From 82105b1cd7bcfa4bdf3805b4168588700712f7da Mon Sep 17 00:00:00 2001 From: =?UTF-8?q?Stardust=C2=B7=E5=87=8F?= Date: Tue, 5 Sep 2023 00:54:00 +0800 Subject: [PATCH] Create japanese_bert.py --- text/japanese_bert.py | 26 ++++++++++++++++++++++++++ 1 file changed, 26 insertions(+) create mode 100644 text/japanese_bert.py diff --git a/text/japanese_bert.py b/text/japanese_bert.py new file mode 100644 index 0000000..304d03a --- /dev/null +++ b/text/japanese_bert.py @@ -0,0 +1,26 @@ +import torch +from transformers import AutoTokenizer, AutoModelForMaskedLM + +device = torch.device("cuda" if torch.cuda.is_available() else "cpu") + +tokenizer = AutoTokenizer.from_pretrained("./bert/bert-base-japanese-v3") +model = AutoModelForMaskedLM.from_pretrained("./bert/bert-base-japanese-v3").to(device) + +def get_bert_feature(text, word2ph): + with torch.no_grad(): + 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() + 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