To load the previous model when doing double finetuning,

[Peter-SungwooCho]

Hello! Thank you for providing great work.

I'm trying to fine-tune and saved checkpoint of a newly fine-tuned model again for an experiment!
For example, after saving a model fine-tuned with the LRS3 dataset, I want to fine-tune it again with the LRS2 dataset.

I set the checkpoint I created in the model.w2v_path section of the config and run it, but for some reason, the model seems to reload the existing large_vox_iter5.pt.

Since load my checkpoints is impossible, I am asking how can I change to apply it. (Maybe need to change hubert_asr.py file)

Best regards,
Peter