search

ttengwang / PDVC

End-to-End Dense Video Captioning with Parallel Decoding (ICCV 2021)
MIT License
200 stars 23 forks source link

Always predicting <end token> for captions for validation loop #25

Closed arnavsshah closed 2 years ago

arnavsshah commented 2 years ago

The model trains fine, but during the validation loop, the model only predicts instead of other words? Did you face this issue too?

ttengwang commented 2 years ago

I didn't face this issue before. Btw, it is strange, as I didn't use a word like "" in this repo.