ZeroSpeech-TTS-without-T
ZeroSpeech-TTS-without-T copied to clipboard
argument ‘train_p‘ and 'train_tgat' are the same
Hi,
For training, the step2 '--train_p' and the step3 '--train_tgat' is not the same training process?
if args.train or args.train_p or args.train_tgat: trainer.add_duo_loader(source_loader, target_loader) trainer.train(model_path, args.flag, mode='patchGAN', target_guided=args.train_tgat) # Stage 2 training trainer.reset_keep()