Canlin GuoandNick Hill 455edc022b [ModelRunnerV2] Support prompt embeds (#42963) ...

Signed-off-by: gcanlin <canlinguosdu@gmail.com>
Signed-off-by: Canlin Guo <961750412@qq.com>
Signed-off-by: Nick Hill <nickhill123@gmail.com>
Co-authored-by: Nick Hill <nickhill123@gmail.com>

2026-08-17 21:58:50 +00:00

..

transformers

[Core] Check for GPU<->CPU syncs during CI (#43107)

2026-08-14 17:42:05 -07:00

__init__.py

[Bugfix][Model Runner V2] Restore multimodal draft capability detection (#50417)

2026-08-03 14:18:14 -07:00

adapters.py

[Mypy Fix] Mypy fix for "vllm/model_executor/models/[aA][bB]" (#48977)

2026-08-06 19:30:15 +00:00

afmoe.py

[Refactor] Delete dead code in models (#51838)

2026-08-11 15:44:26 -04:00

aimv2.py

[Mypy Fix] Mypy fix for "vllm/model_executor/models/[aA][bB]" (#48977)

2026-08-06 19:30:15 +00:00

apertus.py

[Perf] RMSNorm uncontiguous support, 1.2~3.1x kernel performance improvement (#49750)

2026-07-29 13:45:38 -07:00

arcee.py

Remove unnecessary load_weights methods (#44589)

2026-06-29 01:52:23 -07:00

arctic.py

Standardise weight tying on ParallelLMHead.tie_weights (#52147)

2026-08-13 15:10:09 +00:00

aria.py

[Mypy Fix] Mypy fix for "vllm/model_executor/models/[aA][bB]" (#48977)

2026-08-06 19:30:15 +00:00

audioflamingo3.py

[Core] Check for GPU<->CPU syncs during CI (#43107)

2026-08-14 17:42:05 -07:00

AXK1.py

[Refactor] Delete dead code in models (#51838)

2026-08-11 15:44:26 -04:00

bagel.py

[Mypy Fix] Mypy fix for "vllm/model_executor/models/[aA][bB]" (#48977)

2026-08-06 19:30:15 +00:00

bailing_moe_linear.py

[Mypy Fix] Mypy fix for "vllm/model_executor/models/[aA][bB]" (#48977)

2026-08-06 19:30:15 +00:00

bailing_moe_mtp.py

[Mypy Fix] Mypy fix for "vllm/model_executor/models/[aA][bB]" (#48977)

2026-08-06 19:30:15 +00:00

bailing_moe_v3_mtp.py

[Model] [Quantization] Add Ling hybrid MXFP4 routed experts support (#52114)

2026-08-13 21:46:53 +08:00

bailing_moe_v3.py

[Model] [Quantization] Add Ling hybrid MXFP4 routed experts support (#52114)

2026-08-13 21:46:53 +08:00

bailing_moe.py

Standardise weight tying on ParallelLMHead.tie_weights (#52147)

2026-08-13 15:10:09 +00:00

bee.py

[Mypy Fix] Mypy fix for "vllm/model_executor/models/[aA][bB]" (#48977)

2026-08-06 19:30:15 +00:00

bert_with_rope.py

[Mypy Fix] Mypy fix for "vllm/model_executor/models/[aA][bB]" (#48977)

2026-08-06 19:30:15 +00:00

bert.py

[Mypy Fix] Mypy fix for "vllm/model_executor/models/[aA][bB]" (#48977)

2026-08-06 19:30:15 +00:00

blip2.py

[Mypy Fix] Mypy fix for "vllm/model_executor/models/[aA][bB]" (#48977)

2026-08-06 19:30:15 +00:00

blip.py

Remove more unnecessary load_weights methods (#47058)

2026-06-30 15:22:16 +01:00

bloom.py

Standardise weight tying on ParallelLMHead.tie_weights (#52147)

2026-08-13 15:10:09 +00:00

chameleon.py

Standardise weight tying on ParallelLMHead.tie_weights (#52147)

2026-08-13 15:10:09 +00:00

chatglm.py

Remove unnecessary load_weights methods (#44589)

2026-06-29 01:52:23 -07:00

cheers.py

[Model] Add support for Cheers multimodal model (#38788)

2026-04-02 21:01:40 +08:00

clip.py

[Mypy Fix] Mypy fix for "vllm/model_executor/models/[cC][dD]" (#52003)

2026-08-13 11:33:55 +08:00

cohere2_moe.py

[Mypy Fix] Mypy fix for "vllm/model_executor/models/[cC][dD]" (#52003)

2026-08-13 11:33:55 +08:00

cohere2_vision.py

[Mypy Fix] Mypy fix for "vllm/model_executor/models/[cC][dD]" (#52003)

2026-08-13 11:33:55 +08:00

cohere_asr.py

[Core] Check for GPU<->CPU syncs during CI (#43107)

