Files
sbv2-v2/bert_gen.py
2023-09-05 01:59:36 -04:00

53 lines
1.6 KiB
Python

import torch
from multiprocessing import Pool
import commons
import utils
from tqdm import tqdm
from text import cleaned_text_to_sequence, get_bert
config_path = "configs/config.json"
hps = utils.get_hparams_from_file(config_path)
total_gpus = torch.cuda.device_count()
def process_line(line):
wav_path, _, language_str, text, phones, tone, word2ph = line.strip().split("|")
phone = phones.split(" ")
tone = [int(i) for i in tone.split(" ")]
word2ph = [int(i) for i in word2ph.split(" ")]
word2ph = [i for i in word2ph]
phone, tone, language = cleaned_text_to_sequence(phone, tone, language_str)
if hps.data.add_blank:
phone = commons.intersperse(phone, 0)
tone = commons.intersperse(tone, 0)
language = commons.intersperse(language, 0)
for i in range(len(word2ph)):
word2ph[i] = word2ph[i] * 2
word2ph[0] += 1
bert_path = wav_path.replace(".wav", ".bert.pt")
try:
bert = torch.load(bert_path)
assert bert.shape[-1] == len(phone)
except Exception:
bert = get_bert(text, word2ph, language_str)
assert bert.shape[-1] == len(phone)
torch.save(bert, bert_path)
if __name__ == "__main__":
lines = []
with open(hps.data.training_files, encoding="utf-8") as f:
lines.extend(f.readlines())
with open(hps.data.validation_files, encoding="utf-8") as f:
lines.extend(f.readlines())
# A100 40GB suitable config, if coom, please decrease the processes number.
with Pool(processes=6) as pool:
for _ in tqdm(pool.imap_unordered(process_line, lines), total=len(lines)):
pass