class DecoderModel(ModelBase) using padding token for prediction
richardbaihe opened this issue · 1 comments
richardbaihe commented
Hi There,
In the t0/model.py's DecoderModel class, I noticed that there are padding tokens between the input_ids and labels. It seems padding tokens are used when predicting the labels. Is this an issue?
VictorSanh commented
fixed with #32 thanks to @thomasw21