Struct VocoderInferenceConfig 

pub struct VocoderInferenceConfig<'a> {
    pub checkpoint: &'a Path,
    pub mel_path: Option<&'a Path>,
    pub output: &'a Path,
    pub steps: usize,
    pub quality: Option<&'a str>,
    pub batch_input: Option<&'a PathBuf>,
    pub batch_output: Option<&'a PathBuf>,
    pub show_metrics: bool,
}

Configuration for vocoder inference operations

Consolidates parameters for converting mel spectrograms to audio waveforms using trained vocoder models. Supports both single-file and batch processing.

Examples

Single file inference:

use voirs_cli::commands::vocoder_inference::VocoderInferenceConfig;
use std::path::Path;

let config = VocoderInferenceConfig {
    checkpoint: Path::new("./checkpoints/vocoder.safetensors"),
    mel_path: Some(Path::new("./input.mel")),
    output: Path::new("./output.wav"),
    steps: 50,
    quality: Some("balanced"),
    batch_input: None,
    batch_output: None,
    show_metrics: false,
};

Batch processing:

use voirs_cli::commands::vocoder_inference::VocoderInferenceConfig;
use std::path::{Path, PathBuf};

let config = VocoderInferenceConfig {
    checkpoint: Path::new("./checkpoints/vocoder.safetensors"),
    mel_path: None,
    output: Path::new("./output_dir"),
    steps: 50,
    quality: Some("high"),
    batch_input: Some(&PathBuf::from("./mel_dir")),
    batch_output: Some(&PathBuf::from("./audio_dir")),
    show_metrics: true,
};

Fields

checkpoint: &'a Path

Path to trained vocoder checkpoint file

mel_path: Option<&'a Path>

Optional path to input mel spectrogram file (single file mode)

output: &'a Path

Output path for generated audio file or batch directory

steps: usize

Number of diffusion steps for generation (higher = better quality, slower)

quality: Option<&'a str>

Quality preset: “fast” (20 steps), “balanced” (50 steps), or “high” (100 steps)

batch_input: Option<&'a PathBuf>

Optional directory for batch input (batch mode)

batch_output: Option<&'a PathBuf>

Optional directory for batch output (batch mode)

show_metrics: bool

Whether to display performance metrics after inference

Trait Implementations

impl<'a> Debug for VocoderInferenceConfig<'a>

fn fmt(&self, f: &mut Formatter<'_>) -> Result

Formats the value using the given formatter.