026-05-21 16:16:27 +08:00
0a54df2847 [XPU] add setuptools-rust for xpu dependency (#43287) Kunshang Ji 2026-05-21 07:14:13 +00:00
a950e9447e [CI] De-flake test_models for bigscience/bloom-560m (#43197) haosdent 2026-05-21 14:30:14 +08:00
050611a3dd [Bugfix] Fix glm4_moe_tool_parser._is_string_type for /v1/responses FunctionTool format (#39601)
Yiyang "Ian" Liu Copilot Chauncey sfeng33
2026-05-20 22:58:59 -07:00
905b97adfa [Benchmark] Add num-warmup to vllm bench throughput (#43245) yzong-rh 2026-05-21 01:13:15 -04:00
a6682d1d25 [Bugfix] Warn when renderer_num_workers has no effect on offline LLM (#42905) Daoyuan Li 2026-05-20 21:35:08 -07:00
f2ace1d57d [Frontend][RFC] Rust front-end integration (#40848) Nick HillandBugen Zhao 2026-05-20 21:24:48 -07:00
d97ba29fdc [ToolParser][Bugfix] Re-land: Fix anyOf/oneOf/$ref type resolution in Qwen3CoderToolParser (#37831) (#38973)
손세정 세덩 sejung-son sfeng33
2026-05-21 13:24:08 +09:00
6441cf4a44 [Refactor] Use shared coerce_to_schema_type in Seed-OSS tool parser (#43140) Flora Feng 2026-05-21 00:24:06 -04:00
346cf163a1 [Frontend] Normalize reasoning_content to reasoning for client compatibility (#42664) Ben Browningandmergify[bot] <37929162+mergify[bot]@users.noreply.github.com> 2026-05-21 00:23:47 -04:00
7e5070934e [CI] Fix "test_vit_cudagraph_[image|video][step3_vl]" failure (#43082) haosdent 2026-05-21 12:22:10 +08:00
2b75a73b8e [Perf][Gemma4] Batch vision encoder calls for image and video processing (#43169) Luciano MartinsandLuciano Martins 2026-05-21 01:22:06 -03:00
e45df8c3f7 [Bugfix] Fix Qwen3.5 GatedDeltaNet in_proj_ba Marlin failure at TP>=2 (#36329)
sonusflow Isotr0py Claude Opus 4.7
2026-05-21 06:22:01 +02:00
ee05e8137e [Minor] Bigger overlap for FI AR (#43103) Jee Jee Li 2026-05-21 12:20:57 +08:00
5d041cc1fe update GPU json file based on h200 recipes (#43262) Louie Tsai 2026-05-20 20:57:48 -07:00
9640970de2 [Model Runner V2] Fix lora Triton Error [CUDA]: device-side assert triggered (#43139) Wentao YeandNick Hill 2026-05-20 21:00:30 -04:00
63ea11709b [CI] Add composed-schema regression tests for DeepSeek V3.2/V4 parsers (#43255) Ace EldeibandFlora Feng 2026-05-21 02:36:16 +02:00
bde560ed6e [ROCm] Add QuickReduce min-size override and codec threshold (#41675) akii96 2026-05-21 01:46:51 +03:00
6dc0a71843 [Misc] downgrade nvidia-cutlass-dsl to 4.5.0 (#43230) Jiangyun Zhu 2026-05-21 05:19:50 +08:00
5774aad9c5 [Perf][gpt-oss] Downgrade triton_kernels to v3.5.1 (#43135) Michael Goin 2026-05-20 17:13:12 -04:00
452baa860b Add dllehr-amd to CODEOWNERS and committers list (#42772) Douglas Lehr 2026-05-20 16:10:44 -05:00
2a43b407c5 [Bugfix][CI] Add missing import of pad_nvfp4_activation_for_cutlass in flashinfer (#43237) Flora Feng 2026-05-20 14:59:12 -04:00
53ff50fcd3 [Perf] Optimize CutlassFP8ScaledMMLinearKernel when padding needed by pre-weight processing, 13.5% TTFT improvement (#42651) Wentao YeandMatthew Bonanni 2026-05-20 14:57:42 -04:00
363fc84407 Integrate flashinfer b12x MoE and FP4 GEMM kernels for SM120/121 (#40082) meena-at-workandClaude Sonnet 4.6 2026-05-20 10:21:11 -07:00
f2d5e3d3ae [CI] Lower granite-4.0-h-tiny gsm8k threshold for Hybrid SSM NixlConnector PD accuracy tests (4 GPUs) (#43186) haosdentandNickLucche 2026-05-21 01:00:24 +08:00
