Nicolò LucchesiandNickLucche 69ff99fdcd [Core] Optimizing cross-attention QKVParallelLinear computation (#12325) ...

Signed-off-by: NickLucche <nlucches@redhat.com>
Signed-off-by: NickLucche <nick@nlucches-4xa100.c.openshift-330514.internal>
Co-authored-by: NickLucche <nick@nlucches-4xa100.c.openshift-330514.internal>

2025-03-06 09:37:26 +00:00

..

__init__.py

[V1]SupportsV0Only protocol for model definitions (#13959)

2025-02-27 20:02:15 -05:00

adapters.py

Remove unused kwargs from model definitions (#13555)

2025-02-24 17:13:52 -08:00

arctic.py

Clean up unused padding_idx variables across many model definitions (#13240)

2025-03-04 21:27:00 +00:00

aria.py

[V1] EP/TP MoE + DP Attention (#13931)

2025-03-04 21:27:26 -08:00

baichuan.py

Clean up unused padding_idx variables across many model definitions (#13240)

2025-03-04 21:27:00 +00:00

bamba.py

[V1]SupportsV0Only protocol for model definitions (#13959)

2025-02-27 20:02:15 -05:00

bart.py

[Core] Optimizing cross-attention QKVParallelLinear computation (#12325)

2025-03-06 09:37:26 +00:00

bert.py

[V1]SupportsV0Only protocol for model definitions (#13959)

2025-02-27 20:02:15 -05:00

blip2.py

[V1][Molmo] Fix get_multimodal_embeddings() in molmo.py (#14161)

2025-03-04 15:43:59 +00:00

blip.py

[Misc] Add SPDX-License-Identifier headers to python source files (#12628)

2025-02-02 11:58:18 -08:00

bloom.py

Remove unused kwargs from model definitions (#13555)

2025-02-24 17:13:52 -08:00

chameleon.py

Clean up unused padding_idx variables across many model definitions (#13240)

2025-03-04 21:27:00 +00:00

chatglm.py

Remove unused kwargs from model definitions (#13555)

2025-02-24 17:13:52 -08:00

clip.py

[Bugfix] Fix Positive Feature Layers in Llava Models (#13514)

2025-02-19 08:50:07 +00:00

commandr.py

Remove unused kwargs from model definitions (#13555)

2025-02-24 17:13:52 -08:00

dbrx.py

[V1] EP/TP MoE + DP Attention (#13931)

2025-03-04 21:27:26 -08:00

decilm.py

[Misc] Add SPDX-License-Identifier headers to python source files (#12628)

2025-02-02 11:58:18 -08:00

deepseek_mtp.py

[Model][Speculative Decoding] Expand DeepSeek MTP code to support k > n_predict (#13626)

2025-02-27 15:28:08 -08:00

deepseek_v2.py

Clean up unused padding_idx variables across many model definitions (#13240)

2025-03-04 21:27:00 +00:00

deepseek_vl2.py

[V1][Molmo] Fix get_multimodal_embeddings() in molmo.py (#14161)

2025-03-04 15:43:59 +00:00

deepseek.py

Clean up unused padding_idx variables across many model definitions (#13240)

2025-03-04 21:27:00 +00:00

eagle.py

Remove unused kwargs from model definitions (#13555)

2025-02-24 17:13:52 -08:00

exaone.py

Clean up unused padding_idx variables across many model definitions (#13240)

2025-03-04 21:27:00 +00:00

fairseq2_llama.py

[Misc] Add SPDX-License-Identifier headers to python source files (#12628)

2025-02-02 11:58:18 -08:00

falcon.py

Remove unused kwargs from model definitions (#13555)

2025-02-24 17:13:52 -08:00

florence2.py

Clean up unused padding_idx variables across many model definitions (#13240)

2025-03-04 21:27:00 +00:00

fuyu.py

Clean up unused padding_idx variables across many model definitions (#13240)

2025-03-04 21:27:00 +00:00

gemma2.py

Remove unused kwargs from model definitions (#13555)

2025-02-24 17:13:52 -08:00

gemma.py

Remove unused kwargs from model definitions (#13555)

2025-02-24 17:13:52 -08:00

glm4v.py

[V1][Molmo] Fix get_multimodal_embeddings() in molmo.py (#14161)

2025-03-04 15:43:59 +00:00

glm.py

[Misc] Add SPDX-License-Identifier headers to python source files (#12628)

2025-02-02 11:58:18 -08:00

gpt2.py

Remove unused kwargs from model definitions (#13555)

2025-02-24 17:13:52 -08:00

gpt_bigcode.py

Remove unused kwargs from model definitions (#13555)

2025-02-24 17:13:52 -08:00

gpt_j.py

Remove unused kwargs from model definitions (#13555)

2025-02-24 17:13:52 -08:00

gpt_neox.py

Remove unused kwargs from model definitions (#13555)

2025-02-24 17:13:52 -08:00

granite.py

Clean up unused padding_idx variables across many model definitions (#13240)

