ML/Image_GPT_Sample_with_Conditioning.ipynb

{
  "nbformat": 4,
  "nbformat_minor": 0,
  "metadata": {
    "colab": {
      "name": "Image-GPT_Sample_with_Conditioning.ipynb",
      "provenance": [],
      "collapsed_sections": [
        "OoWYiUzT_lLs",
        "8ukKzRMrFBYb",
        "UjVa2W5f-6BB",
        "70aBx7tw_Ukc"
      ],
      "toc_visible": true,
      "machine_shape": "hm"
    },
    "kernelspec": {
      "name": "python3",
      "display_name": "Python 3"
    },
    "accelerator": "GPU"
  },
  "cells": [
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "oGskUEXmJIzs"
      },
      "source": [
        "Image GPT [https://openai.com/blog/image-gpt/](https://openai.com/blog/image-gpt/)\n",
        "\n",
        "Barebones demo, this just samples 32x32 images. That site shows lovely 64x64 images but the XL sized model isn't available. (Yet?)\n",
        "\n",
        "(Runtime)->(Run All) will work unless you get really unlucky with the GPU.\n",
        "\n",
        "Notebook by [https://twitter.com/jonathanfly](https://twitter.com/jonathanfly)\n",
        "\n",
        "Notebook modified to include conditional input by [Alfredo Peguero](https://twitter.com/alfredompeguero)"
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "OoWYiUzT_lLs"
      },
      "source": [
        "# Download Image-GPT"
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "TssZNq1DXarb",
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "outputId": "b8d63160-e247-42de-9463-0f7f955206f6"
      },
      "source": [
        "!nvidia-smi #OpenAI says you need 16GB GPU for the large model, but it may work if you lower n_sub_batch on the others."
      ],
      "execution_count": 1,
      "outputs": [
        {
          "output_type": "stream",
          "text": [
            "Tue Mar  2 18:35:06 2021       \n",
            "+-----------------------------------------------------------------------------+\n",
            "| NVIDIA-SMI 460.39       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"
          ],
          "name": "stdout"
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "P8iKEZQri1C4",
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "outputId": "5262e611-48a0-45f8-bae8-5981435f0abf"
      },
      "source": [
        "!git clone https://github.com/openai/image-gpt.git\n",
        "!pip install tensorflow-gpu==1.15.0"
      ],
      "execution_count": 2,
      "outputs": [
        {
          "output_type": "stream",
          "text": [
            "fatal: destination path 'image-gpt' already exists and is not an empty directory.\n",
            "Requirement already satisfied: tensorflow-gpu==1.15.0 in /usr/local/lib/python3.7/dist-packages (1.15.0)\n",
            "Requirement already satisfied: termcolor>=1.1.0 in /usr/local/lib/python3.7/dist-packages (from tensorflow-gpu==1.15.0) (1.1.0)\n",
            "Requirement already satisfied: absl-py>=0.7.0 in /usr/local/lib/python3.7/dist-packages (from tensorflow-gpu==1.15.0) (0.10.0)\n",
            "Requirement already satisfied: protobuf>=3.6.1 in /usr/local/lib/python3.7/dist-packages (from tensorflow-gpu==1.15.0) (3.12.4)\n",
            "Requirement already satisfied: keras-preprocessing>=1.0.5 in /usr/local/lib/python3.7/dist-packages (from tensorflow-gpu==1.15.0) (1.1.2)\n",
            "Requirement already satisfied: tensorboard<1.16.0,>=1.15.0 in /usr/local/lib/python3.7/dist-packages (from tensorflow-gpu==1.15.0) (1.15.0)\n",
            "Requirement already satisfied: google-pasta>=0.1.6 in /usr/local/lib/python3.7/dist-packages (from tensorflow-gpu==1.15.0) (0.2.0)\n",
            "Requirement already satisfied: six>=1.10.0 in /usr/local/lib/python3.7/dist-packages (from tensorflow-gpu==1.15.0) (1.15.0)\n",
            "Requirement already satisfied: tensorflow-estimator==1.15.1 in /usr/local/lib/python3.7/dist-packages (from tensorflow-gpu==1.15.0) (1.15.1)\n",
            "Requirement already satisfied: wheel>=0.26 in /usr/local/lib/python3.7/dist-packages (from tensorflow-gpu==1.15.0) (0.36.2)\n",
