Commit Graph

Select branches

Hide Pull Requests

main

patches

sm86

Mono Color

• b136cc2c2c [Bugfix][Model] Add stability window to DiffusionGemma to match HF stability_threshold semantics (#45965) Nathaniel McVicarandHarry Mellor 2026-07-06 12:39:12 -07:00
• 9fde043f54 [Kernel][Helion][1/N] Add Helion kernel for silu_and_mul_per_block_quant (#43994) 3 people 2026-07-06 12:19:01 -04:00
• 24dd2aec81 [Bugfix] Preserve FP8 indexer WK pairs across incremental load_weights (#46168) 4 people 2026-07-07 00:16:46 +08:00
• 3ee9eea928 [macOS][CPU][Installation] Fix the broken installation of vllm 0.24.0 in macos + cpu (#47457) Ranran 2026-07-06 10:59:16 -05:00
• 5bce653e09 Make the Transformers modeling backend as fast as native vLLM (#47187) Harry Mellorandmergify[bot] <37929162+mergify[bot]@users.noreply.github.com> 2026-07-06 16:59:14 +01:00
• 5ad11172b7 [perf]Add fused Kimi image preprocessing (#47416) 6 people 2026-07-06 23:46:32 +08:00
• f70caef48b [Perf] Cache token_to_req_indices for dsv4, 5x~6x kernel performance improvement (#47474) Wentao Ye 2026-07-06 11:17:46 -04:00
• 8d8ec38361 [Bugfix][Spec Decode] Add missing draft_id_to_target_id to DSparkDeepseekV4ForCausalLM (#47429) 3 people 2026-07-06 22:55:47 +08:00
• b1c6dba558 [Refactor] Remove multiple dead code (#47329) Wentao Ye 2026-07-06 10:54:08 -04:00
• 598d51153a [Bugfix][Distributed] Delegate MNNVL allreduce one-shot selection (#47589) jescoandmergify[bot] <37929162+mergify[bot]@users.noreply.github.com> 2026-07-06 10:47:06 -04:00
• 095adf1fdc [Bugfix] Fix int32 overflow in triton_decode_attention page offsets (#47671) Yifan Qiao 2026-07-06 07:36:15 -07:00
• 51ee564e56 [CI] Skip test for checkpoint that was deleted (#47748) Harry Mellor 2026-07-06 15:24:09 +01:00
• 373eb314af [Bugfix][Core] Fix num_output_placeholders underflow with async scheduling + spec decode (#46066) Ting SUNandmergify[bot] <37929162+mergify[bot]@users.noreply.github.com> 2026-07-06 20:50:38 +07:00
• 641cb59592 [Doc] Clarify fastokens availability (#45813) LiJzdandOpenAI Codex 2026-07-06 21:33:05 +08:00
• 07f9baf756 Revert "[Platform] Replace torch.cuda.Event with torch.Event (#47140)" (#47668) Kunshang JiandHarry Mellor 2026-07-06 21:18:33 +08:00
• 7a90eb98ab [Bugfix] [Gemma4] Fix Gemma4 MTP draft model layers ignoring quant_config (#47091) Ayushman SinghandBenjamin Chislett 2026-07-06 09:04:00 -04:00
• 8f4c69b222 [Rust Frontend] Cache metric handles for scheduler & request stats (#47444) Bugen Zhaoandmergify[bot] <37929162+mergify[bot]@users.noreply.github.com> 2026-07-06 21:02:59 +08:00
• 8b79971bb9 attention: pass None for unused args in unified attention TD path (#43597) Artur Fierkaandquinnlp 2026-07-06 15:01:21 +02:00
• f676808ba0 [CI] Use TTY for AMD CI tests for colored buildkite logs (#47730) Nick Hill 2026-07-06 13:50:29 +01:00
• 98e4726a14 [fix][run_batch]: respect proxy env vars when downloading media URLs (#47697) Qiming Zhang 2026-07-06 05:45:48 -07:00
• 740f379fae [ROCm][AITER] Directly Implement AITER Custom All-reduce in CudaCommunicator (#46065) BadrBasowid 2026-07-06 20:16:32 +08:00
• 40cc2e8327 [Bugfix] Return HTTP 422 for unprocessable image URLs instead of 500 (#47165) Alexis K. 2026-07-06 13:56:23 +02:00
• ba22152096 fix(security): block request-level GPU video backend selection withou… (#47259) Juan Pérez de Algabaandmergify[bot] <37929162+mergify[bot]@users.noreply.github.com> 2026-07-06 11:36:49 +02:00