2025-03-04 21:27:00 +00:00

granitemoe.py

Clean up unused padding_idx variables across many model definitions (#13240)

2025-03-04 21:27:00 +00:00

granitemoeshared.py

[Model] Add support for GraniteMoeShared models (#13313)

2025-03-04 08:04:52 +08:00

gritlm.py

[V1]SupportsV0Only protocol for model definitions (#13959)

2025-02-27 20:02:15 -05:00

grok1.py

[Model] Support Grok1 (#13795)

2025-02-26 01:07:12 +00:00

h2ovl.py

[VLM] Generalized prompt updates for multi-modal processor (#13964)

2025-02-27 17:44:25 +00:00

idefics2_vision_model.py

[Misc] Add SPDX-License-Identifier headers to python source files (#12628)

2025-02-02 11:58:18 -08:00

idefics3.py

Clean up unused padding_idx variables across many model definitions (#13240)

2025-03-04 21:27:00 +00:00

interfaces_base.py

Remove unused kwargs from model definitions (#13555)

2025-02-24 17:13:52 -08:00

interfaces.py

[V1][Molmo] Fix get_multimodal_embeddings() in molmo.py (#14161)

2025-03-04 15:43:59 +00:00

intern_vit.py

[Misc] Add SPDX-License-Identifier headers to python source files (#12628)

2025-02-02 11:58:18 -08:00

internlm2_ve.py

Remove unused kwargs from model definitions (#13555)

2025-02-24 17:13:52 -08:00

internlm2.py

Clean up unused padding_idx variables across many model definitions (#13240)

2025-03-04 21:27:00 +00:00

internvl.py

[V1][Molmo] Fix get_multimodal_embeddings() in molmo.py (#14161)

2025-03-04 15:43:59 +00:00

jais.py

Remove unused kwargs from model definitions (#13555)

2025-02-24 17:13:52 -08:00

jamba.py

[V1] EP/TP MoE + DP Attention (#13931)

2025-03-04 21:27:26 -08:00

llama.py

Clean up unused padding_idx variables across many model definitions (#13240)

2025-03-04 21:27:00 +00:00

llava_next_video.py

[V1][Molmo] Fix get_multimodal_embeddings() in molmo.py (#14161)

2025-03-04 15:43:59 +00:00

llava_next.py

[V1][Molmo] Fix get_multimodal_embeddings() in molmo.py (#14161)

2025-03-04 15:43:59 +00:00

llava_onevision.py

[VLM] Generalized prompt updates for multi-modal processor (#13964)

2025-02-27 17:44:25 +00:00

llava.py

[V1][VLM][Pixtral-HF] Support Pixtral-HF on V1 (#14275)

2025-03-06 08:58:41 +00:00

mamba2.py

[V1]SupportsV0Only protocol for model definitions (#13959)

2025-02-27 20:02:15 -05:00

mamba_cache.py

[Model] Support Mamba2 (Codestral Mamba) (#9292)

2025-02-17 20:17:50 +08:00

mamba.py

Clean up unused padding_idx variables across many model definitions (#13240)

2025-03-04 21:27:00 +00:00

medusa.py

[Misc] Add SPDX-License-Identifier headers to python source files (#12628)

2025-02-02 11:58:18 -08:00

minicpm3.py

Remove unused kwargs from model definitions (#13555)

2025-02-24 17:13:52 -08:00

minicpm.py

Clean up unused padding_idx variables across many model definitions (#13240)

2025-03-04 21:27:00 +00:00

minicpmo.py

[VLM] Generalized prompt updates for multi-modal processor (#13964)

2025-02-27 17:44:25 +00:00

minicpmv.py

[Misc][V1] Avoid using envs.VLLM_USE_V1 in mm processing (#14256)

2025-03-05 07:37:16 +00:00

mixtral_quant.py

Clean up unused padding_idx variables across many model definitions (#13240)

2025-03-04 21:27:00 +00:00

mixtral.py

[V1] EP/TP MoE + DP Attention (#13931)

2025-03-04 21:27:26 -08:00

mllama.py

[Core] Optimizing cross-attention QKVParallelLinear computation (#12325)

2025-03-06 09:37:26 +00:00

mlp_speculator.py

[Doc] Replace ibm-fms with ibm-ai-platform (#12709)

2025-02-04 07:05:04 +00:00

module_mapping.py

[Misc] Add SPDX-License-Identifier headers to python source files (#12628)

2025-02-02 11:58:18 -08:00

molmo.py

[V1][VLM][Pixtral-HF] Support Pixtral-HF on V1 (#14275)

2025-03-06 08:58:41 +00:00

mpt.py

Remove unused kwargs from model definitions (#13555)

2025-02-24 17:13:52 -08:00

nemotron.py

Clean up unused padding_idx variables across many model definitions (#13240)

2025-03-04 21:27:00 +00:00

nvlm_d.py

[VLM] Generalized prompt updates for multi-modal processor (#13964)

2025-02-27 17:44:25 +00:00

olmo2.py

Remove unused kwargs from model definitions (#13555)

