How to load the quantized model pth file in OpenCLIP?

[xiaohoua]

I used the pytorch_quantization tool to quantify the open_clip model, and saved the quantized model weight as checkpoint.pth
when I used

model, _, preprocess = open_clip.create_model_and_transforms(
            args.model_name, 
            pretrained='my_path_topth/checkpoint.pth'
        )

Attempt to load the quantized model weights，An error occurs ：

Traceback (most recent call last):
  File "predict_accurancy.py", line 169, in <module>
    model, _, preprocess = open_clip.create_model_and_transforms(
  File "/usr/local/lib/python3.8/dist-packages/open_clip/factory.py", line 384, in create_model_and_transforms
    model = create_model(
  File "/usr/local/lib/python3.8/dist-packages/open_clip/factory.py", line 290, in create_model
    load_checkpoint(model, checkpoint_path)
  File "/usr/local/lib/python3.8/dist-packages/open_clip/factory.py", line 160, in load_checkpoint
    resize_text_pos_embed(state_dict, model)
  File "/usr/local/lib/python3.8/dist-packages/open_clip/model.py", line 573, in resize_text_pos_embed
    assert old_width == width, 'text pos_embed width changed!'
AssertionError: text pos_embed width changed!

How to load the quantized model pth file in OpenCLIP?Can you give me some advices?

[LXY-19]

Hello, have you solved it?