Matthew Bonanni 98f60e5acb [6/N][Attention] Move utils to more appropriate locations (#32215) ...

Signed-off-by: Matthew Bonanni <mbonanni@redhat.com>

2026-01-13 05:38:52 -08:00

..

transformers

[Model] Avoid hardcoding pooling type (#32119)

2026-01-11 21:28:12 -08:00

__init__.py

…

adapters.py

[Bugfix] Fix Qwen3-VL-Reranker model loading for sequence classification (#32089)

2026-01-10 12:40:09 -08:00

afmoe.py

[1/N][Attention] Restructure attention: move files (#31916)

2026-01-09 13:10:24 -08:00

aimv2.py

[1/N][Attention] Restructure attention: move files (#31916)

2026-01-09 13:10:24 -08:00

apertus.py

[1/N][Attention] Restructure attention: move files (#31916)

2026-01-09 13:10:24 -08:00

arcee.py

…

arctic.py

Standardise get_rope to use rope_parameters["partial_rotary_factor"], not rotary_dim (#30389)

2025-12-11 20:45:23 +00:00

aria.py

[Misc][Model][Refactor] Pass the prefix into Linear layers (#31669)

2026-01-05 20:03:18 +00:00

audioflamingo3.py

[Bugfix] Enable dynamic_dims for different embeds shape (#31223)

2025-12-24 10:15:47 +08:00

aya_vision.py

…

bagel.py

[Bugfix] Fix BAGEL online serving for text and image understanding (#31546)

2025-12-31 14:46:10 -08:00

baichuan.py

Standardise get_rope to use rope_parameters["partial_rotary_factor"], not rotary_dim (#30389)

2025-12-11 20:45:23 +00:00

bailing_moe.py

[BugFix] LoRA: Support loading base_layer of experts (#31104)

2026-01-07 14:49:39 +08:00

bamba.py

Standardise get_rope to use rope_parameters["partial_rotary_factor"], not rotary_dim (#30389)

2025-12-11 20:45:23 +00:00

bee.py

…

bert_with_rope.py

[Model] Standardize pooling heads (#32148)

2026-01-12 17:01:49 +00:00

bert.py

[Bugfix] Replace PoolingParams.normalize with use_activation (#32243)

2026-01-13 10:45:42 +00:00

blip2.py

[Model] Enable LoRA support for BLIP2 (#31620)

2026-01-05 08:02:24 +00:00

blip.py

[1/N][Attention] Restructure attention: move files (#31916)

2026-01-09 13:10:24 -08:00

bloom.py

…

chameleon.py

Standardise get_rope to use rope_parameters["partial_rotary_factor"], not rotary_dim (#30389)

2025-12-11 20:45:23 +00:00

chatglm.py

Standardise get_rope to use rope_parameters["partial_rotary_factor"], not rotary_dim (#30389)

2025-12-11 20:45:23 +00:00

clip.py

[Refactor] Separate sequence and token pooling types (#32026)

2026-01-10 04:53:24 +00:00

cohere2_vision.py

…

commandr.py

Standardise get_rope to use rope_parameters["partial_rotary_factor"], not rotary_dim (#30389)

2025-12-11 20:45:23 +00:00

config.py

[Bugfix] Fix Qwen3-VL-Reranker model loading for sequence classification (#32089)

2026-01-10 12:40:09 -08:00

dbrx.py

Standardise get_rope to use rope_parameters["partial_rotary_factor"], not rotary_dim (#30389)

2025-12-11 20:45:23 +00:00

deepencoder.py

[1/N][Attention] Restructure attention: move files (#31916)

2026-01-09 13:10:24 -08:00

deepseek_eagle.py

[BugFix] LoRA: Support loading base_layer of experts (#31104)

2026-01-07 14:49:39 +08:00

deepseek_mtp.py

[Model] Remove incorrect SupportsPP from MTP models (#32150)

2026-01-12 01:19:30 -08:00

deepseek_ocr.py

[Model] Remove redundant None check in DeepSeekOCR image input processing (#32016)

2026-01-09 06:12:44 -08:00

deepseek_v2.py

[1/N][Attention] Restructure attention: move files (#31916)

2026-01-09 13:10:24 -08:00

deepseek_vl2.py

Revert "[Renderer] Separate out RendererConfig from ModelConfig (#30145)" (#30199)

