def encoder_train(run_id, user_folder): parser = argparse.ArgumentParser( description= "Trains the speaker encoder. You must have run encoder_preprocess.py first.", formatter_class=argparse.ArgumentDefaultsHelpFormatter) parser.add_argument("--run_id", type=str, default=run_id, help= \ "Name for this model instance. If a model state from the same run ID was previously " "saved, the training will restart from there. Pass -f to overwrite saved states and " "restart from scratch.") parser.add_argument("--clean_data_root", type=Path, default='/home/ubuntu/VC_dataset/SV2TTS/encoder/', help= \ "Path to the output directory of encoder_preprocess.py. If you left the default " "output directory when preprocessing, it should be <datasets_root>/SV2TTS/encoder/.") parser.add_argument("-m", "--models_dir", type=Path, default="/home/ubuntu/Real-Time-Voice-Cloning/encoder/saved_models/", help=\ "Path to the output directory that will contain the saved model weights, as well as " "backups of those weights and plots generated during training.") parser.add_argument("-v", "--vis_every", type=int, default=20, help= \ "Number of steps between updates of the loss and the plots.") parser.add_argument("-u", "--umap_every", type=int, default=100, help= \ "Number of steps between updates of the umap projection. Set to 0 to never update the " "projections.") parser.add_argument("-s", "--save_every", type=int, default=500, help= \ "Number of steps between updates of the model on the disk. Set to 0 to never save the " "model.") parser.add_argument("-b", "--backup_every", type=int, default=10, help= \ "Number of steps between backups of the model. Set to 0 to never make backups of the " "model.") parser.add_argument("-f", "--force_restart", action="store_true", help= \ "Do not load any saved model.") parser.add_argument("--visdom_server", type=str, default="http://localhost") parser.add_argument("--no_visdom", action="store_true", help= \ "Disable visdom.") parser.add_argument( "--fine_tune_pretrained", type=Path, default= '/home/ubuntu/Real-Time-Voice-Cloning/encoder/saved_models/pretrained_cp.pt' ) parser.add_argument("--user_folder", type=Path, default=user_folder) args = parser.parse_args() # Process the arguments args.models_dir.mkdir(exist_ok=True) # Run the training print_args(args, parser) ckpt = train(**vars(args)) return ckpt
"output directory when preprocessing, it should be <datasets_root>/SV2TTS/encoder/.") parser.add_argument("-m", "--models_dir", type=Path, default="encoder/saved_models/", help=\ "Path to the output directory that will contain the saved model weights, as well as " "backups of those weights and plots generated during training.") parser.add_argument("-v", "--vis_every", type=int, default=10, help= \ "Number of steps between updates of the loss and the plots.") parser.add_argument("-u", "--umap_every", type=int, default=100, help= \ "Number of steps between updates of the umap projection. Set to 0 to never update the " "projections.") parser.add_argument("-s", "--save_every", type=int, default=500, help= \ "Number of steps between updates of the model on the disk. Set to 0 to never save the " "model.") parser.add_argument("-b", "--backup_every", type=int, default=7500, help= \ "Number of steps between backups of the model. Set to 0 to never make backups of the " "model.") parser.add_argument("-f", "--force_restart", action="store_true", help= \ "Do not load any saved model.") parser.add_argument("--visdom_server", type=str, default="http://10.4.208.57:6006") parser.add_argument("--no_visdom", action="store_true", help= \ "Disable visdom.") args = parser.parse_args() # Process the arguments args.models_dir.mkdir(exist_ok=True) # Run the training print_args(args, parser) train(**vars(args))