Questions about text feature dimenstion #6
Description
When I was loading the iemocap_features_4 pickle file. The video_text data is shown to be size 100. But you mentioned in your paper, the textual features' size is 768 extracted using sBERT. Could you please explain this?