2025-02-24 17:13:52 -08:00

olmo.py

Remove unused kwargs from model definitions (#13555)

2025-02-24 17:13:52 -08:00

olmoe.py

[V1] EP/TP MoE + DP Attention (#13931)

2025-03-04 21:27:26 -08:00

opt.py

Clean up unused padding_idx variables across many model definitions (#13240)

2025-03-04 21:27:00 +00:00

orion.py

Clean up unused padding_idx variables across many model definitions (#13240)

2025-03-04 21:27:00 +00:00

paligemma.py

[Model] Update Paligemma multimodal processing with PromptUpdate (#14015)

2025-03-06 08:31:38 +00:00

persimmon.py

Remove unused kwargs from model definitions (#13555)

2025-02-24 17:13:52 -08:00

phi3_small.py

Remove unused kwargs from model definitions (#13555)

2025-02-24 17:13:52 -08:00

phi3.py

[Misc] Add SPDX-License-Identifier headers to python source files (#12628)

2025-02-02 11:58:18 -08:00

phi3v.py

[V1][Molmo] Fix get_multimodal_embeddings() in molmo.py (#14161)

2025-03-04 15:43:59 +00:00

phi4mm_audio.py

[Model] New model support for Phi-4-multimodal-instruct (#14119)

2025-03-04 20:57:01 -08:00

phi4mm_utils.py

[Model] New model support for Phi-4-multimodal-instruct (#14119)

2025-03-04 20:57:01 -08:00

phi4mm.py

[Model] New model support for Phi-4-multimodal-instruct (#14119)

2025-03-04 20:57:01 -08:00

phi.py

Remove unused kwargs from model definitions (#13555)

2025-02-24 17:13:52 -08:00

phimoe.py

[V1] EP/TP MoE + DP Attention (#13931)

2025-03-04 21:27:26 -08:00

pixtral.py

[V1][VLM][Pixtral-HF] Support Pixtral-HF on V1 (#14275)

2025-03-06 08:58:41 +00:00

prithvi_geospatial_mae.py

[Misc][V1] Avoid using envs.VLLM_USE_V1 in mm processing (#14256)

2025-03-05 07:37:16 +00:00

qwen2_5_vl.py

[Bugfix] Initialize attention bias on the same device as Query/Key/Value for QwenVL Series (#14031)

2025-02-28 07:36:08 -08:00

qwen2_audio.py

[V1][Molmo] Fix get_multimodal_embeddings() in molmo.py (#14161)

2025-03-04 15:43:59 +00:00

qwen2_moe.py

[V1] EP/TP MoE + DP Attention (#13931)

2025-03-04 21:27:26 -08:00

qwen2_rm.py

[V1]SupportsV0Only protocol for model definitions (#13959)

2025-02-27 20:02:15 -05:00

qwen2_vl.py

[Bugfix] Initialize attention bias on the same device as Query/Key/Value for QwenVL Series (#14031)

2025-02-28 07:36:08 -08:00

qwen2.py

Clean up unused padding_idx variables across many model definitions (#13240)

2025-03-04 21:27:00 +00:00

qwen_vl.py

[V1][Molmo] Fix get_multimodal_embeddings() in molmo.py (#14161)

2025-03-04 15:43:59 +00:00

qwen.py

Remove unused kwargs from model definitions (#13555)

2025-02-24 17:13:52 -08:00

registry.py

[Model] New model support for Phi-4-multimodal-instruct (#14119)

2025-03-04 20:57:01 -08:00

roberta.py

[V1]SupportsV0Only protocol for model definitions (#13959)

2025-02-27 20:02:15 -05:00

siglip.py

[Bugfix] Fix Positive Feature Layers in Llava Models (#13514)

2025-02-19 08:50:07 +00:00

solar.py

Clean up unused padding_idx variables across many model definitions (#13240)

2025-03-04 21:27:00 +00:00

stablelm.py

Remove unused kwargs from model definitions (#13555)

2025-02-24 17:13:52 -08:00

starcoder2.py

Clean up unused padding_idx variables across many model definitions (#13240)

2025-03-04 21:27:00 +00:00

telechat2.py

[Misc] Add SPDX-License-Identifier headers to python source files (#12628)

2025-02-02 11:58:18 -08:00

transformers.py

[LoRA] Remove linear hack outside transformers backend (#14177)

2025-03-05 15:06:28 +00:00

ultravox.py

[V1][Molmo] Fix get_multimodal_embeddings() in molmo.py (#14161)

2025-03-04 15:43:59 +00:00

utils.py

[Core] Optimizing cross-attention QKVParallelLinear computation (#12325)

2025-03-06 09:37:26 +00:00

vision_siglip_navit.py

[Model] New model support for Phi-4-multimodal-instruct (#14119)

2025-03-04 20:57:01 -08:00

vision.py

[Bugfix] Fix Positive Feature Layers in Llava Models (#13514)

2025-02-19 08:50:07 +00:00

whisper.py

Clean up unused padding_idx variables across many model definitions (#13240)

2025-03-04 21:27:00 +00:00