search

OFA-Sys / ONE-PEACE

A general representation model across vision, audio, language modalities. Paper: ONE-PEACE: Exploring One General Representation Model Toward Unlimited Modalities
Apache License 2.0
981 stars 64 forks source link

Vision data in datasets #44

Closed kkontras closed 1 month ago

kkontras commented 11 months ago

Hi,

I would like to replicate some of the results and I started exploring the datasets. Its actually very nice that you have shared it but I was surprised to find only the audio data in some of them, for example AVQA, VGGSound and Kinetics400. Am I missing something there?

logicwong commented 11 months ago

In the initial version of our paper, we did not employ video data for experiments involving the VGGSound and AVQA datasets, so the initial open-source versions of these two datasets did not include video data. In addition, the Kinetics400 dataset should contain video data. You can check it carefully.