min-dalle-test/min_dalle.ipynb

{
  "cells": [
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "view-in-github",
        "colab_type": "text"
      },
      "source": [
        "<a href=\"https://colab.research.google.com/github/kuprel/min-dalle/blob/main/min_dalle.ipynb\" target=\"_parent\"><img src=\"https://colab.research.google.com/assets/colab-badge.svg\" alt=\"Open In Colab\"/></a>"
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "3WL-G_f2_ld8"
      },
      "source": [
        "# min(DALL·E)"
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "Zl_ZFisFApeh"
      },
      "source": [
        "### Install"
      ]
    },
    {
      "cell_type": "code",
      "execution_count": 3,
      "metadata": {
        "cellView": "code",
        "id": "ix_xt4X1_6F4",
        "outputId": "918d48d7-9d8e-48fa-ab32-9fbc142e0074",
        "colab": {
          "base_uri": "https://localhost:8080/"
        }
      },
      "outputs": [
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "Looking in indexes: https://pypi.org/simple, https://us-python.pkg.dev/colab-wheels/public/simple/\n",
            "Collecting min-dalle==0.2.22\n",
            "  Downloading min-dalle-0.2.22.tar.gz (11 kB)\n",
            "Requirement already satisfied: torch>=1.10.0 in /usr/local/lib/python3.7/dist-packages (from min-dalle==0.2.22) (1.11.0+cu113)\n",
            "Requirement already satisfied: typing_extensions>=4.1.0 in /usr/local/lib/python3.7/dist-packages (from min-dalle==0.2.22) (4.1.1)\n",
            "Requirement already satisfied: numpy>=1.21 in /usr/local/lib/python3.7/dist-packages (from min-dalle==0.2.22) (1.21.6)\n",
            "Requirement already satisfied: pillow>=7.1 in /usr/local/lib/python3.7/dist-packages (from min-dalle==0.2.22) (7.1.2)\n",
            "Requirement already satisfied: requests>=2.23 in /usr/local/lib/python3.7/dist-packages (from min-dalle==0.2.22) (2.23.0)\n",
            "Requirement already satisfied: chardet<4,>=3.0.2 in /usr/local/lib/python3.7/dist-packages (from requests>=2.23->min-dalle==0.2.22) (3.0.4)\n",
            "Requirement already satisfied: urllib3!=1.25.0,!=1.25.1,<1.26,>=1.21.1 in /usr/local/lib/python3.7/dist-packages (from requests>=2.23->min-dalle==0.2.22) (1.24.3)\n",
            "Requirement already satisfied: idna<3,>=2.5 in /usr/local/lib/python3.7/dist-packages (from requests>=2.23->min-dalle==0.2.22) (2.10)\n",
            "Requirement already satisfied: certifi>=2017.4.17 in /usr/local/lib/python3.7/dist-packages (from requests>=2.23->min-dalle==0.2.22) (2022.6.15)\n",
            "Building wheels for collected packages: min-dalle\n",
            "  Building wheel for min-dalle (setup.py) ... \u001b[?25l\u001b[?25hdone\n",
            "  Created wheel for min-dalle: filename=min_dalle-0.2.22-py3-none-any.whl size=11263 sha256=84726f3fdf87ac6ccf6731fefadcc29154efee2a26020b771dcd08d4724ccf01\n",
            "  Stored in directory: /root/.cache/pip/wheels/7b/db/df/ec2e6cb890f0f527e178401b3f2fecaa7106748f2bdbc69bb9\n",
            "Successfully built min-dalle\n",
            "Installing collected packages: min-dalle\n",
            "Successfully installed min-dalle-0.2.22\n",
            "Mon Jul  4 21:23:14 2022       \n",
            "+-----------------------------------------------------------------------------+\n",
            "| NVIDIA-SMI 460.32.03    Driver Version: 460.32.03    CUDA Version: 11.2     |\n",
            "|-------------------------------+----------------------+----------------------+\n",
            "| GPU  Name        Persistence-M| Bus-Id        Disp.A | Volatile Uncorr. ECC |\n",
            "| Fan  Temp  Perf  Pwr:Usage/Cap|         Memory-Usage | GPU-Util  Compute M. |\n",
            "|                               |                      |               MIG M. |\n",
            "|===============================+======================+======================|\n",
            "|   0  Tesla P100-PCIE...  Off  | 00000000:00:04.0 Off |                    0 |\n",
            "| N/A   37C    P0    28W / 250W |      0MiB / 16280MiB |      0%      Default |\n",
            "|                               |                      |                  N/A |\n",
            "+-------------------------------+----------------------+----------------------+\n",
            "                                                                               \n",
