tuna2134/sbv2-v2
1
0
Fork 0
Code Issues Pull Requests 1 Actions Packages Projects Releases Wiki Activity

Apply Code Formatter Change

Browse Source
This commit is contained in:
Stardust-minus
2023-09-06 13:42:05 +00:00
committed by github-actions[bot]
parent d82ba3457a
commit eef212fd00
26 changed files with 2938 additions and 1453 deletions
Download Patch File Download Diff File Expand all files Collapse all files

63
text/chinese_bert.py
View File

@@ -4,20 +4,27 @@ from transformers import AutoTokenizer, AutoModelForMaskedLM
tokenizer = AutoTokenizer.from_pretrained("./bert/chinese-roberta-wwm-ext-large")
def get_bert_feature(text, word2ph, device=None):
if sys.platform == "darwin" and torch.backends.mps.is_available() and device == "cpu":
if (
sys.platform == "darwin"
and torch.backends.mps.is_available()
and device == "cpu"
):
device = "mps"
if not device:
device = "cuda"
model = AutoModelForMaskedLM.from_pretrained("./bert/chinese-roberta-wwm-ext-large").to(device)
model = AutoModelForMaskedLM.from_pretrained(
"./bert/chinese-roberta-wwm-ext-large"
).to(device)
with torch.no_grad():
inputs = tokenizer(text, return_tensors='pt')
inputs = tokenizer(text, return_tensors="pt")
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()
res = torch.cat(res["hidden_states"][-3:-2], -1)[0].cpu()
assert len(word2ph) == len(text)+2
assert len(word2ph) == len(text) + 2
word2phone = word2ph
phone_level_feature = []
for i in range(len(word2phone)):
@@ -26,14 +33,53 @@ def get_bert_feature(text, word2ph, device=None):
phone_level_feature = torch.cat(phone_level_feature, dim=0)
return phone_level_feature.T
if __name__ == '__main__':
if __name__ == "__main__":
import torch
word_level_feature = torch.rand(38, 1024) # 12个词,每个词1024维特征
word2phone = [1, 2, 1, 2, 2, 1, 2, 2, 1, 2, 2, 1, 2, 2, 2, 2, 2, 1, 1, 2, 2, 1, 2, 2, 2, 2, 1, 2, 2, 2, 2, 2, 1, 2, 2, 2, 2, 1]
word2phone = [
1,
2,
1,
2,
2,
1,
2,
2,
1,
2,
2,
1,
2,
2,
2,
2,
2,
1,
1,
2,
2,
1,
2,
2,
2,
2,
1,
2,
2,
2,
2,
2,
1,
2,
2,
2,
2,
1,
]
# 计算总帧数
total_frames = sum(word2phone)
@@ -49,4 +95,3 @@ if __name__ == '__main__':
phone_level_feature = torch.cat(phone_level_feature, dim=0)
print(phone_level_feature.shape) # torch.Size([36, 1024])