2025-12-07 00:00:22 -08:00

dots1.py

[BugFix] LoRA: Support loading base_layer of experts (#31104)

2026-01-07 14:49:39 +08:00

dots_ocr.py

[1/N][Attention] Restructure attention: move files (#31916)

2026-01-09 13:10:24 -08:00

ernie45_moe.py

[BugFix] LoRA: Support loading base_layer of experts (#31104)

2026-01-07 14:49:39 +08:00

ernie45_vl_moe.py

[BugFix] LoRA: Support loading base_layer of experts (#31104)

2026-01-07 14:49:39 +08:00

ernie45_vl.py

[1/N][Attention] Restructure attention: move files (#31916)

2026-01-09 13:10:24 -08:00

ernie45.py

…

ernie_mtp.py

[Model] Remove incorrect SupportsPP from MTP models (#32150)

2026-01-12 01:19:30 -08:00

exaone4.py

Add K-EXAONE-236B-A23B (#31621)

2026-01-12 16:30:50 +00:00

exaone_moe_mtp.py

Add K-EXAONE-236B-A23B (#31621)

2026-01-12 16:30:50 +00:00

exaone_moe.py

[BugFix] fix FusedMoE.make_expert_params_mapping in EXAONE-MoE (#32196)

2026-01-12 10:00:45 -08:00

exaone.py

Standardise get_rope to use rope_parameters["partial_rotary_factor"], not rotary_dim (#30389)

2025-12-11 20:45:23 +00:00

fairseq2_llama.py

…

falcon_h1.py

Standardise get_rope to use rope_parameters["partial_rotary_factor"], not rotary_dim (#30389)

2025-12-11 20:45:23 +00:00

falcon.py

Standardise get_rope to use rope_parameters["partial_rotary_factor"], not rotary_dim (#30389)

2025-12-11 20:45:23 +00:00

flex_olmo.py

…

fuyu.py

…

gemma2.py

Standardise get_rope to use rope_parameters["partial_rotary_factor"], not rotary_dim (#30389)

2025-12-11 20:45:23 +00:00

gemma3_mm.py

[Bugfix] Schedule failure due to wrong get_image_size_with_most_features (#29692)

2025-12-12 02:27:20 -08:00

gemma3.py

[1/N][Attention] Restructure attention: move files (#31916)

2026-01-09 13:10:24 -08:00

gemma3n_audio_utils.py

[Bugfix] Fix integer overflow in Gemma3n audio processing (#31657)

2026-01-10 17:52:53 +08:00

gemma3n_mm.py

[Bugfix] Fix integer overflow in Gemma3n audio processing (#31657)

2026-01-10 17:52:53 +08:00

gemma3n.py

Standardise get_rope to use rope_parameters["partial_rotary_factor"], not rotary_dim (#30389)

2025-12-11 20:45:23 +00:00

gemma.py

Standardise get_rope to use rope_parameters["partial_rotary_factor"], not rotary_dim (#30389)

2025-12-11 20:45:23 +00:00

glm4_1v.py

[1/N][Attention] Restructure attention: move files (#31916)

2026-01-09 13:10:24 -08:00

glm4_moe_mtp.py

[Model] Remove incorrect SupportsPP from MTP models (#32150)

2026-01-12 01:19:30 -08:00

glm4_moe.py

[Model] Cleanup: Remove redundant manual definition of make_empty_intermediate_tensors in GLM-4-MoE (#31869)

2026-01-07 08:18:28 +00:00

glm4.py

[1/N][Attention] Restructure attention: move files (#31916)

2026-01-09 13:10:24 -08:00

glm4v.py

[1/N][Attention] Restructure attention: move files (#31916)

2026-01-09 13:10:24 -08:00

glm.py

…

glmasr_utils.py

[Refactor] GLM-ASR Modeling (#31779)

2026-01-07 13:08:29 +00:00

glmasr.py

[1/N][Attention] Restructure attention: move files (#31916)

2026-01-09 13:10:24 -08:00

gpt2.py

[Model] Reorganize pooling layers (#31973)

2026-01-09 11:02:14 +00:00

gpt_bigcode.py

…

gpt_j.py

Standardise get_rope to use rope_parameters["partial_rotary_factor"], not rotary_dim (#30389)

2025-12-11 20:45:23 +00:00

gpt_neox.py

[Misc][Model][Refactor] Pass the prefix into Linear layers (#31669)

2026-01-05 20:03:18 +00:00

gpt_oss.py

[1/N][Attention] Restructure attention: move files (#31916)

2026-01-09 13:10:24 -08:00

granite_speech.py

[Bugfix] Fix Var Length Batched Padding in Granite Speech (#31906)

2026-01-09 10:28:43 +00:00

granite.py

Standardise get_rope to use rope_parameters["partial_rotary_factor"], not rotary_dim (#30389)

2025-12-11 20:45:23 +00:00

granitemoe.py

[BugFix] LoRA: Support loading base_layer of experts (#31104)

2026-01-07 14:49:39 +08:00

granitemoehybrid.py

Standardise get_rope to use rope_parameters["partial_rotary_factor"], not rotary_dim (#30389)

2025-12-11 20:45:23 +00:00

granitemoeshared.py

…

gritlm.py

[Model] Standardize pooling heads (#32148)

2026-01-12 17:01:49 +00:00

grok1.py

[Model] Add Grok-2 (#31847)

2026-01-08 04:59:48 -08:00

h2ovl.py

…

hunyuan_v1.py

[1/N][Attention] Restructure attention: move files (#31916)

2026-01-09 13:10:24 -08:00

hunyuan_vision.py

[1/N][Attention] Restructure attention: move files (#31916)

2026-01-09 13:10:24 -08:00

hyperclovax_vision.py

[Model] Standardize common vision encoders (#31947)

2026-01-08 02:33:16 -08:00

idefics2_vision_model.py

[1/N][Attention] Restructure attention: move files (#31916)

2026-01-09 13:10:24 -08:00

idefics3.py

[Core] Initialize LoRA support for tower and connector in multi-modal models (#26674)

2025-12-26 04:48:20 -08:00

interfaces_base.py

[Refactor] Separate sequence and token pooling types (#32026)

2026-01-10 04:53:24 +00:00

interfaces.py

[Core] Initialize LoRA support for tower and connector in multi-modal models (#26674)

2025-12-26 04:48:20 -08:00

intern_vit.py

[1/N][Attention] Restructure attention: move files (#31916)

2026-01-09 13:10:24 -08:00

internlm2_ve.py

…

internlm2.py

[Refactor] Separate sequence and token pooling types (#32026)

2026-01-10 04:53:24 +00:00

interns1_vit.py

[1/N][Attention] Restructure attention: move files (#31916)

2026-01-09 13:10:24 -08:00

interns1.py

Revert "[Renderer] Separate out RendererConfig from ModelConfig (#30145)" (#30199)

2025-12-07 00:00:22 -08:00

internvl.py

[Misc] Set default torch num threads for input processing (#31879)

2026-01-12 10:28:16 -08:00

iquest_loopcoder.py

[1/N][Attention] Restructure attention: move files (#31916)

2026-01-09 13:10:24 -08:00

isaac.py

feature/issac 0.2 (#31550)

2026-01-10 03:18:05 +00:00

jais2.py

[Bugfix] Fix Jais2ForCausalLM (#31198)

2025-12-23 07:44:01 +00:00

jais.py

…

jamba.py

[Model] Reorganize pooling layers (#31973)

2026-01-09 11:02:14 +00:00

jina_vl.py

[Model] Reorganize pooling layers (#31973)

2026-01-09 11:02:14 +00:00

kanana_v.py

[MODEL] New model support for kakaocorp/kanana-1.5-v-3b-instruct (#29384)

2026-01-12 16:39:02 +00:00

keye_vl1_5.py

…

keye.py

[1/N][Attention] Restructure attention: move files (#31916)

2026-01-09 13:10:24 -08:00

kimi_linear.py

[BugFix] LoRA: Support loading base_layer of experts (#31104)

2026-01-07 14:49:39 +08:00

kimi_vl.py

[BugFix] LoRA: Support loading base_layer of experts (#31104)

2026-01-07 14:49:39 +08:00

lfm2_moe.py

[BugFix] LoRA: Support loading base_layer of experts (#31104)

2026-01-07 14:49:39 +08:00

lfm2_vl.py

[Model] Add LFM2-VL model support (#31758)

