Auto download missing model for bert_gen.py (fishaudio#146)

* auto download missing model

* support openi

* [pre-commit.ci] auto fixes from pre-commit.com hooks

for more information, see https://pre-commit.ci

* fix wrong delete

* [pre-commit.ci] auto fixes from pre-commit.com hooks

for more information, see https://pre-commit.ci

* pass pre-commit

* [pre-commit.ci] auto fixes from pre-commit.com hooks

for more information, see https://pre-commit.ci

* fix repeat login

---------

Co-authored-by: pre-commit-ci[bot] <66853113+pre-commit-ci[bot]@users.noreply.github.com>

bert/bert_models.json

@@ -0,0 +1,14 @@
+{
+    "deberta-v2-large-japanese": {
+        "repo_id": "ku-nlp/deberta-v2-large-japanese",
+        "files": ["spm.model", "pytorch_model.bin"]
+    },
+    "chinese-roberta-wwm-ext-large": {
+        "repo_id": "hfl/chinese-roberta-wwm-ext-large",
+        "files": ["pytorch_model.bin"]
+    },
+    "deberta-v3-large": {
+        "repo_id": "microsoft/deberta-v3-large",
+        "files": ["spm.model", "pytorch_model.bin"]
+    }
+}

bert_gen.py

@@ -3,7 +3,7 @@
 import commons
 import utils
 from tqdm import tqdm
-from text import cleaned_text_to_sequence, get_bert
+from text import check_bert_models, cleaned_text_to_sequence, get_bert
 import argparse
 import torch.multiprocessing as mp
 from config import config
@@ -57,6 +57,7 @@ def process_line(line):
     args, _ = parser.parse_known_args()
     config_path = args.config
     hps = utils.get_hparams_from_file(config_path)
+    check_bert_models()
     lines = []
     with open(hps.data.training_files, encoding="utf-8") as f:
         lines.extend(f.readlines())

config.py

@@ -195,7 +195,10 @@ def __init__(self, config_path: str):
         with open(file=config_path, mode="r", encoding="utf-8") as file:
             yaml_config: Dict[str, any] = yaml.safe_load(file.read())
             dataset_path: str = yaml_config["dataset_path"]
+            openi_token: str = yaml_config["openi_token"]
             self.dataset_path: str = dataset_path
+            self.mirror: str = yaml_config["mirror"]
+            self.openi_token: str = openi_token
             self.resample_config: Resample_config = Resample_config.from_dict(
                 dataset_path, yaml_config["resample"]
             )

default_config.yml

@@ -5,7 +5,8 @@
 # 每个数据集与其对应的模型存放至统一路径下，后续所有的路径配置均为相对于datasetPath的路径
 # 不填或者填空则路径为相对于项目根目录的路径
 dataset_path: "Data/你的数据集"
-
+mirror: "openi" # 模型镜像源
+openi_token: "1145141919810"  # openi token
 
 # resample 音频重采样配置
 # 注意， “:” 后需要加空格

text/__init__.py

@@ -26,3 +26,23 @@ def get_bert(norm_text, word2ph, language, device):
     lang_bert_func_map = {"ZH": zh_bert, "EN": en_bert, "JP": jp_bert}
     bert = lang_bert_func_map[language](norm_text, word2ph, device)
     return bert
+
+
+def check_bert_models():
+    import json
+    from pathlib import Path
+
+    from config import config
+    from .bert_utils import _check_bert
+
+    if config.mirror.lower() == "openi":
+        import openi
+
+        kwargs = {"token": config.openi_token} if config.openi_token else {}
+        openi.login(**kwargs)
+
+    with open("./bert/bert_models.json", "r") as fp:
+        models = json.load(fp)
+        for k, v in models.items():
+            local_path = Path("./bert").joinpath(k)
+            _check_bert(v["repo_id"], v["files"], local_path)