2026-08-14 17:42:05 -07:00

cohere_eagle.py

[Mypy Fix] Mypy fix for "vllm/model_executor/models/[cC][dD]" (#52003)

2026-08-13 11:33:55 +08:00

colbert.py

Model/colbert autoweightsloader (#44999)

2026-06-09 23:58:50 -07:00

colmodernvbert.py

[Mypy Fix] Mypy fix for "vllm/model_executor/models/[cC][dD]" (#52003)

2026-08-13 11:33:55 +08:00

colpali.py

[Mypy Fix] Mypy fix for "vllm/model_executor/models/[cC][dD]" (#52003)

2026-08-13 11:33:55 +08:00

colqwen3_5.py

[EPD] Remove duplicate image preprocessing in EPD and enable preprocess on GPU (#50390)

2026-08-05 15:59:54 -07:00

colqwen3.py

[EPD] Remove duplicate image preprocessing in EPD and enable preprocess on GPU (#50390)

2026-08-05 15:59:54 -07:00

commandr.py

[Mypy Fix] Mypy fix for "vllm/model_executor/models/[cC][dD]" (#52003)

2026-08-13 11:33:55 +08:00

config.py

[Mypy Fix] Mypy fix for "vllm/model_executor/models/[cC][dD]" (#52003)

2026-08-13 11:33:55 +08:00

conformer_encoder.py

[model] support FireRedLID (#39290)

2026-04-10 08:43:58 +00:00

cosmos3_edge.py

[Mypy Fix] Mypy fix for "vllm/model_executor/models/[cC][dD]" (#52003)

2026-08-13 11:33:55 +08:00

cosmos3.py

Cosmos3 FP8 ModelOpt/Diffusers remapping (#48952)

2026-07-20 11:31:48 -07:00

dbrx.py

[Mypy Fix] Mypy fix for "vllm/model_executor/models/[cC][dD]" (#52003)

2026-08-13 11:33:55 +08:00

deepencoder2.py

[Mypy Fix] Mypy fix for "vllm/model_executor/models/[cC][dD]" (#52003)

2026-08-13 11:33:55 +08:00

deepencoder.py

[Mypy Fix] Mypy fix for "vllm/model_executor/models/[cC][dD]" (#52003)

2026-08-13 11:33:55 +08:00

deepseek_eagle3.py

[Mypy Fix] Mypy fix for "vllm/model_executor/models/[cC][dD]" (#52003)

2026-08-13 11:33:55 +08:00

deepseek_eagle.py

[Mypy Fix] Mypy fix for "vllm/model_executor/models/[cC][dD]" (#52003)

2026-08-13 11:33:55 +08:00

deepseek_mtp.py

[Mypy Fix] Mypy fix for "vllm/model_executor/models/[cC][dD]" (#52003)

2026-08-13 11:33:55 +08:00

deepseek_ocr2.py

[Mypy Fix] Mypy fix for "vllm/model_executor/models/[cC][dD]" (#52003)

2026-08-13 11:33:55 +08:00

deepseek_ocr.py

[Mypy Fix] Mypy fix for "vllm/model_executor/models/[cC][dD]" (#52003)

2026-08-13 11:33:55 +08:00

deepseek_v2.py

[Mypy Fix] Mypy fix for "vllm/model_executor/models/[cC][dD]" (#52003)

2026-08-13 11:33:55 +08:00

deepseek_vl2.py

[Mypy Fix] Mypy fix for "vllm/model_executor/models/[cC][dD]" (#52003)

2026-08-13 11:33:55 +08:00

diffusion_gemma.py

[ModelRunnerV2] Support prompt embeds (#42963)

2026-08-17 21:58:50 +00:00

dots_ocr.py

[Mypy Fix] Mypy fix for "vllm/model_executor/models/[cC][dD]" (#52003)

2026-08-13 11:33:55 +08:00

eagle2_5_vl.py

…

ernie45_moe.py

Standardise weight tying on ParallelLMHead.tie_weights (#52147)

2026-08-13 15:10:09 +00:00

ernie45_vl_moe.py

Standardise weight tying on ParallelLMHead.tie_weights (#52147)

2026-08-13 15:10:09 +00:00

ernie45_vl.py

[Core] Check for GPU<->CPU syncs during CI (#43107)

2026-08-14 17:42:05 -07:00

ernie45.py

…

ernie_mtp.py

Standardise weight tying on ParallelLMHead.tie_weights (#52147)

2026-08-13 15:10:09 +00:00

exaone4_5_mtp.py

Standardise weight tying on ParallelLMHead.tie_weights (#52147)

2026-08-13 15:10:09 +00:00

exaone4_5.py

[Bugfix] Fix missing sequence_lengths in EXAONE-4.5 vision encoder (#45073)

2026-06-10 15:44:34 +01:00

exaone4.py

Standardise weight tying on ParallelLMHead.tie_weights (#52147)

2026-08-13 15:10:09 +00:00

exaone_moe_mtp.py

Standardise weight tying on ParallelLMHead.tie_weights (#52147)

