232 lines
8.5 KiB
Python
232 lines
8.5 KiB
Python
import json
|
|
import os
|
|
import shutil
|
|
import subprocess
|
|
import yaml
|
|
|
|
import gradio as gr
|
|
import sys
|
|
|
|
python = sys.executable
|
|
|
|
|
|
def subprocess_wrapper(cmd):
|
|
return subprocess.run(
|
|
cmd,
|
|
stdout=sys.stdout,
|
|
stderr=subprocess.PIPE,
|
|
text=True,
|
|
)
|
|
|
|
|
|
def get_path(model_name):
|
|
assert model_name != "", "モデル名は空にできません"
|
|
dataset_path = os.path.join("Data", model_name)
|
|
lbl_path = os.path.join(dataset_path, "esd.list")
|
|
train_path = os.path.join(dataset_path, "train.list")
|
|
val_path = os.path.join(dataset_path, "val.list")
|
|
config_path = os.path.join(dataset_path, "config.json")
|
|
return dataset_path, lbl_path, train_path, val_path, config_path
|
|
|
|
|
|
def initialize(model_name, batch_size, epochs):
|
|
dataset_path, _, train_path, val_path, config_path = get_path(model_name)
|
|
if os.path.isfile(config_path):
|
|
config = json.load(open(config_path, "r", encoding="utf-8"))
|
|
else:
|
|
# Use default config
|
|
config = json.load(open("configs/config.json", "r", encoding="utf-8"))
|
|
config["model_name"] = model_name
|
|
config["data"]["training_files"] = train_path
|
|
config["data"]["validation_files"] = val_path
|
|
config["train"]["batch_size"] = batch_size
|
|
config["train"]["epochs"] = epochs
|
|
|
|
model_path = os.path.join(dataset_path, "models")
|
|
try:
|
|
shutil.copytree(src="pretrained", dst=model_path)
|
|
except FileExistsError:
|
|
return f"Error: モデルフォルダ {model_path} が既に存在します。問題なければ削除してください。"
|
|
except FileNotFoundError:
|
|
return "Error: pretrainedフォルダが見つかりません。"
|
|
|
|
with open(config_path, "w", encoding="utf-8") as f:
|
|
json.dump(config, f, indent=2)
|
|
if not os.path.exists("config.yml"):
|
|
shutil.copy(src="default_config.yml", dst="config.yml")
|
|
# yml_data = safe_load(open("config.yml", "r", encoding="utf-8"))
|
|
with open("config.yml", "r", encoding="utf-8") as f:
|
|
yml_data = yaml.safe_load(f)
|
|
yml_data["model_name"] = model_name
|
|
yml_data["dataset_path"] = dataset_path
|
|
with open("config.yml", "w", encoding="utf-8") as f:
|
|
yaml.dump(yml_data, f, allow_unicode=True)
|
|
return "Step 1: 初期設定が完了しました"
|
|
|
|
|
|
def resample(model_name):
|
|
dataset_path, _, _, _, _ = get_path(model_name)
|
|
in_dir = os.path.join(dataset_path, "raw")
|
|
out_dir = os.path.join(dataset_path, "wavs")
|
|
result = subprocess_wrapper(
|
|
[
|
|
python,
|
|
"resample.py",
|
|
"--in_dir",
|
|
in_dir,
|
|
"--out_dir",
|
|
out_dir,
|
|
"--sr",
|
|
"44100",
|
|
]
|
|
)
|
|
if result.stderr:
|
|
return f"{result.stderr}"
|
|
return "Step 2: 音声ファイルの前処理が完了しました"
|
|
|
|
|
|
def preprocess_text(model_name):
|
|
dataset_path, lbl_path, train_path, val_path, config_path = get_path(model_name)
|
|
lines = open(lbl_path, "r", encoding="utf-8").readlines()
|
|
with open(lbl_path, "w", encoding="utf-8") as f:
|
|
for line in lines:
|
|
path, spk, language, text = line.strip().split("|")
|
|
path = os.path.join(dataset_path, "wavs", os.path.basename(path)).replace(
|
|
"\\", "/"
|
|
)
|
|
f.writelines(f"{path}|{spk}|{language}|{text}\n")
|
|
result = subprocess_wrapper(
|
|
[
|
|
python,
|
|
"preprocess_text.py",
|
|
"--config-path",
|
|
config_path,
|
|
"--transcription-path",
|
|
lbl_path,
|
|
"--train-path",
|
|
train_path,
|
|
"--val-path",
|
|
val_path,
|
|
]
|
|
)
|
|
|
|
if result.stderr:
|
|
return f"{result.stderr}"
|
|
return "Step 3: 書き起こしファイルの前処理が完了しました"
|
|
|
|
|
|
def bert_gen(model_name):
|
|
_, _, _, _, config_path = get_path(model_name)
|
|
result = subprocess_wrapper([python, "bert_gen.py", "--config", config_path])
|
|
if result.stderr:
|
|
return f"{result.stderr}"
|
|
return "Step 4: BERT特徴ファイルの生成が完了しました"
|
|
|
|
|
|
def style_gen(model_name):
|
|
dataset_path, _, _, _, config_path = get_path(model_name)
|
|
result = subprocess_wrapper(
|
|
[python, "style_gen.py", "--config", config_path, "--model", dataset_path]
|
|
)
|
|
if result.stderr:
|
|
return f"{result.stderr}"
|
|
return "Step 5: スタイル特徴ファイルの生成が完了しました"
|
|
|
|
|
|
def train(model_name):
|
|
dataset_path, _, _, _, config_path = get_path(model_name)
|
|
result = subprocess_wrapper(
|
|
[python, "train_ms.py", "--config", config_path, "--model", dataset_path]
|
|
)
|
|
if result.stderr:
|
|
return f"{result.stderr}"
|
|
return "Final Step: 学習が完了しました!"
