Commit Graph

Select branches

Hide Pull Requests

main

patches

sm86

Mono Color

• c60f8e3b49 [Bugfix][ROCm][GPT-OSS] Use old triton_kernels implementation on ROCm if the new API is not available (#34153) Gregory Shtrasberg 2026-02-09 17:38:54 -06:00
• 5e75a14a66 [Doc] Add DCP support to attention backend doc (#33936) Michael Goin 2026-02-09 18:33:43 -05:00
• e7e52781ff [ModelRunner V2][BugFix] Fix max_query_len calculation (#34167) Nick Hill 2026-02-09 13:47:17 -08:00
• bb9f97308d [torch.compile][Fusion] Fix attention fusion pass removing kv_udpate op. (#33945) Charlie Fu 2026-02-09 15:15:43 -06:00
• 4d39650961 [ROCm] update triton branch to support gpt-oss models for gfx11xx devices (#34032) Hongxia Yang 2026-02-09 14:36:30 -05:00
• 8fd31f6245 [Bugfix] Voxtral prompt/audio placeholder alignment (#34140) Artus Krohn-Grimberghe 2026-02-09 20:30:38 +01:00
• eadb4e868b [Bugfix] Avoid duplicate k-proj weight emission in helper (#34142) Artus Krohn-Grimberghe 2026-02-09 20:17:44 +01:00
• 285bab4752 [Kernel] use flashinfer for gdn prefill (#32846) Jiangyun Zhu 2026-02-10 01:17:25 +08:00
• 995bbf38f1 [Bugfix] Fix shared expert input for latent MoE in EP+DP (Nemotron-H) (#34087) TomerBN-NvidiaandCursor 2026-02-09 18:44:18 +02:00
• d4f123cc48 [Kernel] FlashInfer: switch allreduce fusion to unified API (#33985) Mohammad Miadh Angkad 2026-02-09 23:43:24 +08:00
• cb62e86f83 Add NUMA Core binding in nixl_connector for CPU xPyD (#32365) ZhengHongming888andgemini-code-assist[bot] <176961590+gemini-code-assist[bot]@users.noreply.github.com> 2026-02-09 07:39:12 -08:00
• 781ddf7868 [CI][torch.compile] Fix incorrect filtering for E2E fusion tests on B200 (#34031) Luka Govedič 2026-02-09 10:05:14 -05:00
• 64a9c2528b [UX] Add --language-model-only for hybrid models (#34120) Roger Wang 2026-02-09 06:57:33 -08:00
• d0d97e2974 [Misc] Fix up attention benchmarks (#33810) Lucas WilkinsonandMatthew Bonanni 2026-02-09 06:42:03 -08:00
• 9562912cea [MODEL] Adding Support for Qwen3.5 Models (#34110) 6 people 2026-02-09 21:12:58 +08:00
• 9bdb06b436 [XPU][6/N] add xpu scaled_mm kernel (#34117) zofia 2026-02-09 20:17:35 +08:00
• caad9f1e01 [Fix] [CPU Backend] : Prepack weights for w8a8 oneDNN matmul (#33901) Nikhil Gupta 2026-02-09 10:04:41 +00:00
• 1d5922fade [ASR] Fix audio benchmark and add RTFx metric (#32300) Ekagra RanjanandNicolò Lucchesi 2026-02-09 05:02:37 -05:00
• 3025b3cebb [CI] Remove empty image_size_factors for fuyu, glm4_1v, glm_ocr (#34107) Andreas Karatzas 2026-02-09 03:37:04 -06:00
• 978a37c823 [Model] GLM adaptation (#34124) Jee Jee Li 2026-02-09 17:32:52 +08:00
• 5a5c43511a fix(cpu): fix mla_decode compilation on x86 without AVX512 (#34052) ihb2032androot 2026-02-09 16:55:41 +08:00
• d9bede0314 [BugFix] Fix fastsafetensors TP all procs using all GPUs (#34070) Nick HillandCyrus Leung 2026-02-08 23:15:46 -08:00
• 22b64948f6 [Frontend][last/5] Make pooling entrypoints request schema consensus. (#31127) wang.yuqi 2026-02-09 14:42:38 +08:00
