How can I reimplement results on CMU-MOSEI dataset?
sailist opened this issue · comments
Thanks again for your great work, I noticed that in this code you only provide features and scripts for iemocap setting, but not CMU-MOSEI. Could you kindly provide necessary data and code to reproduce results on CMU-MOSEI setting?
@iabhinavjoshi Would you mind providing any explaination?
Did you use preprocess file for the train test split at your end?
I mean, where did the author split the pickle file features into train and test?
Also are you able to extract this pickle file.
I have created a dataset with audio video and test scripts. Can I use this model ?