MeloTTS / melo /text /japanese_bert.py
mrfakename's picture
Init
4300fed
raw
history blame
1.63 kB
import torch
from transformers import AutoTokenizer, AutoModelForMaskedLM
import sys
# model = None
# model_id = 'cl-tohoku/bert-base-japanese-v3'
# tokenizer = AutoTokenizer.from_pretrained(model_id)
models = {}
tokenizers = {}
def get_bert_feature(text, word2ph, device=None, model_id='cl-tohoku/bert-base-japanese-v3'):
global model
global tokenizer
if (
sys.platform == "darwin"
and torch.backends.mps.is_available()
and device == "cpu"
):
device = "mps"
if not device:
device = "cuda"
if model_id not in models:
model = AutoModelForMaskedLM.from_pretrained(model_id).to(
device
)
models[model_id] = model
tokenizer = AutoTokenizer.from_pretrained(model_id)
tokenizers[model_id] = tokenizer
else:
model = models[model_id]
tokenizer = tokenizers[model_id]
with torch.no_grad():
inputs = tokenizer(text, return_tensors="pt")
tokenized = tokenizer.tokenize(text)
for i in inputs:
inputs[i] = inputs[i].to(device)
res = model(**inputs, output_hidden_states=True)
res = torch.cat(res["hidden_states"][-3:-2], -1)[0].cpu()
assert inputs["input_ids"].shape[-1] == len(word2ph), f"{inputs['input_ids'].shape[-1]}/{len(word2ph)}"
word2phone = word2ph
phone_level_feature = []
for i in range(len(word2phone)):
repeat_feature = res[i].repeat(word2phone[i], 1)
phone_level_feature.append(repeat_feature)
phone_level_feature = torch.cat(phone_level_feature, dim=0)
return phone_level_feature.T