Open praveena2j opened 2 years ago
When the mode is one, I see that you are passing the mean of the modality (text or sound), may i know where do you define the tensor of the mean of the modality ?
Also could you please let me know where did you use reconstruct.py snippet in your code for training ?
Exactly same problem on my side. Have you solved it? @praveena2j Also, can the author somehow answer the question plz @mengmenm? If the code to get the sound mean is not released, it is not possible to reproduce your results.
Thanks
Could not find the script to train the feature reconstruction model