139 lines
4.2 KiB
Plaintext
139 lines
4.2 KiB
Plaintext
{
|
||
"cells": [
|
||
{
|
||
"cell_type": "markdown",
|
||
"metadata": {},
|
||
"source": [
|
||
"# Style-Bert-VITS2ライブラリの使用例\n",
|
||
"\n",
|
||
"`pip install style-bert-vits2`を使った、jupyter notebookでの使用例です。Google colab等でも動きます。"
|
||
]
|
||
},
|
||
{
|
||
"cell_type": "code",
|
||
"execution_count": null,
|
||
"metadata": {},
|
||
"outputs": [],
|
||
"source": [
|
||
"# PyTorch環境の構築（ない場合）\n",
|
||
"# 参照: https://pytorch.org/get-started/locally/\n",
|
||
"\n",
|
||
"!pip install torch torchvision torchaudio --index-url https://download.pytorch.org/whl/cu118"
|
||
]
|
||
},
|
||
{
|
||
"cell_type": "code",
|
||
"execution_count": null,
|
||
"metadata": {
|
||
"id": "LLrngKcQEAyP"
|
||
},
|
||
"outputs": [],
|
||
"source": [
|
||
"# style-bert-vits2のインストール\n",
|
||
"\n",
|
||
"!pip install style-bert-vits2"
|
||
]
|
||
},
|
||
{
|
||
"cell_type": "code",
|
||
"execution_count": null,
|
||
"metadata": {
|
||
"id": "9xRtfUg5EZkx"
|
||
},
|
||
"outputs": [],
|
||
"source": [
|
||
"# BERTモデルをロード（ローカルに手動でダウンロードする必要はありません）\n",
|
||
"\n",
|
||
"from style_bert_vits2.nlp import bert_models\n",
|
||
"from style_bert_vits2.constants import Languages\n",
|
||
"\n",
|
||
"\n",
|
||
"bert_models.load_model(Languages.JP, \"ku-nlp/deberta-v2-large-japanese-char-wwm\")\n",
|
||
"bert_models.load_tokenizer(Languages.JP, \"ku-nlp/deberta-v2-large-japanese-char-wwm\")\n",
|
||
"# bert_models.load_model(Languages.EN, \"microsoft/deberta-v3-large\")\n",
|
||
"# bert_models.load_tokenizer(Languages.EN, \"microsoft/deberta-v3-large\")\n",
|
||
"# bert_models.load_model(Languages.ZH, \"hfl/chinese-roberta-wwm-ext-large\")\n",
|
||
"# bert_models.load_tokenizer(Languages.ZH, \"hfl/chinese-roberta-wwm-ext-large\")"
|
||
]
|
||
},
|
||
{
|
||
"cell_type": "code",
|
||
"execution_count": null,
|
||
"metadata": {
|
||
"id": "q2V9d3HyFAr_"
|
||
},
|
||
"outputs": [],
|
||
"source": [
|
||
"# Hugging Faceから試しにデフォルトモデルをダウンロードしてみて、それを音声合成に使ってみる\n",
|
||
"# model_assetsディレクトリにダウンロードされます\n",
|
||
"\n",
|
||
"from pathlib import Path\n",
|
||
"from huggingface_hub import hf_hub_download\n",
|
||
"\n",
|
||
"\n",
|
||
"model_file = \"jvnv-F1-jp/jvnv-F1-jp_e160_s14000.safetensors\"\n",
|
||
"config_file = \"jvnv-F1-jp/config.json\"\n",
|
||
"style_file = \"jvnv-F1-jp/style_vectors.npy\"\n",
|
||
"\n",
|
||
"for file in [model_file, config_file, style_file]:\n",
|
||
" print(file)\n",
|
||
" hf_hub_download(\n",
|
||
" \"litagin/style_bert_vits2_jvnv\",\n",
|
||
" file,\n",
|
||
" local_dir=\"model_assets\"\n",
|
||
" )"
|
||
]
|
||
},
|
||
{
|
||
"cell_type": "code",
|
||
"execution_count": null,
|
||
"metadata": {
|
||
"id": "hJa31MEUFhe4"
|
||
},
|
||
"outputs": [],
|
||
"source": [
|
||
"# 上でダウンロードしたモデルファイルを指定して音声合成のテスト\n",
|
||
"\n",
|
||
"from style_bert_vits2.tts_model import TTSModel\n",
|
||
"\n",
|
||
"assets_root = Path(\"model_assets\")\n",
|
||
"\n",
|
||
"model = TTSModel(\n",
|
||
" model_path=assets_root / model_file,\n",
|
||
" config_path=assets_root / config_file,\n",
|
||
" style_vec_path=assets_root / style_file,\n",
|
||
" device=\"cpu\"\n",
|
||
")"
|
||
]
|
||
},
|
||
{
|
||
"cell_type": "code",
|
||
"execution_count": null,
|
||
"metadata": {
|
||
"id": "Gal0tqrtGXZx"
|
||
},
|
||
"outputs": [],
|
||
"source": [
|
||
"from IPython.display import Audio, display\n",
|
||
"\n",
|
||
"sr, audio = model.infer(text=\"こんにちは\")\n",
|
||
"display(Audio(audio, rate=sr))"
|
||
]
|
||
}
|
||
],
|
||
"metadata": {
|
||
"colab": {
|
||
"provenance": []
|
||
},
|
||
"kernelspec": {
|
||
"display_name": "Python 3",
|
||
"name": "python3"
|
||
},
|
||
"language_info": {
|
||
"name": "python"
|
||
}
|
||
},
|
||
"nbformat": 4,
|
||
"nbformat_minor": 0
|
||
}
|