Open UchihaBarry opened 11 months ago
Hi there, thanks for your interest in our work. We trained the identity and expression encoder simultaneously, first at the ICT dataset and then ICT + Multiface. The training part requires a complex data-preprocessing pipeline. We will release the code once we find a good way to document the processing precedure.
Were you able document the processing procedure?
i wonder the details during the training the Identity Encoder and Expression Encoder. And when will the training code be released as mentioned in the readme document.