This issue aims at giving the possibility to evaluate few-shot learning performances. To do that, a parameter could represent the percentage of a whole data given for training. With that, it would be possible to evaluate the quantity of data necessary for a defined model, and its performance.
This issue aims at giving the possibility to evaluate few-shot learning performances. To do that, a parameter could represent the percentage of a whole data given for training. With that, it would be possible to evaluate the quantity of data necessary for a defined model, and its performance.