AutoProcessor loading error

[JJJYmmm]

System Info

Related Issues and PR: #34307 #36184

• transformers version: 4.49.0.dev0
• Platform: Linux-5.15.0-131-generic-x86_64-with-glibc2.35
• Python version: 3.10.16
• Huggingface_hub version: 0.27.1
• Safetensors version: 0.5.2
• Accelerate version: 1.0.1
• Accelerate config: not found
• PyTorch version (GPU?): 2.6.0+cu126 (True)
• Tensorflow version (GPU?): not installed (NA)
• Flax version (CPU?/GPU?/TPU?): not installed (NA)
• Jax version: not installed
• JaxLib version: not installed
• Using distributed or parallel set-up in script?:
• Using GPU in script?:
• GPU type: NVIDIA H100 80GB HBM3

Who can help?

@Rocketknight1

Information

• The official example scripts
• My own modified scripts

Tasks

• An officially supported task in the examples folder (such as GLUE/SQuAD, ...)
• My own task or dataset (give details below)

Reproduction

Here are the reprodution steps

1. choose a mllm like qwen2.5vl, and download it's config file
2. derive its images processor, processor and model
3. modify the config file and try to use AutoProcessor to load_from_pretrain
4. and the error occurs like Initializing via AutoImageProcessor before AutoProcessor is imported causes AttributeError #34307

from transformers import Qwen2_5_VLProcessor, Qwen2_5_VLImageProcessor, Qwen2_5_VLForConditionalGeneration, Qwen2_5_VLConfig

class NewProcessor(Qwen2_5_VLProcessor):

    image_processor_class = "NewImageProcessor"

    def __init__(self, *args, **kwargs):
        super().__init__(*args, **kwargs)

class NewImageProcessor(Qwen2_5_VLImageProcessor):

    def __init__(self, *args, **kwargs):
        super().__init__(*args, **kwargs)

class NewConfig(Qwen2_5_VLConfig):
    model_type = "new_model"

class NewModel(Qwen2_5_VLForConditionalGeneration):

    config_class = NewConfig

    def __init__(self, *args, **kwargs):
        super().__init__(*args, **kwargs)



from transformers import AutoModel, AutoImageProcessor, AutoProcessor

AutoImageProcessor.register(NewModel.config_class, NewImageProcessor)
AutoProcessor.register(NewModel.config_class, NewProcessor)
AutoModel.register(NewModel.config_class, NewModel)


if __name__ == "__main__":
    processor = NewProcessor.from_pretrained("path/to/NewModel_config/")

modified config

config.json:

"architectures": [
    "NewModel"
],
"model_type": "new_model",

preprocessor_config.json:

"image_processor_type": "NewImageProcessor",
"processor_class": "NewProcessor"

I also check the pr #36184, it didn't work, because the func _get_class_from_class_name use mapping but the key is string rather than Config class

Expected behavior

None

[JJJYmmm]

fix by #36184