2026-08-13 15:10:09 +00:00

exaone_moe.py

Standardise weight tying on ParallelLMHead.tie_weights (#52147)

2026-08-13 15:10:09 +00:00

exaone.py

Standardise weight tying on ParallelLMHead.tie_weights (#52147)

2026-08-13 15:10:09 +00:00

extract_hidden_states.py

[Spec Decode] Fix hidden-state extraction block size for hybrid verifiers (#46301)

2026-06-30 08:19:51 -07:00

fairseq2_llama.py

Remove unnecessary load_weights methods (#44589)

2026-06-29 01:52:23 -07:00

falcon_h1.py

Remove more unnecessary load_weights methods (#47058)

2026-06-30 15:22:16 +01:00

falcon.py

Standardise weight tying on ParallelLMHead.tie_weights (#52147)

2026-08-13 15:10:09 +00:00

fireredasr2.py

[Audio] Bundle get_generation_prompt() params into SpeechToTextParams (#36268)

2026-04-22 12:24:18 +08:00

fireredlid.py

[model] support FireRedLID (#39290)

2026-04-10 08:43:58 +00:00

flex_olmo.py

[MoE Refactor] Rename FusedMoE to FusedMoEFactory (#44941)

2026-07-31 11:28:18 +01:00

funasr.py

Standardise weight tying on ParallelLMHead.tie_weights (#52147)

2026-08-13 15:10:09 +00:00

funaudiochat.py

Preserve revision pins in secondary artifact loaders (#51446)

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

gemma2.py

Remove unnecessary load_weights methods (#44589)

2026-06-29 01:52:23 -07:00

gemma3_mm.py

[CI] Fix various failures on main (#47197)

2026-07-01 10:35:34 +01:00

gemma3.py

[Bugfix] Default Gemma3 Model intermediate_tensors to None (#50777)

2026-08-03 15:07:59 -04:00

gemma3n_audio_utils.py

…

gemma3n_mm.py

[Perf] Avoid some more unnecessary GPU<->CPU syncs (#51458)

2026-08-08 17:13:26 -07:00

gemma3n.py

Remove even more unnecessary load_weights methods (#48496)

2026-07-18 08:40:27 +00:00

gemma4_dspark.py

Fix Gemma 4 for upcoming Transformers version (#49797)

2026-08-10 07:48:26 +00:00

gemma4_mm.py

[Bugfix][Multimodal] Keep Gemma 4 video frame counts on CPU (#52441)

2026-08-16 09:17:04 +00:00

gemma4_mtp.py

Standardise weight tying on ParallelLMHead.tie_weights (#52147)

2026-08-13 15:10:09 +00:00

gemma4_unified.py

[Bugfix][Model] Gemma3n/Gemma4: pad variable-length audio batches (#50958)

2026-08-05 12:46:00 +08:00

gemma4.py

Fix Gemma 4 for upcoming Transformers version (#49797)

2026-08-10 07:48:26 +00:00

gemma.py

Remove unnecessary load_weights methods (#44589)

2026-06-29 01:52:23 -07:00

glm4_1v.py

[Perf] Avoid more GPU<->CPU syncs in multimodal encoders (#52369)

2026-08-14 10:40:14 -07:00

glm4_moe_lite_mtp.py

Remove even more unnecessary load_weights methods (#48496)

2026-07-18 08:40:27 +00:00

glm4_moe_lite.py

Remove even more unnecessary load_weights methods (#48496)

2026-07-18 08:40:27 +00:00

glm4_moe_mtp.py

Remove even more unnecessary load_weights methods (#48496)

2026-07-18 08:40:27 +00:00

glm4_moe.py

[Refactor] Delete dead code in models (#51838)

2026-08-11 15:44:26 -04:00

glm4.py

Standardise weight tying on ParallelLMHead.tie_weights (#52147)

2026-08-13 15:10:09 +00:00

glm4v.py

Remove unnecessary load_weights methods (#44589)

2026-06-29 01:52:23 -07:00

glm_ocr_mtp.py

Remove even more unnecessary load_weights methods (#48496)

2026-07-18 08:40:27 +00:00

glm_ocr.py

[Refactor] Remove multiple dead codes (#50285)

2026-08-03 10:59:08 -04:00

glm.py

…

glmasr_utils.py

…

glmasr.py

Remove even more unnecessary load_weights methods (#48496)

2026-07-18 08:40:27 +00:00

gpt2.py

Remove even more unnecessary load_weights methods (#48496)

2026-07-18 08:40:27 +00:00

gpt_j.py

Remove unnecessary load_weights methods (#44589)

2026-06-29 01:52:23 -07:00

gpt_neox.py

Standardise weight tying on ParallelLMHead.tie_weights (#52147)

2026-08-13 15:10:09 +00:00

gpt_oss.py

[MoE Refactor] Rename FusedMoE to FusedMoEFactory (#44941)

2026-07-31 11:28:18 +01:00

granite4_vision.py

Standardise weight tying on ParallelLMHead.tie_weights (#52147)

2026-08-13 15:10:09 +00:00

granite_speech_plus.py

Add model support for granite speech plus (#43519)

2026-06-04 14:47:48 +00:00

granite_speech.py

[Core] Check for GPU<->CPU syncs during CI (#43107)

2026-08-14 17:42:05 -07:00

granite.py

Standardise weight tying on ParallelLMHead.tie_weights (#52147)

2026-08-13 15:10:09 +00:00

granitemoe.py

Standardise weight tying on ParallelLMHead.tie_weights (#52147)

2026-08-13 15:10:09 +00:00

granitemoehybrid.py

Standardise weight tying on ParallelLMHead.tie_weights (#52147)

2026-08-13 15:10:09 +00:00

granitemoeshared.py

Standardise weight tying on ParallelLMHead.tie_weights (#52147)

2026-08-13 15:10:09 +00:00

gritlm.py

…

h2ovl.py

[Model] Remove Mono-InternVL (InternLM2VEForCausalLM) (#45129)

2026-06-12 10:41:09 +00:00

hrm_text.py

Standardise weight tying on ParallelLMHead.tie_weights (#52147)

2026-08-13 15:10:09 +00:00

hunyuan_v1.py

Standardise weight tying on ParallelLMHead.tie_weights (#52147)

2026-08-13 15:10:09 +00:00

hunyuan_vision.py

[EPD] Remove duplicate image preprocessing in EPD and enable preprocess on GPU (#50390)

2026-08-05 15:59:54 -07:00

hy_v3_mtp.py

Remove even more unnecessary load_weights methods (#48496)

2026-07-18 08:40:27 +00:00

hy_v3.py

Standardise weight tying on ParallelLMHead.tie_weights (#52147)

2026-08-13 15:10:09 +00:00

hyperclovax_vision_v2.py

…

hyperclovax_vision.py

[Refactor] Remove dead code in multiple files (#49745)

2026-07-27 15:58:26 -04:00

hyperclovax.py

[ROCm][CI] fix test_common.py (#48676)

2026-07-15 06:42:06 +00:00

idefics2_vision_model.py

[Core] Check for GPU<->CPU syncs during CI (#43107)

2026-08-14 17:42:05 -07:00

idefics3.py

[Core] Check for GPU<->CPU syncs during CI (#43107)

2026-08-14 17:42:05 -07:00

interfaces_base.py

…

interfaces.py

[K3] support recoverssm for K3 (#51855)

2026-08-17 19:34:14 +08:00

intern_vit.py

Remove even more unnecessary load_weights methods (#48496)

2026-07-18 08:40:27 +00:00

internlm2.py

Standardise weight tying on ParallelLMHead.tie_weights (#52147)

2026-08-13 15:10:09 +00:00

interns1_pro.py

Standardise weight tying on ParallelLMHead.tie_weights (#52147)

2026-08-13 15:10:09 +00:00

interns1_vit.py

Remove even more unnecessary load_weights methods (#48496)

2026-07-18 08:40:27 +00:00

interns1.py

[Perf] Avoid more GPU<->CPU syncs in multimodal encoders (#52369)

2026-08-14 10:40:14 -07:00

interns2_mobius.py

[Bugfix][Refactor] Keep Qwen3Next layer boundaries sequence parallel (#50685)

2026-08-14 11:58:32 +08:00

interns2_preview.py

[Model] Support InternS2 Preview (#42705)

2026-05-14 21:30:26 -07:00

internvl.py

[Perf] Avoid more GPU<->CPU syncs on the model execution path (#51738)

2026-08-11 19:30:34 -07:00

iquest_loopcoder.py

Standardise weight tying on ParallelLMHead.tie_weights (#52147)

2026-08-13 15:10:09 +00:00

isaac.py

Remove more unnecessary load_weights methods (#47058)

2026-06-30 15:22:16 +01:00

jais2.py

Remove unnecessary load_weights methods (#44589)

2026-06-29 01:52:23 -07:00

jamba.py

[MoE Refactor] Rename FusedMoE to FusedMoEFactory (#44941)

2026-07-31 11:28:18 +01:00

jina_vl.py

…

jina.py

[Core] Check for GPU<->CPU syncs during CI (#43107)

2026-08-14 17:42:05 -07:00

kanana_v.py

[Bugfix][Multimodal] Fix video temporal padding estimates (#49030)

2026-07-29 01:57:31 +08:00

keye_vl1_5.py

[EPD] Remove duplicate image preprocessing in EPD and enable preprocess on GPU (#50390)

2026-08-05 15:59:54 -07:00

keye.py

[Perf] Avoid more GPU<->CPU syncs in multimodal encoders (#52369)

2026-08-14 10:40:14 -07:00

kimi_audio.py

Remove more unnecessary load_weights methods (#47058)

2026-06-30 15:22:16 +01:00

kimi_k25_vit.py

[Kimi][MM] disable kimi_vit's dynamic torch.compile for TPU (#51196)