2d6b3489b9 [R3] Add routed experts to openai entrypoint (#38939) Aaron HaoandIsotr0py 2026-05-20 09:07:59 -07:00
9c78c99995 [MISC] Fix symm_mem cap-equal gate; log AR backend selection (#42993) Vadim Gimpelson 2026-05-20 19:50:24 +04:00
a10d69116c [Bugfix] Use shared coerce_to_schema_type in DeepSeekV32 tool parser (#43019) Flora Feng 2026-05-20 10:21:00 -04:00
644b2a28e7 [Bugfix] Use enable_sm120_family for per-tensor FP8 CUTLASS kernels on SM12.1 (#41215) Joel SmithandShengqi Chen 2026-05-20 15:10:01 +01:00
ded871201a [Bug][Structured Outputs] Fix bug that leads to unconstrained generations with structural tags (#42452) rishitdholakia13andCursor 2026-05-20 10:08:58 -04:00
df84fb07a6 Remove additional dead code as a follow-up to #42889 (#43144) Dipika Sikka 2026-05-20 10:01:45 -04:00
0a508743d4 [Spec Decode] Support non-MTP speculation for NemotronH (#43130) Benjamin Chislett 2026-05-20 09:15:52 -04:00
19cf334207 [Feature] Support manually enabling the cumem allocator (#33648) Kebe 2026-05-20 21:58:30 +09:00
87e31455b0 [Doc] Sync CLI guide with actual help modes and launch subcommand (#40326) Ray WangandRui Wang 2026-05-20 02:32:03 -07:00
cb600d1cdb [Frontend] Forward X-data-parallel-rank header on /inference/v1/generate (#42330) halleriteandClaude Opus 4.7 2026-05-20 10:58:46 +02:00
6f21558da1 [XPU][CI] Add 2 server model test files in Intel GPU CI (#42499) xiangdong 2026-05-20 16:54:58 +08:00
1cb224430b [GDN] Enable FI Blackwell GDN prefill kernel (#40717) Artem Perevedentsev 2026-05-20 11:46:55 +03:00
9b343dd4f5 Enable mermaid diagrams in the docs (#43192) Harry Mellor 2026-05-20 17:10:00 +09:00
07aeaf9d4d [6/n] Migrate activation kernels, gptq, gguf, non cutlass w8a8 to libtorch stable ABI (continued) (#42663)
Chris Leonard Mikayla Gawarecki Shengqi Chen
2026-05-20 03:18:12 -04:00
40651c0207 [Docs][PD][NIXL] Bidirectional kv-cache transfer (#43097) Nicolò Lucchesi 2026-05-20 09:02:36 +02:00
7e4bc2cecb [Docs][PD][NIXL] Lease extension mechanism for blocks on P (#43099) Nicolò Lucchesi 2026-05-20 08:58:25 +02:00
85959567c3 [ci] Revert model executor test back to L4 (#43188) Kevin H. Luu 2026-05-19 23:01:41 -07:00
4f940896a3 [KV Offload] Pass OffloadingSpec instead of VllmConfig to secondary tiers (#43076) Ronen Schaffer 2026-05-20 06:32:08 +03:00
cd0ff26e7a [CI] Add DSV4-Flash to gsm8k moe-refactor/config-b200.txt (#42111) Michael Goin 2026-05-19 23:21:01 -04:00
2ae910ed88 [Perf] Avoid forward scan for async output placeholders (#42938) Izik Golan 2026-05-20 06:16:07 +03:00
fadf5d332c add enqueue all option to throughput benchmark (#42975)
pmaybankgemini-code-assist[bot] <176961590+gemini-code-assist[bot]@users.noreply.github.com>mergify[bot] <37929162+mergify[bot]@users.noreply.github.com>
2026-05-19 23:16:02 -04:00
c628a93a64 [Perf][Bugfix] Update dflash aux layer indexing (#40727) Benjamin Chislett 2026-05-19 23:15:57 -04:00
5774aaed0c [Cohere] Enable Cohere MoE (#43143) Terrence Zhao 2026-05-19 22:32:06 -04:00
39bba710be [MRV2][BugFix] Fix default-stream CG capture in P/W LoRA case (#43160) Nick Hill 2026-05-19 19:19:05 -07:00