            "Requirement already satisfied: gast==0.2.2 in /usr/local/lib/python3.7/dist-packages (from tensorflow-gpu==1.15.0) (0.2.2)\n",
            "Requirement already satisfied: numpy<2.0,>=1.16.0 in /usr/local/lib/python3.7/dist-packages (from tensorflow-gpu==1.15.0) (1.19.5)\n",
            "Requirement already satisfied: wrapt>=1.11.1 in /usr/local/lib/python3.7/dist-packages (from tensorflow-gpu==1.15.0) (1.12.1)\n",
            "Requirement already satisfied: astor>=0.6.0 in /usr/local/lib/python3.7/dist-packages (from tensorflow-gpu==1.15.0) (0.8.1)\n",
            "Requirement already satisfied: opt-einsum>=2.3.2 in /usr/local/lib/python3.7/dist-packages (from tensorflow-gpu==1.15.0) (3.3.0)\n",
            "Requirement already satisfied: grpcio>=1.8.6 in /usr/local/lib/python3.7/dist-packages (from tensorflow-gpu==1.15.0) (1.32.0)\n",
            "Requirement already satisfied: keras-applications>=1.0.8 in /usr/local/lib/python3.7/dist-packages (from tensorflow-gpu==1.15.0) (1.0.8)\n",
            "Requirement already satisfied: setuptools in /usr/local/lib/python3.7/dist-packages (from protobuf>=3.6.1->tensorflow-gpu==1.15.0) (53.0.0)\n",
            "Requirement already satisfied: markdown>=2.6.8 in /usr/local/lib/python3.7/dist-packages (from tensorboard<1.16.0,>=1.15.0->tensorflow-gpu==1.15.0) (3.3.4)\n",
            "Requirement already satisfied: werkzeug>=0.11.15 in /usr/local/lib/python3.7/dist-packages (from tensorboard<1.16.0,>=1.15.0->tensorflow-gpu==1.15.0) (1.0.1)\n",
            "Requirement already satisfied: h5py in /usr/local/lib/python3.7/dist-packages (from keras-applications>=1.0.8->tensorflow-gpu==1.15.0) (2.10.0)\n",
            "Requirement already satisfied: importlib-metadata; python_version < \"3.8\" in /usr/local/lib/python3.7/dist-packages (from markdown>=2.6.8->tensorboard<1.16.0,>=1.15.0->tensorflow-gpu==1.15.0) (3.7.0)\n",
            "Requirement already satisfied: zipp>=0.5 in /usr/local/lib/python3.7/dist-packages (from importlib-metadata; python_version < \"3.8\"->markdown>=2.6.8->tensorboard<1.16.0,>=1.15.0->tensorflow-gpu==1.15.0) (3.4.0)\n",
            "Requirement already satisfied: typing-extensions>=3.6.4; python_version < \"3.8\" in /usr/local/lib/python3.7/dist-packages (from importlib-metadata; python_version < \"3.8\"->markdown>=2.6.8->tensorboard<1.16.0,>=1.15.0->tensorflow-gpu==1.15.0) (3.7.4.3)\n"
          ],
          "name": "stdout"
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "46kpjAMhXu6U",
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "outputId": "5865bb9e-12ef-471a-ab69-99b646b011fb"
      },
      "source": [
        "%cd /content/image-gpt"
      ],
      "execution_count": 3,
      "outputs": [
        {
          "output_type": "stream",
          "text": [
            "/content/image-gpt\n"
          ],
          "name": "stdout"
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "e0DcaUYv8LYf"
      },
      "source": [
        "model_sizes = [\"s\", \"m\", \"l\"] #small medium large, xl not available\n",
        "model_sizes = [\"s\"] #actually just download one\n",
        "n_sub_batch = 8 #8 is default, trying lowering if this doesn't work.\n",
        "n_px = 32\n",
        "n_gpu = 1"
      ],
      "execution_count": 4,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "ylcjIJcwXsFw",
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "outputId": "ed33204a-6b5b-4b69-ddc4-674bc85a248b"
      },
      "source": [
        "!mkdir -p /content/image-gpt/models\n",
        "!mkdir -p /content/image-gpt/clusters\n",
        "!mkdir -p /content/image-gpt/datasets\n",
        "\n",
        "for model_size in model_sizes:\n",
        "    !mkdir -p ./models/{model_size}\n",
        "    !python download.py --model {model_size} --ckpt 1000000 --download_dir ./models/{model_size} #models\n",
        "    #!python download.py --dataset imagenet --download_dir ./datasets/{model_size} #dataset\n",
        "    !python download.py --clusters --download_dir ./clusters/{model_size} #color clusters"