2026-08-08 00:42:51 +00:00

kimi_k25.py

[MM][CG] Support ViT full CUDA graph for Kimi-K2.5 (#50929)

2026-08-04 17:15:56 +00:00

kimi_vl.py

[MM][Perf][CG] Support ViT full CUDA graph for Kimi-VL (#41992)

2026-06-17 12:14:01 +00:00

laguna_dflash.py

Add Laguna XS.2.1 DFlash drafter support (#46853)

2026-07-02 18:09:27 -07:00

laguna.py

[Refactor] Delete dead code in models (#51838)

2026-08-11 15:44:26 -04:00

lfm2_moe.py

[Refactor] Delete dead code in models (#51838)

2026-08-11 15:44:26 -04:00

lfm2_siglip2.py

Remove more unnecessary load_weights methods (#47058)

2026-06-30 15:22:16 +01:00

lfm2_vl.py

[MM][Perf][CG] Support dual-path ViT full CUDA graph for DeepSeek-OCR (#43586)

2026-06-16 04:35:20 -07:00

lfm2.py

[Bugfix] Fix LFM2 ShortConv prefix breaking quant ignore list (#51495)

2026-08-08 11:27:13 +00:00

lightonocr.py

…

llama4_eagle.py

[Bugfix][Model Runner V2] Restore multimodal draft capability detection (#50417)

2026-08-03 14:18:14 -07:00

llama4.py

[MoE Refactor] Rename FusedMoE to FusedMoEFactory (#44941)

2026-07-31 11:28:18 +01:00

llama_eagle3.py

[Bugfix][Spec Decode] Fix eagle3 first-layer qkv_proj prefix for quantized drafts (#48068)

2026-07-16 11:34:41 +08:00

llama_eagle.py

Remove more unnecessary load_weights methods (#47058)

2026-06-30 15:22:16 +01:00

llama.py

[Mypy Fix] Mypy fix for "vllm/model_executor/models/[aA][bB]" (#48977)

2026-08-06 19:30:15 +00:00

llava_next_video.py

[Model] Enable LoRA support for tower and connector in LlavaNextVideo (#48594)

2026-07-14 20:09:38 +08:00

llava_next.py

[Mypy Fix] Mypy fix for "vllm/model_executor/models/[aA][bB]" (#48977)

2026-08-06 19:30:15 +00:00

llava_onevision2.py

[EPD] Remove duplicate image preprocessing in EPD and enable preprocess on GPU (#50390)

2026-08-05 15:59:54 -07:00

llava_onevision.py

[Perf][4/n] Eliminate various GPU<->CPU syncs (#42347)

2026-05-19 10:35:54 -04:00

llava.py

Remove mantis (#46806)

2026-07-01 07:13:58 +00:00

longcat_flash_mtp.py

[Model] Add LongCat-Flash-Lite (n-gram embedding) (#47857)

2026-07-10 07:17:50 -07:00

longcat_flash_ngram.py

[ModelRunnerV2] Support prompt embeds (#42963)

2026-08-17 21:58:50 +00:00

longcat_flash.py

[MoE Refactor] Rename FusedMoE to FusedMoEFactory (#44941)

2026-07-31 11:28:18 +01:00

mamba2.py

Remove unnecessary load_weights methods (#44589)

2026-06-29 01:52:23 -07:00

mamba.py

Standardise weight tying on ParallelLMHead.tie_weights (#52147)

2026-08-13 15:10:09 +00:00

medusa.py

Add Medusa speculative decoding e2e test (#41396)

2026-06-30 18:02:22 +00:00

mellum.py

Standardise weight tying on ParallelLMHead.tie_weights (#52147)

2026-08-13 15:10:09 +00:00

midashenglm.py

Fix MiDashengLM TP>1 crash in audio encoder attention (#44408)

2026-06-09 09:29:14 -07:00

mimo_audio.py

[Model] Add MiMo-V2.5 support (#40967)

2026-04-27 13:26:51 +00:00

mimo_mtp.py

Remove more unnecessary load_weights methods (#47058)

2026-06-30 15:22:16 +01:00

mimo_v2_mtp.py

[SpecDecoding] extend mtp support for mimo 2.5 (#41905)

2026-05-09 18:22:59 +00:00

mimo_v2_omni.py

[Bugfix][MiMo] Apply vision attention sinks in the window attention path (#49815)

2026-08-10 22:36:42 +00:00

mimo_v2.py

[Refactor] Delete dead code in models (#51838)

2026-08-11 15:44:26 -04:00

mimo.py

Standardise weight tying on ParallelLMHead.tie_weights (#52147)

2026-08-13 15:10:09 +00:00

minicpm3.py

[Perf] RMSNorm uncontiguous support, 1.2~3.1x kernel performance improvement (#49750)

2026-07-29 13:45:38 -07:00

minicpm_eagle.py

[Bugfix][UT]Fix EagleMiniCPMForCausalLM meet TypeError (#48452)

2026-07-13 04:37:23 +00:00

minicpm.py

[MoE] Remove inplace fused experts mechanism (#43727)

2026-05-27 20:00:19 -07:00

minicpmo.py

[Bugfix] Fixes MiniCPM-O resampler device placement to avoid tensor device mismatch (#42332)

2026-06-17 08:35:27 +00:00

minicpmv4_6.py

[Revert][Bugfix] Restore MiniCPM-V 4.6 ViT QKV weight loader (#49193)

2026-07-20 18:17:46 +00:00

minicpmv.py

[Core] Check for GPU<->CPU syncs during CI (#43107)

2026-08-14 17:42:05 -07:00

minimax_m2.py

[MoE Refactor] Rename FusedMoE to FusedMoEFactory (#44941)

2026-07-31 11:28:18 +01:00

mistral3.py

[Bugfix][Model] Fix Devstral Small 2 HF format weight loading (#39293)

2026-04-14 10:11:18 +00:00

mistral_eagle.py

[Bugfix][Model Runner V2] Restore multimodal draft capability detection (#50417)

2026-08-03 14:18:14 -07:00

mistral_large_3_eagle.py

[Bugfix][Model Runner V2] Restore multimodal draft capability detection (#50417)

2026-08-03 14:18:14 -07:00

mistral_large_3.py

[Model] Migrate MistralLarge3ForCausalLM to AutoWeightsLoader (#48153)

2026-07-10 12:27:58 +00:00

mistral.py

…

mixtral.py

Standardise weight tying on ParallelLMHead.tie_weights (#52147)

2026-08-13 15:10:09 +00:00

mllama4.py

Remove boilerplate missed by #46820 (#46956)

2026-06-29 08:11:17 +08:00

mlp_speculator.py

[Refactor] Remove unused dead code (#38842)

2026-04-06 11:52:05 -04:00

modernbert.py

Correct model layer aliasing for Bert style models (#43896)

2026-07-09 19:46:22 -04:00

module_mapping.py

…

molmo2.py

Remove even more unnecessary load_weights methods (#48496)

2026-07-18 08:40:27 +00:00

molmo.py

Standardise weight tying on ParallelLMHead.tie_weights (#52147)

2026-08-13 15:10:09 +00:00

moondream3.py

[Model] Add Moondream3 model support(only query and caption skills) (#32325)

2026-05-01 10:06:48 +08:00

moonvit.py

[Core] Check for GPU<->CPU syncs during CI (#43107)

2026-08-14 17:42:05 -07:00

moss_audio.py

[Core] Check for GPU<->CPU syncs during CI (#43107)

2026-08-14 17:42:05 -07:00

moss_transcribe_diarize.py

[Frontend] Add diarized_json support for MOSS-Transcribe-Diarize (#48543)

2026-07-29 23:24:30 +00:00

mpt.py

Remove unnecessary load_weights methods (#44589)

2026-06-29 01:52:23 -07:00

muse_glimmer.py

Add Muse Glimmer model support (#51655)

2026-08-14 12:28:56 +08:00

nano_nemotron_vl.py

fix(security): enforce audio decode duration limit in NanoNemotronVL (#50221)

2026-08-13 03:31:55 -07:00

nemotron_h_mtp.py

fix: ensure no double load of lm head in nemotron mtp (#47440)

2026-07-07 12:01:45 +00:00

nemotron_h.py

[Refactor] Delete dead code in models (#51838)

2026-08-11 15:44:26 -04:00

nemotron_nas.py

Remove unnecessary load_weights methods (#44589)

2026-06-29 01:52:23 -07:00

nemotron_parse.py

[Refactor] Remove multiple dead codes (#50285)

2026-08-03 10:59:08 -04:00

nemotron_vl.py

fix(quantization): Fix AWQ dequantize on Intel XPU and refactor AutoAWQ config (#42727)

2026-06-18 20:12:28 +08:00

nemotron.py

Standardise weight tying on ParallelLMHead.tie_weights (#52147)

2026-08-13 15:10:09 +00:00

nvlm_d.py

[Model] Remove Mono-InternVL (InternLM2VEForCausalLM) (#45129)

2026-06-12 10:41:09 +00:00

olmo3.py

Standardise weight tying on ParallelLMHead.tie_weights (#52147)

2026-08-13 15:10:09 +00:00

olmo_hybrid.py

Standardise weight tying on ParallelLMHead.tie_weights (#52147)

2026-08-13 15:10:09 +00:00

olmoe.py

[MoE Refactor] Rename FusedMoE to FusedMoEFactory (#44941)

2026-07-31 11:28:18 +01:00

openai_privacy_filter.py

[Model] Add support for openai/privacy-filter (#41026)

2026-06-28 16:18:22 -07:00

opencua.py

[EPD] Remove duplicate image preprocessing in EPD and enable preprocess on GPU (#50390)

2026-08-05 15:59:54 -07:00

openpangu_mtp.py

[MoE Refactor] Rename FusedMoE.make_expert_params_mapping to fused_moe_make_expert_params_mapping (#40671)