|
|
|
|
|
|
prepare_md = """
|
|
次のようにデータを置いてください。
|
|
```
|
|
├── Data
|
|
│ ├── {モデルの名前}
|
|
│ │ ├── esd.list
|
|
│ │ ├── raw
|
|
│ │ │ ├── ****.wav
|
|
│ │ │ ├── ****.wav
|
|
│ │ │ ├── ...
|
|
```
|
|
|
|
wavファイル名やモデルの名前は空白を含まない半角で、wavファイルの拡張子は小文字`.wav`である必要があります。
|
|
`raw` フォルダにはすべてのwavファイルを入れ、`esd.list` ファイルには、以下のフォーマットで各wavファイルの情報を記述してください。
|
|
```
|
|
****.wav|{話者名}|{言語ID、ZHかJPかEN}|{書き起こしテキスト}
|
|
```
|
|
|
|
例：
|
|
```
|
|
wav_number1.wav|hanako|JP|こんにちは、聞こえて、いますか？
|
|
wav_next.wav|taro|JP|はい、聞こえています……。
|
|
english_teacher.wav|Mary|EN|How are you? I'm fine, thank you, and you?
|
|
...
|
|
```
|
|
もちろん日本語話者の単一話者データセットでも構いません。
|
|
"""
|
|
|
|
if __name__ == "__main__":
|
|
with gr.Blocks(theme="NoCrypt/miku") as app:
|
|
gr.Markdown("# Style Bert-VITS2 データ前処理")
|
|
gr.Markdown("途中から学習を再開する場合は、モデル名を入力してFinal Stepだけ実行すればよいです。")
|
|
with gr.Accordion(label="データの前準備", open=False):
|
|
gr.Markdown(prepare_md)
|
|
model_name = gr.Textbox(
|
|
label="モデル名",
|
|
)
|
|
info = gr.Textbox(label="状況")
|
|
with gr.Row():
|
|
with gr.Column():
|
|
gr.Markdown(value="### Step 1: 設定ファイルの生成")
|
|
with gr.Row():
|
|
batch_size = gr.Slider(
|
|
label="バッチサイズ",
|
|
info="VRAM 12GBで4くらい",
|
|
value=4,
|
|
minimum=1,
|
|
maximum=64,
|
|
step=1,
|
|
)
|
|
epochs = gr.Slider(
|
|
label="エポック数",
|
|
info="100もあれば十分そう",
|
|
value=100,
|
|
minimum=1,
|
|
maximum=1000,
|
|
step=1,
|
|
)
|
|
generate_config_btn = gr.Button(value="実行", variant="primary")
|
|
with gr.Column():
|
|
gr.Markdown(value="### Step 2: 音声ファイルの前処理")
|
|
resample_btn = gr.Button(value="実行", variant="primary")
|
|
with gr.Column():
|
|
gr.Markdown(value="### Step 3: 書き起こしファイルの前処理")
|
|
preprocess_text_btn = gr.Button(value="実行", variant="primary")
|
|
with gr.Column():
|
|
gr.Markdown(value="### Step 4: BERT特徴ファイルの生成")
|
|
bert_gen_btn = gr.Button(value="実行", variant="primary")
|
|
with gr.Column():
|
|
gr.Markdown(value="### Step 5: スタイル特徴ファイルの生成")
|
|
style_gen_btn = gr.Button(value="実行", variant="primary")
|
|
with gr.Row():
|
|
with gr.Column():
|
|
gr.Markdown(value="### Final Step: 学習")
|
|
train_btn = gr.Button(value="学習", variant="primary")
|
|
|
|
generate_config_btn.click(
|
|
initialize, inputs=[model_name, batch_size, epochs], outputs=[info]
|
|
)
|
|
resample_btn.click(resample, inputs=[model_name], outputs=[info])
|
|
preprocess_text_btn.click(preprocess_text, inputs=[model_name], outputs=[info])
|
|
bert_gen_btn.click(bert_gen, inputs=[model_name], outputs=[info])
|
|
style_gen_btn.click(style_gen, inputs=[model_name], outputs=[info])
|
|
train_btn.click(train, inputs=[model_name], outputs=[info])
|
|
|
|
app.launch(share=False, server_port=7860, inbrowser=True)
|