text/bert_utils.py

@@ -0,0 +1,23 @@
+from pathlib import Path
+
+from huggingface_hub import hf_hub_download
+
+from config import config
+
+
+MIRROR: str = config.mirror
+
+
+def _check_bert(repo_id, files, local_path):
+    for file in files:
+        if not Path(local_path).joinpath(file).exists():
+            if MIRROR.lower() == "openi":
+                import openi
+
+                openi.model.download_model(
+                    "Stardust_minus/Bert-VITS2", repo_id.split("/")[-1], "./bert"
+                )
+            else:
+                hf_hub_download(
+                    repo_id, file, local_dir=local_path, local_dir_use_symlinks=False
+                )

text/chinese_bert.py

@@ -1,9 +1,13 @@
-import torch
 import sys
-from transformers import AutoTokenizer, AutoModelForMaskedLM
+
+import torch
+from transformers import AutoModelForMaskedLM, AutoTokenizer
+
 from config import config
 
-tokenizer = AutoTokenizer.from_pretrained("./bert/chinese-roberta-wwm-ext-large")
+LOCAL_PATH = "./bert/chinese-roberta-wwm-ext-large"
+
+tokenizer = AutoTokenizer.from_pretrained(LOCAL_PATH)
 
 models = dict()
 
@@ -18,9 +22,7 @@ def get_bert_feature(text, word2ph, device=config.bert_gen_config.device):
     if not device:
         device = "cuda"
     if device not in models.keys():
-        models[device] = AutoModelForMaskedLM.from_pretrained(
-            "./bert/chinese-roberta-wwm-ext-large"
-        ).to(device)
+        models[device] = AutoModelForMaskedLM.from_pretrained(LOCAL_PATH).to(device)
     with torch.no_grad():
         inputs = tokenizer(text, return_tensors="pt")
         for i in inputs:
@@ -41,8 +43,6 @@ def get_bert_feature(text, word2ph, device=config.bert_gen_config.device):
 
 
 if __name__ == "__main__":
-    import torch
-
     word_level_feature = torch.rand(38, 1024)  # 12个词,每个词1024维特征
     word2phone = [
         1,

text/english_bert_mock.py

@@ -1,9 +1,14 @@
+import sys
+
 import torch
 from transformers import DebertaV2Model, DebertaV2Tokenizer
+
 from config import config
-import sys
 
-tokenizer = DebertaV2Tokenizer.from_pretrained("./bert/deberta-v3-large")
+
+LOCAL_PATH = "./bert/deberta-v3-large"
+
+tokenizer = DebertaV2Tokenizer.from_pretrained(LOCAL_PATH)
 
 models = dict()
 
@@ -18,9 +23,7 @@ def get_bert_feature(text, word2ph, device=config.bert_gen_config.device):
     if not device:
         device = "cuda"
     if device not in models.keys():
-        models[device] = DebertaV2Model.from_pretrained("./bert/deberta-v3-large").to(
-            device
-        )
+        models[device] = DebertaV2Model.from_pretrained(LOCAL_PATH).to(device)
     with torch.no_grad():
         inputs = tokenizer(text, return_tensors="pt")
         for i in inputs:

text/japanese_bert.py

@@ -1,10 +1,14 @@
-import torch
-from transformers import AutoTokenizer, AutoModelForMaskedLM
 import sys
-from text.japanese import text2sep_kata
+
+import torch
+from transformers import AutoModelForMaskedLM, AutoTokenizer
+
 from config import config
+from text.japanese import text2sep_kata
+
+LOCAL_PATH = "./bert/deberta-v2-large-japanese"
 
-tokenizer = AutoTokenizer.from_pretrained("./bert/deberta-v2-large-japanese")
+tokenizer = AutoTokenizer.from_pretrained(LOCAL_PATH)
 
 models = dict()
 
@@ -27,9 +31,7 @@ def get_bert_feature_with_token(tokens, word2ph, device=config.bert_gen_config.d
     if not device:
         device = "cuda"
     if device not in models.keys():
-        models[device] = AutoModelForMaskedLM.from_pretrained(
-            "./bert/deberta-v2-large-japanese"
-        ).to(device)
+        models[device] = AutoModelForMaskedLM.from_pretrained(LOCAL_PATH).to(device)
     with torch.no_grad():
         inputs = torch.tensor(tokens).to(device).unsqueeze(0)
         token_type_ids = torch.zeros_like(inputs).to(device)