            "+-----------------------------------------------------------------------------+\n",
            "| Processes:                                                                  |\n",
            "|  GPU   GI   CI        PID   Type   Process name                  GPU Memory |\n",
            "|        ID   ID                                                   Usage      |\n",
            "|=============================================================================|\n",
            "|  No running processes found                                                 |\n",
            "+-----------------------------------------------------------------------------+\n"
          ]
        }
      ],
      "source": [
        "! pip install min-dalle\n",
        "! nvidia-smi"
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "kViq2dMbGDKt"
      },
      "source": [
        "### Load Model"
      ]
    },
    {
      "cell_type": "code",
      "execution_count": 1,
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "8W-L2ICFGFup",
        "outputId": "2ea63766-5e01-4073-8b1a-364c6cb9a349"
      },
      "outputs": [
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "initializing MinDalle\n",
            "intializing TextTokenizer\n",
            "initializing DalleBartEncoder\n",
            "initializing DalleBartDecoder\n",
            "initializing VQGanDetokenizer\n"
          ]
        }
      ],
      "source": [
        "from PIL import Image\n",
        "from IPython.display import update_display\n",
        "import numpy\n",
        "from math import log2\n",
        "from min_dalle import MinDalle\n",
        "\n",
        "model = MinDalle(is_mega=True, is_reusable=True)"
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "c52TV1GbBNgS"
      },
      "source": [
        "### Generate Images\n",
        "Note: reduce the grid size if you run out of GPU memory.  4x4 has been tested to work on T4 and P100 (with intermediate_image_count = 1)"
      ]
    },
    {
      "cell_type": "code",
      "execution_count": 3,
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 563
        },
        "id": "nQ0UG05dA4p2",
        "outputId": "c638dd6b-f944-4d94-f357-87d4444ea3c5"
      },
      "outputs": [
        {
          "output_type": "display_data",
          "data": {
            "text/plain": [
              "<PIL.Image.Image image mode=RGB size=512x512 at 0x7F78F138C290>"
            ],
            "image/png": "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
          },
          "metadata": {}
        },
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "CPU times: user 31.2 s, sys: 193 ms, total: 31.4 s\n",
            "Wall time: 31.1 s\n"
          ]
        }
      ],
      "source": [
        "%%time\n",
        "\n",
        "text = \"Dali painting of WALL·E\" #@param {type:\"string\"}\n",
        "grid_size = 3 #@param {type:\"integer\"}\n",
        "seed = -1 #@param {type:\"integer\"}\n",
        "intermediate_image_count = 8 #@param [\"1\", \"2\", \"4\", \"8\", \"16\"] {type:\"raw\"}\n",
        "display_size = 512 #@param {type:\"integer\"}\n",
        "\n",
        "image_shape = (display_size, display_size, 3)\n",
        "zero_image = Image.fromarray(numpy.zeros(image_shape, dtype=numpy.uint8))\n",
        "display(zero_image, display_id=1)\n",
        "\n",
        "def handle_intermediate_image(row_index: int, image: Image.Image):\n",
        "    image = image.resize((display_size, display_size))\n",
        "    update_display(image, display_id=1)\n",
        "\n",
        "image = model.generate_image(\n",
        "    text,\n",
        "    seed,\n",
        "    grid_size,\n",
        "    log2(intermediate_image_count),\n",
        "    handle_intermediate_image\n",
        ")\n",
        "\n",
        "if intermediate_image_count == 1:\n",
        "    handle_intermediate_image(-1, image)"
      ]
    }
  ],
  "metadata": {
    "accelerator": "GPU",
    "colab": {
      "collapsed_sections": [
        "Zl_ZFisFApeh"
      ],
      "name": "min-dalle",
      "provenance": [],
      "include_colab_link": true
    },
    "gpuClass": "standard",
    "kernelspec": {
      "display_name": "Python 3",
      "name": "python3"
    },
    "language_info": {
      "name": "python"
    }
  },
  "nbformat": 4,
  "nbformat_minor": 0
}