{
  "nbformat": 4,
  "nbformat_minor": 0,
  "metadata": {
    "colab": {
      "name": "tortoise-tts.ipynb",
      "provenance": [],
      "collapsed_sections": []
    },
    "kernelspec": {
      "name": "python3",
      "display_name": "Python 3"
    },
    "language_info": {
      "name": "python"
    },
    "accelerator": "GPU"
  },
  "cells": [
    {
      "cell_type": "markdown",
      "source": [
        "Welcome to Tortoise! 🐢🐢🐢🐢\n",
        "\n",
        "Before you begin, I **strongly** recommend you turn on a GPU runtime.\n",
        "\n",
        "There's a reason this is called \"Tortoise\" - this model takes up to a minute to perform inference for a single sentence on a GPU. Expect waits on the order of hours on a CPU."
      ],
      "metadata": {
        "id": "_pIZ3ZXNp7cf"
      }
    },
    {
      "cell_type": "code",
      "execution_count": null,
      "metadata": {
        "id": "JrK20I32grP6"
      },
      "outputs": [],
      "source": [
        "!git clone https://github.com/neonbjb/tortoise-tts.git\n",
        "%cd tortoise-tts\n",
        "!pip install -r requirements.txt"
      ]
    },
    {
      "cell_type": "code",
      "source": [
        "# Imports used through the rest of the notebook.\n",
        "import torch\n",
        "import torchaudio\n",
        "import torch.nn as nn\n",
        "import torch.nn.functional as F\n",
        "\n",
        "from api import TextToSpeech\n",
        "from utils.audio import load_audio, get_voices\n",
        "\n",
        "# This will download all the models used by Tortoise from the HF hub.\n",
        "tts = TextToSpeech()"
      ],
      "metadata": {
        "id": "Gen09NM4hONQ"
      },
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "code",
      "source": [
        "# List all the voices available. These are just some random clips I've gathered\n",
        "# from the internet as well as a few voices from the training dataset.\n",
        "# Feel free to add your own clips to the voices/ folder.\n",
        "%ls voices"
      ],
      "metadata": {
        "id": "SSleVnRAiEE2"
      },
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "code",
      "source": [
        "# This is the text that will be spoken.\n",
        "text = \"Joining two modalities results in a surprising increase in generalization! What would happen if we combined them all?\"\n",
        "\n",
        "# Here's something for the poetically inclined.. (set text=)\n",
        "\"\"\"\n",
        "Then took the other, as just as fair,\n",
        "And having perhaps the better claim,\n",
        "Because it was grassy and wanted wear;\n",
        "Though as for that the passing there\n",
        "Had worn them really about the same,\"\"\"\n",
        "\n",
        "# Pick one of the voices from above\n",
        "voice = 'dotrice'\n",
        "# Pick a \"preset mode\" to determine quality. Options: {\"ultra_fast\", \"fast\" (default), \"standard\", \"high_quality\"}. See docs in api.py\n",
        "preset = \"fast\""
      ],
      "metadata": {
        "id": "bt_aoxONjfL2"
      },
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "code",
      "source": [
        "# Fetch the voice references and forward execute!\n",
        "voices = get_voices()\n",
        "cond_paths = voices[voice]\n",
        "conds = []\n",
        "for cond_path in cond_paths:\n",
        "    c = load_audio(cond_path, 22050)\n",
        "    conds.append(c)\n",
        "\n",
        "gen = tts.tts_with_preset(text, conds, preset)\n",
        "torchaudio.save('generated.wav', gen.squeeze(0).cpu(), 24000)"
      ],
      "metadata": {
        "id": "KEXOKjIvn6NW"
      },
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "code",
      "source": [
        "# You can add as many conditioning voices as you want together. Combining\n",
        "# clips from multiple voices takes the mean of the latent space for all\n",
        "# voices. This creates a novel voice that is a combination of the two inputs.\n",
        "#\n",
        "# Lets see what it would sound like if Picard and Kirk had a kid with a penchant for philosophy:\n",
        "conds = []\n",
        "for v in ['pat', 'william']:\n",
        "  cond_paths = voices[v]\n",
        "  for cond_path in cond_paths:\n",
        "      c = load_audio(cond_path, 22050)\n",
        "      conds.append(c)\n",
        "\n",
        "gen = tts.tts_with_preset(\"They used to say that if man was meant to fly, he’d have wings. But he did fly. He discovered he had to.\", conds, preset)\n",
        "torchaudio.save('captain_kirkard.wav', gen.squeeze(0).cpu(), 24000)"
      ],
      "metadata": {
        "id": "fYTk8KUezUr5"
      },
      "execution_count": null,
      "outputs": []
    }
  ]
}