microsoft/VQ-Diffusion_Not-User-Friendly.ipynb

microsoft-vq-diffusion_not-user-friendly.ipynb

{
  "nbformat": 4,
  "nbformat_minor": 0,
  "metadata": {
    "colab": {
      "name": "microsoft/VQ-Diffusion_Not-User-Friendly.ipynb",
      "private_outputs": true,
      "provenance": [],
      "collapsed_sections": [],
      "machine_shape": "hm",
      "authorship_tag": "ABX9TyMNZdVDz5BQuRmkfA+MOj5m",
      "include_colab_link": true
    },
    "kernelspec": {
      "name": "python3",
      "display_name": "Python 3"
    },
    "language_info": {
      "name": "python"
    },
    "accelerator": "GPU"
  },
  "cells": [
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "view-in-github",
        "colab_type": "text"
      },
      "source": [
        "<a href=\"https://colab.research.google.com/gist/Norod/f66399619447530e878de7f7088de0c2/microsoft-vq-diffusion_not-user-friendly.ipynb\" target=\"_parent\"><img src=\"https://colab.research.google.com/assets/colab-badge.svg\" alt=\"Open In Colab\"/></a>"
      ]
    },
    {
      "cell_type": "code",
      "source": [
        "%cd /content/\n",
        "!git clone https://github.com/cientgu/VQ-Diffusion.git\n",
        "%cd ./VQ-Diffusion"
      ],
      "metadata": {
        "id": "zecLuzidQ-dn"
      },
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "code",
      "source": [
        "!pip install torch torchvision torchtext\n",
        "!pip install omegaconf pytorch-lightning\n",
        "!pip install timm\n",
        "!pip install tensorboard\n",
        "!pip install lmdb tqdm\n",
        "!pip install einops ftfy\n",
        "!pip install git+https://github.com/openai/DALL-E.git"
      ],
      "metadata": {
        "id": "f7CP_5YZROMg"
      },
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "markdown",
      "source": [
        "On the first run, restart the Runtime and run the previous cell again"
      ],
      "metadata": {
        "id": "DD_z0fvSIsjj"
      }
    },
    {
      "cell_type": "code",
      "source": [
        "%cd /content/VQ-Diffusion"
      ],
      "metadata": {
        "id": "lWWA9ouISkKK"
      },
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "code",
      "source": [
        "from google.colab import drive\n",
        "drive.mount('/gdrive')"
      ],
      "metadata": {
        "id": "_cK_ghKuehPb"
      },
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "code",
      "source": [
        "### After you have already ran this Colab and downloaded the rather large model files\n",
        "### then at the end of this notebook there are cells to copy the stuff to your own\n",
        "### Google drive. Anyway, here you can copy it back from your drive.\n",
        "!cp --no-clobber -r '/gdrive/MyDrive/colab_data/MS-VQ-Diffusion/OUTPUT' '/content/VQ-Diffusion/'"
      ],
      "metadata": {
        "id": "aOGO_B-unjYL"
      },
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "code",
      "source": [
        "import multiprocessing\n",
        "import torch\n",
        "from psutil import virtual_memory\n",
        "\n",
        "ram_gb = round(virtual_memory().total / 1024**3, 1)\n",
        "\n",
        "print('CPU:', multiprocessing.cpu_count())\n",
        "print('RAM GB:', ram_gb)\n",
        "print(\"PyTorch version:\", torch.__version__)\n",
        "print(\"CUDA version:\", torch.version.cuda)\n",
        "print(\"cuDNN version:\", torch.backends.cudnn.version())\n",
        "device = torch.device(\"cuda:0\" if torch.cuda.is_available() else \"cpu\")\n",
        "print(\"device:\", device.type)\n"
      ],
      "metadata": {
        "id": "PaeQYus0ROmv"
      },
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "code",
      "source": [
        "#You can safely run these cells as --continue will not re-download if you already have it\n",
        "!wget --continue \"https://facevcstandard.blob.core.windows.net/t-shuygu/release_model/VQ-Diffusion/pretrained_model/taming_dvae/taming_f8_8192_openimages_last.ckpt?sv=2019-12-12&st=2021-12-18T05%3A29%3A52Z&se=2028-02-19T05%3A29%3A00Z&sr=b&sp=r&sig=bpTSMQFFcQWty%2FJd1CbZ3RH4uuMuIaFQEj2XTnU3a%2FI%3D\" -O /content/VQ-Diffusion/OUTPUT/pretrained_model/taming_dvae/taming_f8_8192_openimages_last.ckpt"
      ],
      "metadata": {
        "id": "0r-pXr2bUnSq"
      },
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "code",
      "source": [