2026-01-08 05:00:27 -08:00

lfm2.py

Standardise get_rope to use rope_parameters["partial_rotary_factor"], not rotary_dim (#30389)

2025-12-11 20:45:23 +00:00

lightonocr.py

[Model] Standardize common vision encoders (#31947)

2026-01-08 02:33:16 -08:00

llama4_eagle.py

…

llama4.py

[1/N][Attention] Restructure attention: move files (#31916)

2026-01-09 13:10:24 -08:00

llama_eagle3.py

…

llama_eagle.py

…

llama.py

[Misc][LLaMa4] Compile LLaMa Vision Encoder (#30709)

2026-01-09 22:01:38 -05:00

llava_next_video.py

[Model] Standardize common vision encoders (#31947)

2026-01-08 02:33:16 -08:00

llava_next.py

[Model] Standardize common vision encoders (#31947)

2026-01-08 02:33:16 -08:00

llava_onevision.py

[Model] Standardize common vision encoders (#31947)

2026-01-08 02:33:16 -08:00

llava.py

[Model] Standardize common vision encoders (#31947)

2026-01-08 02:33:16 -08:00

longcat_flash_mtp.py

[Model] Remove incorrect SupportsPP from MTP models (#32150)

2026-01-12 01:19:30 -08:00

longcat_flash.py

[BugFix] LoRA: Support loading base_layer of experts (#31104)

2026-01-07 14:49:39 +08:00

mamba2.py

…

mamba.py

…

medusa.py

…

midashenglm.py

…

mimo_mtp.py

…

mimo_v2_flash.py

[1/N][Attention] Restructure attention: move files (#31916)

2026-01-09 13:10:24 -08:00

mimo.py

…

minicpm3.py

Standardise get_rope to use rope_parameters["partial_rotary_factor"], not rotary_dim (#30389)

2025-12-11 20:45:23 +00:00

minicpm_eagle.py

[Misc][Model][Refactor] Pass the prefix into Linear layers (#31669)

2026-01-05 20:03:18 +00:00

minicpm.py

[Misc][Model][Refactor] Pass the prefix into Linear layers (#31669)

2026-01-05 20:03:18 +00:00

minicpmo.py

…

minicpmv.py

[Bugfix] Enable dynamic_dims for different embeds shape (#31223)

2025-12-24 10:15:47 +08:00

minimax_m2.py

[BugFix] LoRA: Support loading base_layer of experts (#31104)

2026-01-07 14:49:39 +08:00

minimax_text_01.py

[1/N][Attention] Restructure attention: move files (#31916)

2026-01-09 13:10:24 -08:00

minimax_vl_01.py

[Model] Standardize common vision encoders (#31947)

2026-01-08 02:33:16 -08:00

mistral3.py

[Model] Standardize common vision encoders (#31947)

2026-01-08 02:33:16 -08:00

mistral_large_3_eagle.py

[Misc][Model][Refactor] Pass the prefix into Linear layers (#31669)

2026-01-05 20:03:18 +00:00

mistral_large_3.py

…

mixtral.py

[BugFix] LoRA: Support loading base_layer of experts (#31104)

2026-01-07 14:49:39 +08:00

mllama4.py

[Misc][LLaMa4] Compile LLaMa Vision Encoder (#30709)

2026-01-09 22:01:38 -05:00

mlp_speculator.py

…

modernbert.py

[Bugfix] Replace PoolingParams.normalize with use_activation (#32243)

2026-01-13 10:45:42 +00:00

module_mapping.py

…

molmo.py

[1/N][Attention] Restructure attention: move files (#31916)

2026-01-09 13:10:24 -08:00

moonvit.py

[1/N][Attention] Restructure attention: move files (#31916)

2026-01-09 13:10:24 -08:00

mpt.py

…

nano_nemotron_vl.py

[Model] Nemotron Parse 1.1 Support (#30864)

2026-01-05 13:00:14 -08:00

nemotron_h.py

Enable quantized attention in NemotronH models (#31898)

2026-01-07 17:37:19 +00:00

nemotron_nas.py

[1/N][Attention] Restructure attention: move files (#31916)

2026-01-09 13:10:24 -08:00

nemotron_parse.py

[Refactor] Remove get_encoder_dummy_data (#32241)