      ],
      "execution_count": 5,
      "outputs": [
        {
          "output_type": "stream",
          "text": [
            "input args:\n",
            " {\n",
            "    \"download_dir\":\"./models/s\",\n",
            "    \"bert\":false,\n",
            "    \"model\":\"s\",\n",
            "    \"ckpt\":\"1000000\",\n",
            "    \"clusters\":false,\n",
            "    \"dataset\":null\n",
            "}\n",
            "Fetching model.ckpt-1000000.data-00000-of-00032: 1.00kit [00:00, 1.19Mit/s]     \n",
            "Fetching model.ckpt-1000000.data-00001-of-00032: 31.5Mit [00:02, 15.2Mit/s]     \n",
            "Fetching model.ckpt-1000000.data-00002-of-00032: 28.3Mit [00:02, 14.1Mit/s]     \n",
            "Fetching model.ckpt-1000000.data-00003-of-00032: 28.3Mit [00:01, 14.6Mit/s]     \n",
            "Fetching model.ckpt-1000000.data-00004-of-00032: 28.3Mit [00:01, 15.4Mit/s]     \n",
            "Fetching model.ckpt-1000000.data-00005-of-00032: 31.5Mit [00:02, 15.6Mit/s]     \n",
            "Fetching model.ckpt-1000000.data-00006-of-00032: 30.4Mit [00:01, 16.4Mit/s]     \n",
            "Fetching model.ckpt-1000000.data-00007-of-00032: 28.3Mit [00:01, 15.6Mit/s]     \n",
            "Fetching model.ckpt-1000000.data-00008-of-00032: 31.5Mit [00:02, 14.1Mit/s]     \n",
            "Fetching model.ckpt-1000000.data-00009-of-00032: 29.4Mit [00:02, 13.8Mit/s]     \n",
            "Fetching model.ckpt-1000000.data-00010-of-00032: 31.5Mit [00:01, 16.2Mit/s]     \n",
            "Fetching model.ckpt-1000000.data-00011-of-00032: 31.5Mit [00:02, 15.3Mit/s]     \n",
            "Fetching model.ckpt-1000000.data-00012-of-00032: 30.4Mit [00:02, 13.7Mit/s]     \n",
            "Fetching model.ckpt-1000000.data-00013-of-00032: 28.3Mit [00:01, 14.7Mit/s]     \n",
            "Fetching model.ckpt-1000000.data-00014-of-00032: 31.5Mit [00:01, 16.3Mit/s]     \n",
            "Fetching model.ckpt-1000000.data-00015-of-00032: 32.5Mit [00:02, 14.8Mit/s]     \n",
            "Fetching model.ckpt-1000000.data-00016-of-00032: 31.5Mit [00:02, 15.4Mit/s]     \n",
            "Fetching model.ckpt-1000000.data-00017-of-00032: 28.3Mit [00:01, 15.2Mit/s]     \n",
            "Fetching model.ckpt-1000000.data-00018-of-00032: 29.4Mit [00:01, 15.3Mit/s]     \n",
            "Fetching model.ckpt-1000000.data-00019-of-00032: 28.3Mit [00:02, 13.3Mit/s]     \n",
            "Fetching model.ckpt-1000000.data-00020-of-00032: 28.3Mit [00:02, 14.1Mit/s]     \n",
            "Fetching model.ckpt-1000000.data-00021-of-00032: 28.3Mit [00:01, 14.5Mit/s]     \n",
            "Fetching model.ckpt-1000000.data-00022-of-00032: 30.4Mit [00:01, 15.6Mit/s]     \n",
            "Fetching model.ckpt-1000000.data-00023-of-00032: 30.4Mit [00:02, 13.3Mit/s]     \n",
            "Fetching model.ckpt-1000000.data-00024-of-00032: 29.4Mit [00:02, 13.5Mit/s]     \n",
            "Fetching model.ckpt-1000000.data-00025-of-00032: 28.3Mit [00:01, 14.9Mit/s]     \n",
            "Fetching model.ckpt-1000000.data-00026-of-00032: 28.3Mit [00:01, 15.7Mit/s]     \n",
            "Fetching model.ckpt-1000000.data-00027-of-00032: 29.4Mit [00:01, 14.8Mit/s]     \n",
            "Fetching model.ckpt-1000000.data-00028-of-00032: 29.4Mit [00:02, 13.2Mit/s]     \n",
            "Fetching model.ckpt-1000000.data-00029-of-00032: 29.4Mit [00:02, 14.6Mit/s]     \n",
            "Fetching model.ckpt-1000000.data-00030-of-00032: 28.3Mit [00:01, 14.8Mit/s]     \n",
            "Fetching model.ckpt-1000000.data-00031-of-00032: 28.3Mit [00:01, 15.5Mit/s]     \n",
            "Fetching model.ckpt-1000000.index: 10.0kit [00:00, 9.48Mit/s]                   \n",
            "Fetching model.ckpt-1000000.meta: 18.4Mit [00:01, 10.8Mit/s]                    \n",
            "input args:\n",
            " {\n",
            "    \"download_dir\":\"./clusters/s\",\n",
            "    \"bert\":false,\n",
            "    \"model\":null,\n",
            "    \"ckpt\":null,\n",
            "    \"clusters\":true,\n",
            "    \"dataset\":null\n",