73dd2f33b7 [bug] fix WeightTransferConfig.backend to allow for all strings (#43121) Aaron Hao 2026-05-19 18:01:29 -07:00
be16785998 [CPU][DOC] Fix installation commands for Arm CPUs (#43115) Fadi Arafeh 2026-05-20 00:31:15 +01:00
117afeea46 Fix error in Dynamic NTK scaling (#41277)
Max de Baysergemini-code-assist[bot] <176961590+gemini-code-assist[bot]@users.noreply.github.com>wang.yuqi
2026-05-19 17:27:54 -04:00
1242196295 [Model] Support post-norm architecture for EAGLE-3 supeculators (#42764) Doğaç Eldenk 2026-05-19 15:39:00 -05:00
a65093c1a3 [ci] Move language models tests (hybrid) back to L4 (#43129) Kevin H. Luu 2026-05-19 11:51:34 -07:00
9aaf83ef50 [CI failure] Temporarily disable using persistent cache for flashinfer autotune (#43119) Wei Zhaoandgemini-code-assist[bot] <176961590+gemini-code-assist[bot]@users.noreply.github.com> 2026-05-19 14:44:32 -04:00
f54721bcc3 [Bugfix][MoE] FlashInfer one-sided: workspace union across heterogeneous layers (#42976) tomeras91 2026-05-19 21:43:04 +03:00
aed2eb355a [Docs] Fix MooncakeStoreConnector role in disaggregated example (#42994) Dao007foreverandClaude 2026-05-19 11:14:43 -07:00
d247a931cc [feat] Add FP8 per-tensor Q scale support to Triton attention backend (#42080) Dom Brown 2026-05-19 17:02:05 +01:00
8200fbe1ac [Misc] add humming to dependencies (#42540) Jinzhen Lin 2026-05-19 23:36:47 +08:00
42b4f1fdf7 [Refactor] Extract extract_types_from_schema utility from Minimax M2 tool parser (#43025) Flora Feng 2026-05-19 11:21:12 -04:00
1c6158083a [Model] Openvla support (#42654) Wang Yiwen 2026-05-19 23:17:42 +08:00
d740e2c029 [XPU] update xpu graph usage (#43043) Xinyu Chen 2026-05-19 23:09:07 +08:00
b82e908b4c [Perf][4/n] Eliminate various GPU<->CPU syncs (#42347) Nick Hill 2026-05-19 07:35:54 -07:00
a78b842d0e [Bugfix] Fix top logprobs token placeholders in /inference/v1/generate (#42887) Sage 2026-05-19 13:21:49 +03:00
129019f334 [CI] Add MTP + PD disagg test for Qwen3.5 (#42677) zhanqiuhuandNicolò Lucchesi 2026-05-19 05:44:33 -04:00
ef54a4d604 [Misc][MM] Remove redundant code in CLIPAttention (#43046) Shanshan Shen 2026-05-19 16:43:16 +08:00
07beaed842 [Model Refactoring] Rename deepseek_v4.py to model.py [4/N] (#43077) Woosuk Kwon 2026-05-19 01:12:46 -07:00
056bc2e166 [KVConnector][DSV4] HMA support for Mooncake store connector (#42828) Yifan Qiao 2026-05-19 01:07:46 -07:00
f34623bf3c [bug] AsyncScheduler drops first post-resume token after pause_generation + clear_cache (#42117) Aaron Haoandmergify[bot] <37929162+mergify[bot]@users.noreply.github.com> 2026-05-19 01:06:21 -07:00
b14be81c1f [Model Refactoring] Move deepseek_v4_ops to models/deepseek_v4 [3/N] (#43073) Woosuk Kwon 2026-05-19 00:52:54 -07:00
301d986473 [Frontend] Consolidate beam search by BeamSearchMixin. (#42946) wang.yuqi 2026-05-19 15:37:40 +08:00
257af77bc2 [Docs] Reorganize online serving docs. (#41907)
wang.yuqigemini-code-assist[bot] <176961590+gemini-code-assist[bot]@users.noreply.github.com>Cyrus Leung Harry Mellor
2026-05-19 14:43:18 +08:00

... 64 65 66 67 68 ...

Commit Graph Select branches Hide Pull Requests main patches sm86 Mono Color

Commit Graph

Select branches

Hide Pull Requests

main

patches

sm86