• 7c233dbb36 [Tiny] Rename encoder budget file to more specific name (#34103) Reagan LeeandReagan Lee 2026-02-08 19:48:19 -08:00
• a75a5b54c7 [bug-fix] supported_tasks is breaking backward compatibility at init_app_state (#34027) kourosh hakhamaneshiandCyrus Leung 2026-02-08 17:46:46 -08:00
• f97ca67176 [Release 2.10] Update to Torch 2.10 - final release (#30525) Andrey Talman 2026-02-08 16:51:09 -05:00
• 084aa19f02 Add support for ModelOpt MXFP8 dense models (#33786) danisereb 2026-02-08 21:16:48 +02:00
• 1ecfabe525 glm 4.6 fused tuned inference config for B200 (#32958) navmarri14 2026-02-08 10:55:47 -08:00
• 4df841fe75 [torch.compile] Add an option to force-enable the MOE cold start optimization (#33735) Richard Zou 2026-02-08 13:42:56 -05:00
• a263aa6140 [BugFix] Change support no act and mul for marlin (#34088) TomerBN-NvidiaandTomer Natan 2026-02-08 19:18:22 +02:00
• 179ae7da8f [Revert] Fix performance regression for GLM-4.7-GPTQ decode and MTP acceptance rate (#33771) aabbccddwasd 2026-02-09 00:13:24 +08:00
• c4df59ad43 Add embedding input functionality for disabled modalities [remake] (#32493) 3 people 2026-02-08 04:57:16 -08:00
• 785cf28fff [ROCm] [CI] Reduce Resource of two test groups (#34059) TJian 2026-02-08 15:17:26 +08:00
• a96197f564 [Perf] Simplify DeepseekV32 tokenizer, ensure fast detokenization used (#33855) Nick Hill 2026-02-07 23:16:34 -08:00
• ab10d79855 [ROCm][Bugfix] fix act_quant_fusion module import error (#34069) Andreas Karatzas 2026-02-07 21:21:12 -06:00
• 7fcb705b80 [CI/Build] Skip GCS test (#34057) Cyrus Leung 2026-02-08 00:52:38 +08:00
• b956cdf818 [Doc] Fix run_batch docs (#34056) Cyrus Leung 2026-02-07 22:18:16 +08:00
• ed17f54c8b Perf tuning and expansion of cases covered for wvSplitKrc (#33493) Hashem Hashemi 2026-02-07 05:33:11 -08:00
• 860981d8d8 Make directory exist ok for ray spinning up multiple replicas on a single instance (#33604) Jiang Wu 2026-02-07 05:30:49 -08:00
• 52181baaea Update DeepGEMM version pin in Dockerfile to match #32479 (#33935) zifeitongandWentao Ye 2026-02-07 05:30:22 -08:00
• de3869bb4d move checks out of unified_kv_cache_update custom op (#33943) Rohan Potdar 2026-02-07 07:30:09 -06:00
• ce9b3cd3e9 [PluggableLayer][3/N] Apply PluggableLayer to mamba layers. (#33660) whx 2026-02-07 21:26:05 +08:00
• db4ede9743 [Model] Enable Step3p5ForCausalLM testing (#33755) Jee Jee Li 2026-02-07 21:25:24 +08:00
• 2cb2340f7a [Frontend]Add support for transcriptions and translations to run_batch (#33934) Pooya DavoodiandCyrus Leung 2026-02-07 05:24:57 -08:00
• 4df44c16ba Enable Eagle3 speculative decoding for Mistral3ForConditionalGeneration to support eagle3 (#33939) TundeAtSNandgemini-code-assist[bot] <176961590+gemini-code-assist[bot]@users.noreply.github.com> 2026-02-07 08:24:52 -05:00
• 81fe69cae5 [torch.compile] Stop compiling identical artifacts (#34003) Richard Zou 2026-02-07 08:24:48 -05:00
• dd6a6e1190 [Kernel] Add KernelConfig flag to enable/disable FlashInfer autotune (#34006) Mohammad Miadh AngkadandLuka Govedič 2026-02-07 21:24:44 +08:00
