Positional embedding contains the False padding index `1`

by Phando - opened Sep 19, 2023

Sep 19, 2023

Suppose this could result from the vocab converting, since the original tokenizer set the padding index to 1 while this repo's GPT2Tokenizer has a padding index of 50258

Upload images, audio, and videos by dragging in the text input, pasting, or clicking here.

Tap or paste here to upload images

· Sign up or log in to comment