A high-throughput and memory-efficient inference and serving engine for LLMs
Mirrored from https://github.com/tduncan-lum/Nvidia-Optimization-Stack---VLLM.git
Synced Nov 13, 05:45 UTC
[Bugfix] fix dots.ocr pp support (#28705)
Signed-off-by: zjy0516 <riverclouds.zhu@qq.com>
Jiangyun Zhu committed
Nov 14, 2025 at 17:01 UTC
c36bcfe6b37967ab52763f2ddb9400ff4fe3885b
1 file changed
+4
vllm/model_executor/models/dots_ocr.py
+4
@@ -780,6 +780,10 @@ class DotsOCRForCausalLM(nn.Module, SupportsMultiModal, SupportsPP, SupportsLoRA
780
architectures=["Qwen2ForCausalLM"],
781
)
782
783
+ self.make_empty_intermediate_tensors = (
784
+ self.language_model.make_empty_intermediate_tensors
785
+ )
786
+
787
def _parse_and_validate_image_input(
788
self, **kwargs: object
789
) -> DotsOCRImageInputs | None: