I am trying to adapt https://github.com/patil-suraj/exploring-T5/blob/master/t5_fine_tuning.ipynb to my own dataset, a text-to-sql task. I am able to train the model but can’t evaluate it.
I am trying to evaluate the dataset like in the notebook: to a given NL-query I would like to compare the reference SQL with the generate SQL. Currently, I am getting this bos_token_id error:
But in my Dataset Class I have defined the input_ids, or did I not?:
Can anybody help, please? Thank you in advance!