• 90ce3a09be [bugfix] fix MOSS-Audio deepstack_input_embeds initialization in PP (#47607) Yan Ma 2026-07-06 17:15:50 +08:00
• 26c754d847 [XPU][Bugfix] Do not transpose weight_scale_inv at load time (#47116) Ma JianandKunshang Ji 2026-07-06 17:15:26 +08:00
• 3d7f357ebf [Doc] docs: fix note formatting for pooling models (#47701) Sungjae Lee 2026-07-06 18:01:10 +09:00
• 736f1a5907 [XPU] Route mm_prefix models to Triton attention backend (#47688) liuzhenwei 2026-07-06 16:52:44 +08:00
• 344609ab17 [CI/Build] Fix pre-commit check (#47695) Li, Jiang 2026-07-06 16:24:24 +08:00
• d039c17114 [Bugfix] Recycle post-final-norm hidden in GLM MTP (single norm) (#47448) xiaozhoupy 2026-07-06 16:07:56 +08:00
• cdab28319f [XPU][CI]Add agent tags for Basic Models Tests (Initialization) in Intel GPU CI (#47675) xiangdong 2026-07-06 15:15:45 +08:00
• 2fa10566e3 [Core][DP] Rotate load-balancer tie-break to avoid systematic engine bias (#47420) Qiming Zhang 2026-07-06 00:09:16 -07:00
• fb265fc8fb [ROCm][CI] Increasing parallelism in Basic Models Tests (Extra Initialization) (#47591) Andreas Karatzas 2026-07-06 02:06:16 -05:00
• 8f0e75e16b [ROCm][CI] Adding nixl multiconn (#47481) Andreas Karatzas 2026-07-06 02:04:58 -05:00
• 98ba9b9583 [Frontend] Support OpenAI Responses API namespace tools (#47024) Zhongandzhongjing123 2026-07-06 14:21:27 +08:00
• 990c2a0187 [RISC-V] Enable BF16 on VLEN=256 hardware (#45243) velonica0 2026-07-06 14:05:16 +08:00
• e433634c78 [Performance][Hardware][RISC-V] Reduce LMUL pressure in INT4 LUT dequant (#47538) LiuTongandClaude 2026-07-06 13:58:56 +08:00
• 16f8110935 [Bugfix][CPU][RISC-V] Fix VLEN detection for RVV attention path (#47532) LiuTongandClaude 2026-07-06 13:58:03 +08:00
• d9c1767cd4 [INC][ARK] Direct Register Custom Op for ARK (#46361) Zhenzhong Xuandmergify[bot] <37929162+mergify[bot]@users.noreply.github.com> 2026-07-06 13:45:50 +08:00
• e9cc1fd093 [CI/Build][CPU] Remove global extra index (#47687) Li, Jiang 2026-07-06 13:42:01 +08:00
• f1073c050c [CPU][BugFix] Multiple fixes to w4a8_int8 CPU MoE path (#46739) Fadi Arafeh 2026-07-06 06:39:20 +01:00
• 394edc8108 [XPU] limit max-num-seqs in test_lmeval.py for XPU (#47682) Qiming Zhang 2026-07-05 22:34:16 -07:00
• 69715823df [Test][XPU] Skip fork in kv_sharing_fast_prefill test on XPU (#47406) Liangliang MaandKunshang Ji 2026-07-06 11:32:26 +08:00
• 6569df6a3e [Test][LoRA] Use lightweight CPU reference and skip heavy cleanup in punica ops tests (#47534) Chaojun Zhang 2026-07-06 11:29:59 +08:00
• f2aaf59151 [Feature] Support MTP speculative decoding for Bailing hybrid models (#44880) alex101-opsandzc02384840 2026-07-06 10:38:50 +08:00
• 95a248faed [Attention Backend] HPC_ATTN backend support mtp and dynamic scheduled attention (#47433) Cheng Jiangandchengvjiang 2026-07-06 09:18:25 +08:00
• d2ec433e37 [XPU] Fix Eagle3 initialization on XPU (#43957) 4 people 2026-07-06 08:46:05 +08:00
• 78a04c208d [XPU] Fix CUDA API shims breaking Torch Dynamo during AOT compile (#43092) 3 people 2026-07-06 02:29:20 +02:00
• b71218107f [ROCm][Test] Fix test_per_token_group_quant_fp8 tolerance for 1-ULP FP8 rounding on gfx950 (#46944) Spandan Tiwari 2026-07-05 16:02:30 -07:00
• cc1d020d01 [MRV2] Enable mm prefix bidi attention support on MRV2 (#46942) Isotr0pyandNick Hill 2026-07-05 22:45:29 +08:00
