DL-Art-School/codes/scripts/audio/use_vocoder.py

import pathlib

import numpy
import torch
from scipy.io import wavfile
from tqdm import tqdm
import matplotlib.pyplot as plt
import librosa

from models.waveglow.waveglow import WaveGlow


class Vocoder:
    def __init__(self):
        self.model = WaveGlow(n_mel_channels=80, n_flows=12, n_group=8, n_early_size=2, n_early_every=4, WN_config={'n_layers': 8, 'n_channels': 256, 'kernel_size': 3})
        sd = torch.load('../experiments/waveglow_256channels_universal_v5.pth')
        self.model.load_state_dict(sd)
        self.model = self.model.to('cuda')
        self.model.eval()

    def transform_mel_to_audio(self, mel):
        if len(mel.shape) == 2:  # Assume it's missing the batch dimension and fix that.
            mel = mel.unsqueeze(0)
        with torch.no_grad():
            return self.model.infer(mel)


def plot_spectrogram(spec, title=None, ylabel="freq_bin", aspect="auto", xmax=None):
    fig, axs = plt.subplots(1, 1)
    axs.set_title(title or "Spectrogram (db)")
    axs.set_ylabel(ylabel)
    axs.set_xlabel("frame")
    im = axs.imshow(librosa.power_to_db(spec), origin="lower", aspect=aspect)
    if xmax:
        axs.set_xlim((0, xmax))
    fig.colorbar(im, ax=axs)
    plt.show(block=False)


if __name__ == '__main__':
    vocoder = Vocoder()
    m = torch.load('test_mels.pth')
    for i, b in enumerate(m):
        plot_spectrogram(b.cpu())
        wav = vocoder.transform_mel_to_audio(b)
        wavfile.write(f'{i}.wav', 22050, wav[0].cpu().numpy())