search

YicongHong / Recurrent-VLN-BERT

Code of the CVPR 2021 Oral paper: A Recurrent Vision-and-Language BERT for Navigation
Other
150 stars 29 forks source link

The vocab size #10

Closed MarSaKi closed 3 years ago

MarSaKi commented 3 years ago

Hi, yicong,

Thanks for your great work! I found the vocab size of R2R is 991，but the vocab size of Prevalent aug data is 1101. Additionaly, the Prevalent instructions is generated based on a speaker model trained on R2R dataset. Do you have any idea about this?

Thanks,

YicongHong commented 3 years ago

Hi There,

A very interesting point, but I have no idea. 😓

It is probably better to ask the authors of the Speaker-Follower or PREVALENT.

Cheers

MarSaKi commented 3 years ago

Thanks for your reply!

Cheers,