About the token Selection

Open loseevaya opened this issue 2 years ago • 1 comments

Nice work! When selecting tokens from the encoder output, the output dimension of the class_embedding is 91, which includes the category of "no object". Will the tokens selected in this way have an impact on the results?

Mar 05 '23 04:03 loseevaya

We use focal loss, where no "no object" token exists. Or you can view it as multiple binary classifications.

Mar 20 '23 02:03 SlongLiu