codec/scripts/download_gigaspeech.py

from random import sample

import soundfile as sf
from datasets import load_dataset

# dataset = load_dataset("keithito/lj_speech", split="train")
#dataset = load_dataset("parler-tts/mls_eng", split="train")
dataset = load_dataset("speechcolab/gigaspeech", "xl", split="train", token=True)
Is = sample(list(range(len(dataset))), k=100000)
print(dataset)
for i, I in enumerate(Is):
    audio = dataset[I]["audio"]
    wav, sr = audio["array"], audio["sampling_rate"]
    sf.write(f"gigaspeech/{I}.wav", wav, sr)