kanshan/TensorRT-LLMs
1
0
Fork 0
mirror of https://github.com/NVIDIA/TensorRT-LLM.git synced 2026-02-18 08:45:05 +08:00
Code Issues Actions 1 Packages Projects Releases Wiki Activity

[https://nvbugs/5854419][fix] Fix Qwen3-VL-Dense/MoE accuracy drop (#11134)

Browse Source 
Signed-off-by: yechank <161688079+yechank-nvidia@users.noreply.github.com>
Signed-off-by: Wangshanshan <30051912+dominicshanshan@users.noreply.github.com>
This commit is contained in:
Yechan Kim 2026-02-03 18:59:44 +09:00 committed by Yanchao Lu
parent 435ea36977
commit ebd859cf61
2 changed files with 2 additions and 0 deletions
Show Stats Download Patch File Download Diff File Expand all files Collapse all files

1
tensorrt_llm/_torch/models/modeling_qwen3vl.py
View File

@ -1138,6 +1138,7 @@ class Qwen3VLModelBase(PreTrainedModel):
"video": "<|vision_start|><|video_pad|><|vision_end|>",
},
placeholder_placement=MultimodalPlaceholderPlacement.BEFORE_TEXT,
placeholders_separator="",
),
)
class Qwen3VLModel(Qwen3VLModelBase):

1
tensorrt_llm/_torch/models/modeling_qwen3vl_moe.py
View File

@ -41,6 +41,7 @@ from .modeling_utils import ModelConfig, register_auto_model, register_vision_en
"video": "<|vision_start|><|video_pad|><|vision_end|>",
},
placeholder_placement=MultimodalPlaceholderPlacement.BEFORE_TEXT,
placeholders_separator="",
),
)
class Qwen3MoeVLModel(Qwen3VLModelBase):