process_flow.md

graph TD
    A[Start] --> B[Run voice_converter_runner.py]
    B --> C{Select Option}
    
    C -->|Option 1: Train New Model| D[Training Process]
    C -->|Option 2: Resume Training| E[Resume Training]
    C -->|Option 3: Convert Voice| F[Voice Conversion]
    C -->|Option 4: Debug| G[Debug Mode]
    
    D --> D1[Initialize Dataset]
    D1 --> D2[Create DataLoader]
    D2 --> D3[Initialize Models]
    D3 --> D4[Setup Optimizers]
    D4 --> D5[Training Loop]
    D5 --> D6[Save Checkpoint]
    
    E --> E1[Load Latest Checkpoint]
    E1 --> E2[Restore States]
    E2 --> E3[Continue Training]
    
    F --> F1[Load Trained Model]
    F1 --> F2[Select Target Voice]
    F2 --> F3[Select TTS Audio]
    F3 --> F4[Process Audio]
    F4 --> F5[Post-processing]
    F5 --> F6[Save Output]
    
    G --> G1[Create Debug Directory]
    G1 --> G2[Save Original Files]
    G2 --> G3[Extract Features]
    G3 --> G4[Process Test Chunk]
    G4 --> G5[Save Debug Info]
    
    subgraph Training Process
    D1
    D2
    D3
    D4
    D5
    D6
    end
    
    subgraph Voice Conversion
    F1
    F2
    F3
    F4
    F5
    F6
    end
    
    subgraph Debug Mode
    G1
    G2
    G3
    G4
    G5
    end

Loading

Process Flow Description

Main Options

1. Training Process (Option 1)

   • Initialize dataset and data loader
   • Set up neural network models
   • Train with batch processing
   • Save checkpoints periodically

2. Resume Training (Option 2)

   • Load latest checkpoint
   • Restore model and optimizer states
   • Continue training from saved state

3. Voice Conversion (Option 3)

   • Load trained model
   • Process target voice and TTS audio
   • Apply voice conversion
   • Save converted output

4. Debug Mode (Option 4)

   • Create debug environment
   • Save intermediate outputs
   • Analyze conversion process
   • Generate detailed logs

Key Components

• Audio Processing: Handles all audio file operations
• Neural Networks: Generator and discriminator models
• Data Management: Dataset handling and normalization
• Debug Tools: Analysis and logging utilities