2026-01-13 09:21:23 +00:00

nemotron_vl.py

Revert "[Renderer] Separate out RendererConfig from ModelConfig (#30145)" (#30199)

2025-12-07 00:00:22 -08:00

nemotron.py

Standardise get_rope to use rope_parameters["partial_rotary_factor"], not rotary_dim (#30389)

2025-12-11 20:45:23 +00:00

nvlm_d.py

…

olmo2.py

Standardise get_rope to use rope_parameters["partial_rotary_factor"], not rotary_dim (#30389)

2025-12-11 20:45:23 +00:00

olmo.py

Standardise get_rope to use rope_parameters["partial_rotary_factor"], not rotary_dim (#30389)

2025-12-11 20:45:23 +00:00

olmoe.py

[BugFix] LoRA: Support loading base_layer of experts (#31104)

2026-01-07 14:49:39 +08:00

opencua.py

[CustomOp][MM] Extract MMEncoderAttention as CustomOp and replace the backend of QwenVisionAttention with it. (#30125)

2025-12-15 11:13:32 +08:00

openpangu_mtp.py

[Model] Remove incorrect SupportsPP from MTP models (#32150)

2026-01-12 01:19:30 -08:00

openpangu.py

[1/N][Attention] Restructure attention: move files (#31916)

2026-01-09 13:10:24 -08:00

opt.py

…

orion.py

Standardise get_rope to use rope_parameters["partial_rotary_factor"], not rotary_dim (#30389)

2025-12-11 20:45:23 +00:00

ouro.py

[1/N][Attention] Restructure attention: move files (#31916)

2026-01-09 13:10:24 -08:00

ovis2_5.py

[CustomOp][MM] Extract MMEncoderAttention as CustomOp and replace the backend of QwenVisionAttention with it. (#30125)

2025-12-15 11:13:32 +08:00

ovis.py

…

paddleocr_vl.py

[1/N][Attention] Restructure attention: move files (#31916)

2026-01-09 13:10:24 -08:00

paligemma.py

[Model] Enable LoRA support for PaliGemma (#31656)

2026-01-07 10:09:32 +08:00

persimmon.py

Standardise get_rope to use rope_parameters["partial_rotary_factor"], not rotary_dim (#30389)

2025-12-11 20:45:23 +00:00

phi3.py

…

phi3v.py

[Model] Standardize common vision encoders (#31947)

2026-01-08 02:33:16 -08:00

phi4mm_audio.py

…

phi4mm_utils.py

…

phi4mm.py

[Docs] Improve malformed exception caused by backslash line continuations (#31694)

2026-01-05 17:51:54 -08:00

phi.py

Standardise get_rope to use rope_parameters["partial_rotary_factor"], not rotary_dim (#30389)

2025-12-11 20:45:23 +00:00

phimoe.py

[BugFix] LoRA: Support loading base_layer of experts (#31104)

2026-01-07 14:49:39 +08:00

pixtral.py

[Model] Enable LoRA support for Pixtral (#31724)

2026-01-08 05:00:57 -08:00

plamo2.py

[ROCm][Bugfix] Fix Mamba batched decode producing incorrect output (#32099)

2026-01-13 05:46:53 +00:00

plamo3.py

Support LoRA and GPTQModel for PLaMo 2/3 (#31322)

2025-12-26 11:41:33 +08:00

qwen2_5_omni_thinker.py

[Fix] Introduce audio channels spec (#31595)

2026-01-09 19:34:51 +00:00

qwen2_5_vl.py

[Model] Use mm_position to compute mrope positions for Qwen2-VL/2.5-VL (#32126)

2026-01-13 09:04:29 +00:00

qwen2_audio.py

[Fix] Introduce audio channels spec (#31595)

2026-01-09 19:34:51 +00:00

qwen2_moe.py

[BugFix] LoRA: Support loading base_layer of experts (#31104)

2026-01-07 14:49:39 +08:00

qwen2_rm.py

[Refactor] Separate sequence and token pooling types (#32026)

2026-01-10 04:53:24 +00:00

qwen2_vl.py

[Model] Use mm_position to compute mrope positions for Qwen2-VL/2.5-VL (#32126)

2026-01-13 09:04:29 +00:00

qwen2.py

[1/N][Attention] Restructure attention: move files (#31916)

2026-01-09 13:10:24 -08:00

qwen3_moe.py

[BugFix] LoRA: Support loading base_layer of experts (#31104)

