storyteller-ml/voice_conversion/so-vits-svc-paul/notebooks/so-vits-svc-fork-4.0.ipynb

{
 "cells": [
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## Before training\n",
    "\n",
    "This program saves the last 3 generations of models to Google Drive. Since 1 generation of models is >1GB, you should have at least 3GB of free space in Google Drive. If you do not have such free space, it is recommended to create another Google Account.\n",
    "\n",
    "Training requires >10GB VRAM. (T4 should be enough) Inference does not require such a lot of VRAM."
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## Installation"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": [
    "#@title Check GPU\n",
    "!nvidia-smi"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": [
    "#@title Mount Google Drive\n",
    "from google.colab import drive\n",
    "drive.mount('/content/drive')"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": [
    "#@title Install dependencies\n",
    "!python -m pip install -U pip wheel\n",
    "%pip install -U ipython \n",
    "\n",
    "#@markdown Branch (for development)\n",
    "BRANCH = \"none\" #@param {\"type\": \"string\"}\n",
    "if BRANCH == \"none\":\n",
    "    %pip install -U so-vits-svc-fork\n",
    "else:\n",
    "    %pip install -U git+https://github.com/34j/so-vits-svc-fork.git@{BRANCH}"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## Training"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": [
    "#@title Make dataset directory\n",
    "%cd ./content"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": [
    "!mkdir -p \"dataset_raw\""
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": [
    "#!rm -r \"dataset_raw\"\n",
    "#!rm -r \"dataset/44k\""
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": [
    "#@title Copy your dataset\n",
    "DATASET_NAME = \"kiritan\" #@param {type: \"string\"}\n",
    "!cp -R /content/drive/MyDrive/so-vits-svc-fork/dataset/{DATASET_NAME}/ -t \"dataset_raw/\""
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": [
    "#@title Download dataset (Tsukuyomi-chan JVS)\n",
    "#@markdown Make sure you agree to the license when using this dataset.\n",
    "# !wget https://tyc.rei-yumesaki.net/files/sozai-tyc-corpus1.zip\n",
    "# !unzip sozai-tyc-corpus1.zip\n",
    "# !mv \"/content/つくよみちゃんコーパス Vol.1 声優統計コーパス（JVSコーパス準拠）/おまけ：WAV（+12dB増幅＆高音域削減）/WAV（+12dB増幅＆高音域削減）\" \"dataset_raw/tsukuyomi\""
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": [
    "#@title Automatic preprocessing\n",
    "!svc pre-resample"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": [
    "!svc pre-config"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": [
    "#@title Copy configs file\n",
    "!cp configs/44k/config.json drive/MyDrive/so-vits-svc-fork"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": [
    "!svc pre-hubert"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": [
    "#@title Train\n",
    "%load_ext tensorboard\n",
    "%tensorboard --logdir drive/MyDrive/so-vits-svc-fork/logs/44k\n",
    "!svc train --model-path drive/MyDrive/so-vits-svc-fork/logs/44k"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## Inference"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": [
    "#@title Get the author's voice as a source\n",
    "import random\n",
    "NAME = str(random.randint(1, 100))\n",
    "!wget -N \"https://github.com/34j/34j/raw/main/jvs-parallel100/{NAME}.wav\"\n",
    "from IPython.display import Audio, display\n",
    "display(Audio(f\"{NAME}.wav\"))"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": [
    "#@title Use trained model\n",
    "!svc infer {NAME}.wav -m drive/MyDrive/so-vits-svc-fork/logs/44k/ -c drive/MyDrive/so-vits-svc-fork/logs/44k/config.json\n",
    "display(Audio(f\"{NAME}.out.wav\", autoplay=True))"
   ]
  },
  {
   "attachments": {},
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### Pretrained models"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": [
    "#@title https://huggingface.co/TachibanaKimika/so-vits-svc-4.0-models/tree/main\n",
    "!wget -N \"https://huggingface.co/TachibanaKimika/so-vits-svc-4.0-models/resolve/main/riri/G_riri_220.pth\"\n",
    "!wget -N \"https://huggingface.co/TachibanaKimika/so-vits-svc-4.0-models/resolve/main/riri/config.json\""
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": [
    "!svc infer {NAME}.wav -c config.json -m G_riri_220.pth\n",
    "display(Audio(f\"{NAME}.out.wav\", autoplay=True))"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": [
    "#@title https://huggingface.co/therealvul/so-vits-svc-4.0/tree/main\n",
    "!wget -N \"https://huggingface.co/therealvul/so-vits-svc-4.0/resolve/main/Pinkie%20(speaking%20sep)/G_166400.pth\"\n",
    "!wget -N \"https://huggingface.co/therealvul/so-vits-svc-4.0/resolve/main/Pinkie%20(speaking%20sep)/config.json\""
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": [
    "!svc infer {NAME}.wav --speaker \"Pinkie {neutral}\" -c config.json -m G_166400.pth\n",
    "display(Audio(f\"{NAME}.out.wav\", autoplay=True))"
   ]
  }
 ],
 "metadata": {
  "accelerator": "GPU",
  "colab": {
   "provenance": [
    {
     "file_id": "https://github.com/34j/so-vits-svc-fork/blob/feat%2Fmain-feat/so-vits-svc-fork-4.0.ipynb",
     "timestamp": 1678970434570
    }
   ]
  },
  "gpuClass": "standard",
  "kernelspec": {
   "display_name": "Python 3",
   "name": "python3"
  },
  "language_info": {
   "name": "python"
  }
 },
 "nbformat": 4,
 "nbformat_minor": 0
}