        "!wget --continue \"https://facevcstandard.blob.core.windows.net/t-shuygu/release_model/VQ-Diffusion/pretrained_model/ViT-B-32.pt?sv=2019-12-12&st=2021-12-18T05%3A38%3A59Z&se=2028-03-19T05%3A38%3A00Z&sr=b&sp=r&sig=vjKexxaBROiGlS1iUkU7MzC0pMFjJBqSffj0SZCtVw0%3D\" -O /content/VQ-Diffusion/OUTPUT/pretrained_model/ViT-B-32.pt"
      ],
      "metadata": {
        "id": "XRqa-d_vXPIH"
      },
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "code",
      "source": [
        "!wget --continue \"https://facevcstandard.blob.core.windows.net/t-shuygu/release_model/VQ-Diffusion/pretrained_model/human_pretrained.pth?sv=2019-12-12&st=2021-12-18T05%3A41%3A14Z&se=2028-12-19T05%3A41%3A00Z&sr=b&sp=r&sig=tK%2BCuX9jpikRmGKg4lvx2MBOvU2JRwu7sq9pKNAenh4%3D\" -O /content/VQ-Diffusion/OUTPUT/pretrained_model/human_pretrained.pth"
      ],
      "metadata": {
        "id": "2n3O-F-mY9vm"
      },
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "code",
      "source": [
        "from inference_VQ_Diffusion import VQ_Diffusion\n",
        "VQ_Diffusion_model = VQ_Diffusion(config='OUTPUT/pretrained_model/config_text.yaml', path='OUTPUT/pretrained_model/human_pretrained.pth')\n",
        "VQ_Diffusion_model.inference_generate_sample_with_condition(\"a beautiful smiling woman\",truncation_rate=0.85, save_root=\"RESULT\",batch_size=4)\n",
        "VQ_Diffusion_model.inference_generate_sample_with_condition(\"a woman in yellow dress\",truncation_rate=0.85, save_root=\"RESULT\",batch_size=4,fast=2) # for fast inference"
      ],
      "metadata": {
        "id": "6wZWympKR_vR"
      },
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "code",
      "source": [
        "!ls -latr /content/VQ-Diffusion/RESULT/"
      ],
      "metadata": {
        "id": "PbyJi-vDd-Dv"
      },
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "code",
      "source": [
        "!wget --continue \"https://facevcstandard.blob.core.windows.net/t-shuygu/release_model/VQ-Diffusion/pretrained_model/taming_dvae/vqgan_imagenet_f16_16384.ckpt?sv=2019-12-12&st=2021-12-18T05%3A32%3A10Z&se=2028-12-19T05%3A32%3A00Z&sr=b&sp=r&sig=PAQzUkbD9bukKbLwqyuJpuzNfkr52ckwc%2B%2F5GkN7eGo%3D\" -O /content/VQ-Diffusion/OUTPUT/pretrained_model/taming_dvae/vqgan_imagenet_f16_16384.ckpt"
      ],
      "metadata": {
        "id": "As5ZtheAhDft"
      },
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "code",
      "source": [
        "!wget --continue \"https://facevcstandard.blob.core.windows.net/t-shuygu/release_model/VQ-Diffusion/pretrained_model/imagenet_pretrained.pth?sv=2019-12-12&st=2021-12-18T05%3A42%3A24Z&se=2028-12-19T05%3A42%3A00Z&sr=b&sp=r&sig=TOS5fddB5vSGKTLBsPMC5LhpyWRbZTlQKeIJrlQmiyg%3D\" -O /content/VQ-Diffusion/OUTPUT/pretrained_model/imagenet_pretrained.pth"
      ],
      "metadata": {
        "id": "Bo54XsaXh4GB"
      },
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "code",
      "source": [
        "from inference_VQ_Diffusion import VQ_Diffusion\n",
        "VQ_Diffusion_model = VQ_Diffusion(config='OUTPUT/pretrained_model/config_imagenet.yaml', path='OUTPUT/pretrained_model/imagenet_pretrained.pth')\n",
        "VQ_Diffusion_model.inference_generate_sample_with_class(409,truncation_rate=0.86, save_root=\"RESULT\",batch_size=4)\n",
        "\n",
        "# Here someone prepared an \"imagenet 1000 class idx to human readable labels\"\n",
        "# https://gist.github.com/yrevar/942d3a0ac09ec9e5eb3a\n"
      ],
      "metadata": {
        "id": "Kzwth6bpgzni"
      },
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "code",
      "source": [
        "!ls -latr /content/VQ-Diffusion/RESULT/"
      ],
      "metadata": {
        "id": "ve9_HPRvedap"
      },
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "code",
      "source": [
        "### After you have already ran this Colab and downloaded the rather large model files\n",
        "### then here you can copy the stuff into your own Google drive.\n",
        "\n",
        "!mkdir -p '/gdrive/MyDrive/colab_data/MS-VQ-Diffusion/'\n",
        "!cp --no-clobber -r '/content/VQ-Diffusion/OUTPUT' '/gdrive/MyDrive/colab_data/MS-VQ-Diffusion/'"
      ],
      "metadata": {
        "id": "NBtxxxsoehPc"
      },
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "code",
      "source": [
        "!mkdir -p '/gdrive/MyDrive/colab_data/MS-VQ-Diffusion/RESULT'\n",
        "%cd /content/VQ-Diffusion/RESULT/\n",
        "!cp -r '.' '/gdrive/MyDrive/colab_data/MS-VQ-Diffusion/RESULT/'\n",
        "%cd /content/VQ-Diffusion"
      ],
      "metadata": {
        "id": "fWTHrZ3Sfuvn"
      },
      "execution_count": null,
      "outputs": []
    }
  ]
}