2026-04-23 11:22:34 -04:00

openpangu_vl.py

[Quant] Consolidate GPTQ: rename gptq_marlin.py to auto_gptq.py (#38288)

2026-05-15 08:25:52 +08:00

openpangu.py

Standardise weight tying on ParallelLMHead.tie_weights (#52147)

2026-08-13 15:10:09 +00:00

openvla.py

[Model] Openvla support (#42654)

2026-05-19 08:17:42 -07:00

opt.py

Standardise weight tying on ParallelLMHead.tie_weights (#52147)

2026-08-13 15:10:09 +00:00

orion.py

Standardise weight tying on ParallelLMHead.tie_weights (#52147)

2026-08-13 15:10:09 +00:00

ovis2_5.py

[Bugfix] Fix Ovis2_5 special tokens for transformers v5 (#47298)

2026-07-21 08:09:20 +00:00

ovis.py

[Model] Fix missing maybe_prefix (#42280)

2026-05-11 09:04:06 +00:00

paddleocr_vl.py

[Perf] Avoid more GPU<->CPU syncs in multimodal encoders (#52369)

2026-08-14 10:40:14 -07:00

paligemma.py

…

parakeet.py

Bugfix: Parakeet: .conv.pointwise/depthwise_conv1/2.bias weigths can exist even if convolution_bias=False (#40007)

2026-04-16 23:22:05 +00:00

param2moe.py

Standardise weight tying on ParallelLMHead.tie_weights (#52147)

2026-08-13 15:10:09 +00:00

phi3.py

…

phi3v.py

[Perf] Avoid more GPU<->CPU syncs on the model execution path (#51738)

2026-08-11 19:30:34 -07:00

phi4mm_audio.py

[Perf][4/n] Eliminate various GPU<->CPU syncs (#42347)

2026-05-19 10:35:54 -04:00

phi4mm_utils.py

[Refactor] Remove multiple dead codes (#50285)

2026-08-03 10:59:08 -04:00

phi4mm.py

[Core] Check for GPU<->CPU syncs during CI (#43107)

2026-08-14 17:42:05 -07:00

phi4siglip.py

[Model] Add Phi4ForCausalLMV for microsoft/Phi-4-reasoning-vision-15B (#38306)

2026-04-02 21:14:57 -07:00

phi.py

Remove unnecessary load_weights methods (#44589)

2026-06-29 01:52:23 -07:00

phimoe.py

[MoE Refactor] Rename FusedMoE to FusedMoEFactory (#44941)

2026-07-31 11:28:18 +01:00

pixtral.py

[BUGFIX] Fix Pixtral consolidated format vision weight loading (#39916)

2026-04-19 22:25:03 -07:00

plamo3.py

…

qianfan_ocr.py

[Model] support Qianfan-OCR model (#40136)

2026-05-05 10:51:25 +00:00

qwen2_5_omni_thinker.py

[Core] Check for GPU<->CPU syncs during CI (#43107)

2026-08-14 17:42:05 -07:00

qwen2_5_vl.py

[Core] Check for GPU<->CPU syncs during CI (#43107)

2026-08-14 17:42:05 -07:00

qwen2_audio.py

[Core] Check for GPU<->CPU syncs during CI (#43107)

2026-08-14 17:42:05 -07:00

qwen2_moe.py

Standardise weight tying on ParallelLMHead.tie_weights (#52147)

2026-08-13 15:10:09 +00:00

qwen2_rm.py

Remove unnecessary load_weights methods (#44589)

2026-06-29 01:52:23 -07:00

qwen2_vl.py

[Model] Fused mm preprocess normalisation on the Device (#50411)

2026-08-06 03:11:14 -07:00

qwen2.py

Standardise weight tying on ParallelLMHead.tie_weights (#52147)

2026-08-13 15:10:09 +00:00

qwen3_5_mtp.py

[Bugfix][Refactor] Keep Qwen3Next layer boundaries sequence parallel (#50685)

2026-08-14 11:58:32 +08:00

qwen3_5.py

Standardise weight tying on ParallelLMHead.tie_weights (#52147)

2026-08-13 15:10:09 +00:00

qwen3_asr_forced_aligner.py

…

qwen3_asr_realtime.py

…

qwen3_asr.py

[Core] Check for GPU<->CPU syncs during CI (#43107)

2026-08-14 17:42:05 -07:00

qwen3_dflash.py

Add Muse Glimmer model support (#51655)

2026-08-14 12:28:56 +08:00

qwen3_dspark.py

[Spec Decode] DSpark confidence-scheduled verification (#47808)

2026-08-12 12:27:48 -07:00

qwen3_eagle3.py

Added sliding window attention support for qwen-eagle3 architecture (#47568)

2026-07-13 20:20:44 +00:00

qwen3_moe.py

Standardise weight tying on ParallelLMHead.tie_weights (#52147)