2026-01-07 14:49:39 +08:00

qwen3_next_mtp.py

[Model] Remove incorrect SupportsPP from MTP models (#32150)

2026-01-12 01:19:30 -08:00

qwen3_next.py

[BUGFIX] Add missed remaping of the names of fp8 kv-scale (#32199)

2026-01-12 20:42:06 +00:00

qwen3_omni_moe_thinker.py

[1/N][Attention] Restructure attention: move files (#31916)

2026-01-09 13:10:24 -08:00

qwen3_vl_moe.py

[Bugfix] Fix multimodal configuration for Qwen3VL MOE model (#30670)

2025-12-15 14:06:01 +00:00

qwen3_vl.py

[1/N][Attention] Restructure attention: move files (#31916)

2026-01-09 13:10:24 -08:00

qwen3.py

[1/N][Attention] Restructure attention: move files (#31916)

2026-01-09 13:10:24 -08:00

qwen_vl.py

[Misc][Model][Refactor] Pass the prefix into Linear layers (#31669)

2026-01-05 20:03:18 +00:00

qwen.py

[Misc][Model][Refactor] Pass the prefix into Linear layers (#31669)

2026-01-05 20:03:18 +00:00

radio.py

[Model] Nemotron Parse 1.1 Support (#30864)

2026-01-05 13:00:14 -08:00

registry.py

[Model] Handle trust_remote_code for transformers backend (#32194)

2026-01-13 09:30:12 +08:00

roberta.py

[Model] Avoid hardcoding pooling type (#32119)

2026-01-11 21:28:12 -08:00

rvl.py

…

seed_oss.py

[1/N][Attention] Restructure attention: move files (#31916)

2026-01-09 13:10:24 -08:00

siglip2.py

[1/N][Attention] Restructure attention: move files (#31916)

2026-01-09 13:10:24 -08:00

siglip2navit.py

[1/N][Attention] Restructure attention: move files (#31916)

2026-01-09 13:10:24 -08:00

siglip.py

[Refactor] Separate sequence and token pooling types (#32026)

2026-01-10 04:53:24 +00:00

skyworkr1v.py

…

smolvlm.py

…

solar.py

Standardise get_rope to use rope_parameters["partial_rotary_factor"], not rotary_dim (#30389)

2025-12-11 20:45:23 +00:00

stablelm.py

Standardise get_rope to use rope_parameters["partial_rotary_factor"], not rotary_dim (#30389)

2025-12-11 20:45:23 +00:00

starcoder2.py

Standardise get_rope to use rope_parameters["partial_rotary_factor"], not rotary_dim (#30389)

2025-12-11 20:45:23 +00:00

step3_text.py

Standardise get_rope to use rope_parameters["partial_rotary_factor"], not rotary_dim (#30389)

2025-12-11 20:45:23 +00:00

step3_vl.py

[1/N][Attention] Restructure attention: move files (#31916)

2026-01-09 13:10:24 -08:00

swin.py

Remove head_mask from Ultravox and Swin (#30764)

2025-12-16 08:02:41 -08:00

tarsier.py

[Model] Standardize common vision encoders (#31947)

2026-01-08 02:33:16 -08:00

telechat2.py

…

teleflm.py

…

terratorch.py

[Model] Reorganize pooling layers (#31973)

2026-01-09 11:02:14 +00:00

ultravox.py

[Fix] Introduce audio channels spec (#31595)

2026-01-09 19:34:51 +00:00

utils.py

…

vision.py

[1/N][Attention] Restructure attention: move files (#31916)

2026-01-09 13:10:24 -08:00

voxtral_streaming.py

[Voxtral] Fix speech transcription api (#31388)

2026-01-08 18:34:19 +08:00

voxtral.py

[Voxtral] Fix speech transcription api (#31388)

2026-01-08 18:34:19 +08:00

whisper_utils.py

[6/N][Attention] Move utils to more appropriate locations (#32215)

2026-01-13 05:38:52 -08:00

whisper.py

[Refactor] Remove get_encoder_dummy_data (#32241)

2026-01-13 09:21:23 +00:00

zamba2.py

[Misc][Model][Refactor] Pass the prefix into Linear layers (#31669)

2026-01-05 20:03:18 +00:00