• edb359cce4 [Renderer] Define render_cmpl and render_chat (#34039) Cyrus Leung 2026-02-07 21:24:40 +08:00
• 6ed5eda300 [CI][Build] Pin grpcio-tools==1.78.0 (#34048) wang.yuqiandCyrus Leung 2026-02-07 21:24:35 +08:00
• 11a4c9d30d [Misc] Simplify get_max_tokens (#34036) Cyrus Leung 2026-02-07 16:59:49 +08:00
• 15a0b9e570 Fix spelling errors (#33978) lukec 2026-02-07 15:58:50 +08:00
• c490d8cc73 [ROCm][CI] Pinning lm-eval version to resolve multi-modal small eval bug (#34038) Andreas Karatzas 2026-02-07 00:21:08 -06:00
• 48312e579a [Misc] Make PlaceholderRange.get_num_embeds a method (#34035) Cyrus Leung 2026-02-07 13:30:17 +08:00
• bc32444b23 [Kernel] Add enable_sm120_or_later for SM121 (DGX Spark) CUTLASS support (#33517) Vel 2026-02-06 20:28:01 -08:00
• 18e8545297 [Revert] Add util handle_deprecated back (#33998) Wentao Ye 2026-02-06 23:14:45 -05:00
• 6f7adc533a fix description in plugin_system.md (#33999) 果冻虾仁 2026-02-07 11:37:02 +08:00
• 40218a82ba [ModelRunner V2] Revert token rank comparison difference for now (#34017) Nick Hill 2026-02-06 19:11:05 -08:00
• 1c3b22058f [Misc] Add backward-compatible import aliases for renamed translations module (#34015) kourosh hakhamaneshiandCursor 2026-02-06 19:01:41 -08:00
• 3920cafdd6 [Bugfix] Fix _fused_moe_lora_expand signature mismatch (#33821) Xin Yang 2026-02-06 18:45:59 -08:00
• ec28784fdc [CI][AMD]Bugfix] Check that model_config is not None in enable_norm_pad_fusion (#34007) rasmith 2026-02-06 20:43:25 -06:00
• 55aeec04f5 [Bugfix] Fix Whisper tokenization (#34011) Nicolò Lucchesi 2026-02-07 03:42:52 +01:00
• 906077181b [Bugfix] Fix QK Norm+RoPE fusion pattern matching on B200+FP8 (#33967) IkennaandLuka Govedič 2026-02-06 21:27:33 -05:00
• 89a385d79f [Feat][RL] Pause and Resume with keep requests for single engine (#32351) Aaron HaoandRobert Shaw 2026-02-06 16:08:58 -08:00
• 4a2d00eafd [bugfix] [ROCm] Fix premature CUDA initialization in platform detection (#33941) kourosh hakhamaneshi 2026-02-06 14:17:55 -08:00
• 207c3a0c20 Fix RoutingMethodType logic (#33919) 3 people 2026-02-06 23:03:34 +01:00
• ae2e93f89b [Fix] Fix logprobs=0 handling for /inference/v1/generate endpoint (#34010) Sumanth R Hegde 2026-02-06 12:33:40 -08:00
• 9e9acce577 [Bugfix] Fix no attribute error of SharedFusedMoE (DeepSeek-V3.1 as test model) (#33993) xuebwang-amd 2026-02-07 03:11:32 +08:00
• fe5438200b [Rocm][Bugfix] Fix dtype not same for gemm_a4w4 op (#33734) Charlie Fu 2026-02-06 13:09:59 -06:00
• 77c09e1130 [Refactor] Remove align block size logic in moe_permute (#33449) Wentao Ye 2026-02-06 13:57:06 -05:00
• 16786da735 [Model Runner V2] support apply penalty for spec decode (#33251) zhrrr 2026-02-07 02:56:48 +08:00
• aaa2efbe98 [DOC] [ROCm] Update docker deployment doc (#33971) 3 people 2026-02-07 02:05:35 +08:00
• aca5967416 [KV Connector] Add missing method overrides to MultiConnector (#33292) Seiji Eicher 2026-02-06 09:58:21 -08:00
• 67a746e87f [Log] Optimize duplicate startup log (#33944) Wentao Ye 2026-02-06 12:49:56 -05:00
• 7bec435130 [Bugfix] Fix the issue where tool calling does not work when using fast detokenization with dsv32 (#33964) Chauncey 2026-02-07 01:23:44 +08:00
