You signed in with another tab or window. Reload to refresh your session.You signed out in another tab or window. Reload to refresh your session.You switched accounts on another tab or window. Reload to refresh your session.Dismiss alert
Hi,
when I try to do the second step of the training export SPEAKER_PATH_PREFIX=tasks/R2R/speaker/snapshots/speaker_teacher_imagenet_mean_pooled_train_iter_20000
it fails because the files available after training are speaker_teacher_imagenet_mean_pooled_train_iter_20000_enc and speaker_teacher_imagenet_mean_pooled_train_iter_20000_dec, which one should we export to the path? Or did something go wrong during the speaker training?
The text was updated successfully, but these errors were encountered:
Hi @mees, did you manage to get it up and running in the end? The implementations in data_augmentation_from_speaker is really confusing to me as well. Especially with respect to the need for a follower during data generation
Hi,
when I try to do the second step of the training
export SPEAKER_PATH_PREFIX=tasks/R2R/speaker/snapshots/speaker_teacher_imagenet_mean_pooled_train_iter_20000
it fails because the files available after training are speaker_teacher_imagenet_mean_pooled_train_iter_20000_enc and speaker_teacher_imagenet_mean_pooled_train_iter_20000_dec, which one should we export to the path? Or did something go wrong during the speaker training?
The text was updated successfully, but these errors were encountered: