File size: 2,088 Bytes
9bd9742
 
 
 
 
 
f3e00ea
9bd9742
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
f3e00ea
9bd9742
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
import torch
from transformers import AutoTokenizer, AutoModelForMaskedLM
import sys
from text.japanese import text2sep_kata

tokenizer = AutoTokenizer.from_pretrained("ku-nlp/deberta-v2-base-japanese")
# tokenizer = AutoTokenizer.from_pretrained("ku-nlp/deberta-v2-base-japanese-with-auto-jumanpp")

models = dict()


def get_bert_feature(text, word2ph, device=None):
    sep_text, _ = text2sep_kata(text)
    sep_tokens = [tokenizer.tokenize(t) for t in sep_text]
    sep_ids = [tokenizer.convert_tokens_to_ids(t) for t in sep_tokens]
    sep_ids = [2] + [item for sublist in sep_ids for item in sublist] + [3]
    return get_bert_feature_with_token(sep_ids, word2ph, device)


def get_bert_feature_with_token(tokens, word2ph, device=None):
    if (
        sys.platform == "darwin"
        and torch.backends.mps.is_available()
        and device == "cpu"
    ):
        device = "mps"
    if not device:
        device = "cuda"
    if device not in models.keys():
        models[device] = AutoModelForMaskedLM.from_pretrained(
            "ku-nlp/deberta-v2-base-japanese"
            # "ku-nlp/deberta-v2-base-japanese-with-auto-jumanpp"
        ).to(device)
    with torch.no_grad():
        inputs = torch.tensor(tokens).to(device).unsqueeze(0)
        token_type_ids = torch.zeros_like(inputs).to(device)
        attention_mask = torch.ones_like(inputs).to(device)
        inputs = {
            "input_ids": inputs,
            "token_type_ids": token_type_ids,
            "attention_mask": attention_mask,
        }

        # for i in inputs:
        #     inputs[i] = inputs[i].to(device)
        res = models[device](**inputs, output_hidden_states=True)
        res = torch.cat(res["hidden_states"][-3:-2], -1)[0].cpu()
    assert 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