modify

app.py

@@ -5,15 +5,7 @@ import numpy as np
 import yaml
 from inference import MasteringStyleTransfer
 from utils import download_youtube_audio
-from config import config
-
-# Initialize MasteringStyleTransfer
-args = type('Args', (), vars(config))()
-# load network configurations
-with open(args.path_to_config, 'r') as f:
-    configs = yaml.full_load(f)
-args.cfg_converter = configs['TCN']['param_mapping']
-args.cfg_enc = configs['Effects_Encoder']['default']
+from config import args
 
 mastering_transfer = MasteringStyleTransfer(args)
 

config.py

@@ -1,28 +1,23 @@
-import os
+import argparse
+import yaml
+
+parser = argparse.ArgumentParser(description="Mastering Style Transfer")
+parser.add_argument("--model_path", type=str, default=f"models/white_box_converter.pt", help="Path to mastering converter model")
+parser.add_argument("--encoder_path", type=str, default=f"models/mastering_effects_encoder.pt", help="Path to effects encoder model")
+parser.add_argument("--perform_ito", action="store_true", help="Whether to perform ITO")
+parser.add_argument("--optimizer", type=str, default="RAdam", help="Optimizer for ITO")
+parser.add_argument("--learning_rate", type=float, default=0.001, help="Learning rate for ITO")
+parser.add_argument("--num_steps", type=int, default=100, help="Number of optimization steps for ITO")
+parser.add_argument("--af_weights", nargs='+', type=float, default=[0.1, 0.001, 1.0, 1.0, 0.1], help="Weights for AudioFeatureLoss")
+parser.add_argument("--sample_rate", type=int, default=44100, help="Sample rate for AudioFeatureLoss")
+parser.add_argument("--path_to_config", type=str, default='networks/configs.yaml', help="Path to network architecture configuration file")
+
+args = parser.parse_args()
+
+# load network configurations
+with open(args.path_to_config, 'r') as f:
+    configs = yaml.full_load(f)
+args.cfg_converter = configs['TCN']['param_mapping']
+args.cfg_enc = configs['Effects_Encoder']['default']
 
-class Config:
-    model_path = "models/mastering_converter.pt"
-    encoder_path = "models/effects_encoder.pt"
-    sample_rate = 44100
-    path_to_config = "networks/configs.yaml"
-    
-    # Add more configurations as needed
-    inference_only = True
-    evaluate_only = False
-    reload_enc = True
-    reload_converter = True
-    manual_reload_enc = False
-    manual_reload_converter = False
-    manual_reload_name_converter = "dasp_tcn_tuneenc_daspman_loudnessnorm"
-    manual_reload_name_encoder = "dasp_tcn_tuneenc_daspman_loudnessnorm"
-    manual_reload_enc_path = "models/effects_encoder.pt"
-    reload_epoch_converter = 1000
-    reload_epoch_enc = 1000
-    
-    # Add other configurations from the main.py file
-    random_seed = 111
-    train_ito = True
-    max_iter_ito = 101
-    ito_type = "blackbox"
 
-config = Config()