• 8974ed89cd [Bugfix][Voxtral Realtime] Fix token feedback timeout silent hang (#44461) Ting SUN 2026-07-05 19:42:36 +07:00
• fb2faceacd [Bugfix][Model] Fix crash loading Mamba/Mamba2 checkpoints without an architectures field (#46037) 3 people 2026-07-05 19:42:32 +07:00
• b6cc46ec3b [Feature] Support sequence parallel without the need for DP, 1.9%~5.0% E2E Throughput Improvement (#47070) Wentao YeandCanlin Guo 2026-07-05 08:41:30 -04:00
• fa4321de3d [Bugfix][TurboQuant] Preserve KV cache dtype in backend shape (#47609) Lucas Wilkinson 2026-07-05 04:20:48 -04:00
• 9226613043 [Bugfix][Pooling] Forward instruction to Jina reranker scoring prompts (#47590) Ting SUN 2026-07-05 12:39:13 +07:00
• 34b560b725 [Bugfix][Gemma4] Fix FA4 mm_prefix mask: add sliding window and absolute q_idx (#47332) Luciano MartinsandLuciano Martins 2026-07-04 21:46:40 -03:00
• 91b5647300 [Bugfix][Model] Allow Run:ai memory_limit sentinel values (#47337) Ting SUN 2026-07-05 07:08:34 +07:00
• 4a6bf3c77f [ROCm][CI] Fix Kernels and Kernels attention test failures (#47519) Carl Persson 2026-07-04 22:59:51 +02:00
• d2afe39647 [Bugfix][Frontend] Preserve default sampling params in batch chat (#47597) Ting SUN 2026-07-05 02:06:39 +07:00
• 2a9113f998 [Perf] Remove redundant op for GLM 5.2 (#47198) Wentao Ye 2026-07-04 13:25:02 -04:00
• 0cd6f767e3 [Bugfix][Frontend][gpt-oss] Recover raw tail when Harmony parser ends non-terminal (#47379) yzong-rh 2026-07-04 10:46:24 -04:00
• f1445f6dbd [CI] Bump huggingface-hub from v1.10.2 to v1.22.0 (#47551) Harry Mellor 2026-07-04 15:45:45 +01:00
• 1d354c694e [Misc] Validate Pooling cache_salt Values (#46966) Taneem IbrahimandWentao Ye 2026-07-04 09:19:28 -05:00
• 2f21224527 [Misc] Update request-extras parity for batch chat completion (#47333) Taneem Ibrahim 2026-07-04 09:19:04 -05:00
• fa1fa968c4 [Misc] Forward request-level prompt extras for cross-encoder scoring (#46939) 3 people 2026-07-04 09:18:36 -05:00
• 6eac8e0070 [Misc] Preserve cross-encoder pooling extra kwargs (#47082) Taneem Ibrahimandmergify[bot] <37929162+mergify[bot]@users.noreply.github.com> 2026-07-04 07:14:13 -05:00
• 1a308c449c [XPU] Add W8A8 FP8 linear kernel with multi-granularity quant support (#43645) Chaojun ZhangandCopilot 2026-07-04 18:10:01 +08:00
• e7c9df9449 [Bugfix][Structured Output][Spec Decode] Constrain bitmask and trim grammar advance at the reasoning boundary (#44297) yue.yuandBenjamin Chislett 2026-07-04 17:08:45 +08:00
• 26eb87204d [Bugfix] Fix CPU split-KV scratchpad sizing (#45844) gausah01 2026-07-04 07:47:23 +01:00
• 4c3c17d43b [ROCm] Disable persistent sparse-MLA kernel for chunked-prefill continuations (#47567) Rohan PotdarandClaude Opus 4.8 2026-07-04 01:21:43 -05:00
• f329ce405b [ROCm][CI][Bugfix] Use VllmRunner for voxtral_realtime tests to avoid OOM on AMD GPU (#47536) Shanshan ShenandAndreas Karatzas 2026-07-04 12:26:10 +08:00
• 07516fda67 [MRV2][SD] Make Dynamic SD comatible with Full Cuda Graphs (#45953) 3 people 2026-07-03 23:58:27 -04:00
• 67ff0ae30f Support nvfp4 kv with kv-cache-dtype-skip-layers sliding_window (#42890) sychen52andmergify[bot] <37929162+mergify[bot]@users.noreply.github.com> 2026-07-03 19:29:13 -07:00
• ab3b6d97aa [Frontend] Limit SO_REUSEPORT to multi-worker serving (#47529) Bugen Zhao 2026-07-04 09:26:24 +08:00
• fb5291b35b [Frontend] [Parser] Port DeepSeek V4 to streaming parser engine framework (#45877) Ben Browning 2026-07-03 20:55:23 -04:00
• d6d39c111e [GLM4V] Avoid GLM4V processor init during startup metadata reads (#47155) labAxiaoming 2026-07-04 06:03:16 +08:00
• 379950191f [Bugfix][Multimodal] Normalize direct PIL image inputs (#47566) Ting SUNandCyrus Leung 2026-07-04 04:27:14 +07:00
• 576bf75d0e [AMD][EPLB] Enable EPLB for Quark OCP MXFP4 MoE (#47220) 4 people 2026-07-03 13:41:52 -06:00
• f006e5a24c [CI][AMD] Allow git operations on previously created work trees (#47554) Tres 2026-07-03 21:41:01 +02:00
• f63dca6838 [ROCm] Fix encoder-decoder cross-attention KV layout aliasing (#47035) djramicandAndreas Karatzas 2026-07-03 20:53:29 +02:00
• 8651f043b8 [Rust Frontend] Speed up chat roundtrip tests (#47523) Bugen Zhao 2026-07-04 02:25:06 +08:00
• 3775d5fcab [ROCm][CI] Adding test groups for parity with upstream (#47479) Andreas Karatzas 2026-07-03 10:15:01 -05:00
• d7192cfccf [CI Bugfix] Lazily import Qwen warmup dependencies (#47539) Roberto L. Castroandmergify[bot] <37929162+mergify[bot]@users.noreply.github.com> 2026-07-03 17:10:49 +02:00
• 978de83353 [Bugfix][CPU] Ship examples/ in the CPU release image (#47447) AgenticSpark 2026-07-03 04:46:24 -07:00
• a14f57a3ac [Frontend] Refine the entrypoint class's inheritance hierarchy. (#47498) wang.yuqi 2026-07-03 18:50:06 +08:00
• 18f658bb31 [Bugfix][Frontend] Fix batch chat endpoint corrupting logprobs when return_token_ids is set (#47384) David FengandClaude 2026-07-03 06:01:34 -04:00
• 400a9c386d [Rust Frontend] Bump llm-multimodal version (#47530) Isotr0py 2026-07-03 17:48:36 +08:00
• bbdcbe4686 Move Roberta remaining nn.Embedding to VocabParallelEmbedding (#47452) Max de Bayser 2026-07-03 05:47:50 -04:00
• 4875b4456b [Doc] Fix VLM2Vec benchmark chat template path (#47517) Kalyanam Dewri 2026-07-03 01:24:45 -07:00
• 1f486d96a1 Add Triton Backend for Unlimited-OCR R-SWA (#47102) Dakai An 2026-07-03 15:11:50 +08:00
• b790c84cde [CI] Enable sccache for Rust build under CUDA/ROCm (#45246) Bugen Zhao 2026-07-03 14:45:41 +08:00
• 6429d5f527 [Rust Frontend] add repetition_detection support to sampling params (#46684) yangyang-cs95andBugen Zhao 2026-07-03 14:06:23 +08:00
• fbc9ba6d30 New stable abi cleanup (#46656) Chris Leonard 2026-07-03 02:02:26 -04:00
• 2dfaae752b [XPU][CI]Fix dependency typo in Intel GPU CI (#47510) xiangdong 2026-07-03 12:11:47 +08:00
• bd8d9021ce [CPU][Build] Enable oneDNN ITT task collection by default for CPU primitive-level profiling (#47467) Evgeny Parshutin 2026-07-03 06:00:19 +02:00
• 3f0b773b30 [XPU][CI]Mv huggingface cache to larger disk in Intel GPU CI (#47405) xiangdong 2026-07-03 11:56:17 +08:00
• 9b8e76589d [Rust Frontend] Recover buffered text from incomplete tool calls at EOS (#47289) Reid 2026-07-03 11:45:03 +08:00
• 1aeabec355 [Bugfix][Rust Frontend] Tolerate out-of-vocab prompt ids in detokenizer (#44682) Ting SUNandBugen Zhao 2026-07-03 10:41:53 +07:00
• 979f5511d7 [Bugfix][Gemma4] Keep image bidirectional attention within the sliding window (#47217) Luciano MartinsandLuciano Martins 2026-07-02 23:57:41 -03:00
• 41de1380c2 [BugFix] Derive FlashInfer Q dtype from resolved per-group builder state (#47485) Michael GoinandClaude Fable 5 2026-07-02 22:33:28 -04:00
• d85601c20f [CI] Pin modelscope version to fix test breakage (#47465) Nick Hill 2026-07-03 03:33:08 +01:00