• 5c52644b10 [Docs] Update link to Benchmark CLI documentation (#33254) Eldar Kurtić 2026-02-06 17:00:59 +01:00
• 2ce9fe4ad0 [XPU][5/N] add wna16 xpu kernel (#33973) zofia 2026-02-06 23:59:53 +08:00
• cd8b405bd0 [Refactor] Consolidate sequence normalization and enc-dec parsing (#33928) Cyrus Leung 2026-02-06 23:43:47 +08:00
• 4707f7ebb4 [Model] Support MiniCPM-o 4.5 (#33431) tc-mbandmslv 2026-02-06 23:29:10 +08:00
• c39ee9ee2b [Docs] Add sections on process architecture and minimum CPU resources (#33940) Michael Goin 2026-02-06 10:26:43 -05:00
• 350ca72c04 [ROCm][AITER] Fix AITER import regression for explicit backend selection (#33749) Andreas Karatzas 2026-02-06 09:08:16 -06:00
• 1fb0495a72 [FIX] guidance: use max(vocab_size, len(tokenizer)) for n_vocab (#33509) FredericOdermatt 2026-02-06 15:23:03 +01:00
• 85ee1d962b [Bugfix] Fix models and tests for transformers v5 (#33977) 3 people 2026-02-06 14:47:41 +01:00
• 51a7bda625 Update WeightTransferConfig to be more standard like the others (#33989) Harry Mellor 2026-02-06 13:15:00 +00:00
• 6e7b1c4b59 [Docs] Improve documentation (#33799) 3 people 2026-02-06 13:57:09 +01:00
• 2991dd3d22 [Bugfix][Model] Support LoRA on Qwen3 Output Embedding (#29816) Kurt Shuster 2026-02-06 04:25:31 -08:00
• ac32e66cf9 [torch.compile] Reorganize vllm/compilation and tests/compile (0/N for vLLM IR) (#33731) Luka Govedič 2026-02-06 07:19:49 -05:00
• f79d9dce16 [CPU][BugFix] Fix loading of w8a8int models with bias (#33582) Fadi Arafeh 2026-02-06 11:59:20 +00:00
• ba5cbbf107 Bump HF Hub client to get bug fix (#33984) Harry Mellor 2026-02-06 11:25:33 +00:00
• 233b26ab35 [PaddleOCR-VL] Add BC for transformers 5.0 config (#33976) zhang-prog 2026-02-06 18:33:49 +08:00
• 791a94bed0 Consolidate and fix forbidden import pre-commit checks (#33982) Harry Mellor 2026-02-06 09:47:41 +00:00
• e969a169ef support view_from_cpu_tensor on XPU (#33868) Xinyu Chen 2026-02-06 16:34:20 +08:00
• 6d8d34be6d Fix main pre-commit (#33975) Harry Mellor 2026-02-06 08:08:05 +00:00
• 1363e3d6d5 [cpu][performance] CPU Paged Attention NEON BFMMLA BF16 Implementation (#32263) Gassan Salama 2026-02-06 07:01:48 +00:00
• 965525667b Onboard voyage-4-nano (#33720) chengchengpeiandgemini-code-assist[bot] <176961590+gemini-code-assist[bot]@users.noreply.github.com> 2026-02-05 22:23:34 -08:00
• 6550815c3a [XPU]Replace pip in docker.xpu with uv pip (#31112) sihao_li 2026-02-06 14:02:33 +08:00
• 7439e4f41b [XPU][4/N] add mxfp4 moe model support (#33679) Kunshang Ji 2026-02-06 13:03:59 +08:00
• ac04dd374f [CPU] Add BF16 Kernel type for s390x (#33788) R3hankhan 2026-02-06 10:27:02 +05:30
• 035a6cb09a [Misc] Update code for encoder-decoder models (#33900) Cyrus Leung 2026-02-06 11:38:39 +08:00
• a32cb49b60 feat(frontend): early-fail tokenization guard for user requests (#31366) 3 people 2026-02-06 11:38:02 +08:00
• 20d7454c9b fix(ROCm): Make flash_attn import optional in MLA attention (#33511) Rabi Mishra 2026-02-06 07:52:53 +05:30