2026-08-13 15:10:09 +00:00

qwen3_next_mtp.py

[Bugfix][Refactor] Keep Qwen3Next layer boundaries sequence parallel (#50685)

2026-08-14 11:58:32 +08:00

qwen3_next.py

[Bugfix][Refactor] Keep Qwen3Next layer boundaries sequence parallel (#50685)

2026-08-14 11:58:32 +08:00

qwen3_omni_moe_thinker.py

[Core] Check for GPU<->CPU syncs during CI (#43107)

2026-08-14 17:42:05 -07:00

qwen3_vl_moe.py

Standardise weight tying on ParallelLMHead.tie_weights (#52147)

2026-08-13 15:10:09 +00:00

qwen3_vl.py

[Core] Check for GPU<->CPU syncs during CI (#43107)

2026-08-14 17:42:05 -07:00

qwen3.py

Standardise weight tying on ParallelLMHead.tie_weights (#52147)

2026-08-13 15:10:09 +00:00

radio.py

[Refactor] Remove dead code in multiple files (#49745)

2026-07-27 15:58:26 -04:00

registry.py

Add Muse Glimmer model support (#51655)

2026-08-14 12:28:56 +08:00

rnj1.py

Remove unnecessary load_weights methods (#44589)

2026-06-29 01:52:23 -07:00

roberta.py

[Model] Add RobertaForTokenClassification / XLMRobertaForTokenClassification (#47991)

2026-07-15 14:30:33 +00:00

rvl.py

…

sarvam.py

Standardise weight tying on ParallelLMHead.tie_weights (#52147)

2026-08-13 15:10:09 +00:00

seed_oss.py

Standardise weight tying on ParallelLMHead.tie_weights (#52147)

2026-08-13 15:10:09 +00:00

siglip2navit.py

Remove more unnecessary load_weights methods (#47058)

2026-06-30 15:22:16 +01:00

siglip.py

[Core] Check for GPU<->CPU syncs during CI (#43107)

2026-08-14 17:42:05 -07:00

skyworkr1v.py

[Refactor] Remove multiple dead codes (#50285)

2026-08-03 10:59:08 -04:00

smolvlm.py

…

solar.py

Standardise weight tying on ParallelLMHead.tie_weights (#52147)

2026-08-13 15:10:09 +00:00

stablelm.py

Standardise weight tying on ParallelLMHead.tie_weights (#52147)

2026-08-13 15:10:09 +00:00

step1.py

Remove unnecessary load_weights methods (#44589)

2026-06-29 01:52:23 -07:00

step3_text.py

[MoE Refactor] Rename FusedMoE to FusedMoEFactory (#44941)

2026-07-31 11:28:18 +01:00

step3_vl.py

[Perf] Avoid more GPU<->CPU syncs on the model execution path (#51738)

2026-08-11 19:30:34 -07:00

step3p5_mtp.py

fix(step3p5-mtp): honor exclude_modules for the MTP head via prefix (#48883)

2026-07-29 16:58:29 +00:00

step3p5.py

[Refactor] Delete dead code in models (#51838)

2026-08-11 15:44:26 -04:00

step3p7.py

[Model]Support Step-3.7-Flash (#43859)

2026-05-28 17:01:48 -07:00

step_vl.py

[MM][CG] Enable encoder Cudagraph for Step3VL (#42224)

2026-05-17 20:19:13 -07:00

telechat2.py

Remove even more unnecessary load_weights methods (#48496)

2026-07-18 08:40:27 +00:00

teleflm.py

…

terratorch.py

[Core] Check for GPU<->CPU syncs during CI (#43107)

2026-08-14 17:42:05 -07:00

ultravox.py

[Core] Check for GPU<->CPU syncs during CI (#43107)

2026-08-14 17:42:05 -07:00

unlimited_ocr.py

Add Triton Backend for Unlimited-OCR R-SWA (#47102)

2026-07-03 00:11:50 -07:00

utils.py

[Refactor] Remove multiple dead codes (#50285)

2026-08-03 10:59:08 -04:00

vision.py

replace batch_norm to numerically identical without cudnn (#51734)

2026-08-11 12:43:20 +08:00

voxtral_realtime.py

[Bugfix][Voxtral Realtime] Fix token feedback timeout silent hang (#44461)

2026-07-05 05:42:36 -07:00

voxtral.py

[Perf] Avoid some more unnecessary GPU<->CPU syncs (#51458)

2026-08-08 17:13:26 -07:00

voyage.py

[Model] Use AutoWeightsLoader for Voyage (#42972)

2026-05-22 05:28:23 +00:00

whisper_causal.py

[Bugfix] Fix pooled Whisper encoder sliding-window kernel size (#47437)

2026-07-02 10:19:11 -07:00

whisper_utils.py

…

whisper.py

[Model][LoRA] Add tower/connector LoRA support for Ultravox (#48215)

2026-08-12 22:57:59 -07:00

zamba2.py

Remove more unnecessary load_weights methods (#47058)

2026-06-30 15:22:16 +01:00