I am using blog this with custom data to fine-tune wav2vec.
in the training part, the writer passes the tokenizer too. My question is does the tokenizer also get fine-tuned? and if it does how do we load the fine-tuned tokenizer from the drive .
here the writer is saving the output to drive but in the evaluation part of the post, he doesn’t explain how to load the model or tokenizer from the drive and all he says is that they have been uploaded here and you can call them/
I was able to figure out how to load the fine-tuned model :
but while trying the same thing for the tokenizer it gives me the error shown above. Can anyone help?