code-kunkun

/

LamRA-Rank-Qwen2.5VL-7b

like 0

Update image_processor_type to use Qwen2VLImageProcessor in preprocessor_config.json

This aligns with Qwen2.5-VL processor and avoids the Unrecognized image processor error when loading the processor with AutoProcessor.from_pretrained
https://huggingface.co/Qwen/Qwen2.5-VL-7B-Instruct/blob/main/preprocessor_config.json

See related issue
https://github.com/huggingface/transformers/issues/36193