            "}\n",
            "Fetching kmeans_centers.npy: 7.00kit [00:00, 4.32Mit/s]                         \n"
          ],
          "name": "stdout"
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "tauyrqoCGIeD"
      },
      "source": [
        "model_size = \"s\""
      ],
      "execution_count": 6,
      "outputs": []
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "8ukKzRMrFBYb"
      },
      "source": [
        "# Setup cropped images for conditioning"
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "sriUPI4YANqM"
      },
      "source": [
        "#numpy implementation of functions in src/utils which convert pixels of image to nearest color cluster. \n",
        "def normalize_img(img):\n",
        "  return img/127.5 - 1\n",
        "\n",
        "def squared_euclidean_distance_np(a,b):\n",
        "  b = b.T\n",
        "  a2 = np.sum(np.square(a),axis=1)\n",
        "  b2 = np.sum(np.square(b),axis=0)\n",
        "  ab = np.matmul(a,b)\n",
        "  d = a2[:,None] - 2*ab + b2[None,:]\n",
        "  return d\n",
        "\n",
        "def color_quantize_np(x, clusters):\n",
        "    x = x.reshape(-1, 3)\n",
        "    d = squared_euclidean_distance_np(x, clusters)\n",
        "    return np.argmin(d,axis=1)"
      ],
      "execution_count": 7,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "Oj7D8MjkiPRo",
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "outputId": "c7fdc7e5-7782-4483-f06f-8db8c843a205"
      },
      "source": [
        "#get images\n",
        "#!curl https://i.imgur.com/vF56Fsib.jpg > kp.jpg #LUCIA\n",
        "#!curl https://i.imgur.com/rlaBIBV.jpeg > kp.jpg #LUCIA\n",
        "!curl https://i.imgur.com/in9HnCx.jpg> kp.jpg \n",
        "\n",
        "image_paths = [\"kp.jpg\"]*(n_gpu*n_sub_batch)"
      ],
      "execution_count": 8,
      "outputs": [
        {
          "output_type": "stream",
          "text": [
            "  % Total    % Received % Xferd  Average Speed   Time    Time     Time  Current\n",
            "                                 Dload  Upload   Total   Spent    Left  Speed\n",
            "100 3723k  100 3723k    0     0  5755k      0 --:--:-- --:--:-- --:--:-- 5755k\n"
          ],
          "name": "stdout"
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "-zGV59gpC9oH",
        "outputId": "23082c62-6e01-4a54-94f6-34e466e9e484"
      },
      "source": [
        "image_paths"
      ],
      "execution_count": 9,
      "outputs": [
        {
          "output_type": "execute_result",
          "data": {
            "text/plain": [
              "['kp.jpg',\n",
              " 'kp.jpg',\n",
              " 'kp.jpg',\n",
              " 'kp.jpg',\n",
              " 'kp.jpg',\n",
              " 'kp.jpg',\n",
              " 'kp.jpg',\n",
              " 'kp.jpg']"
            ]
          },
          "metadata": {
            "tags": []
          },
          "execution_count": 9
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "khyeR9aC9kOH"
      },
      "source": [
        "#Resize original images to n_px by n_px\n",
        "import cv2\n",
        "import numpy as np\n",
        "dim=(n_px,n_px)\n",
        "\n",
        "x = np.zeros((n_gpu*n_sub_batch,n_px,n_px,3),dtype=np.uint8)\n",
        "\n",
        "for n,image_path in enumerate(image_paths):\n",
        "  img_np = cv2.imread(image_path)   # reads an image in the BGR format\n",
        "  img_np = cv2.cvtColor(img_np, cv2.COLOR_BGR2RGB)   # BGR -> RGB\n",
        "  H,W,C = img_np.shape\n",
        "  D = min(H,W)\n",
        "  img_np = img_np[:D,:D,:C] #get square piece of image\n",
        "  x[n] = cv2.resize(img_np,dim, interpolation = cv2.INTER_AREA) #resize to n_px by n_px\n",
        "  "
      ],
      "execution_count": 10,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "oZJGAwLxGuXi",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 191
        },
        "outputId": "1f39d3a1-04fa-4eed-e346-417c0b67ec84"
      },
      "source": [
        "#visualize resized images\n",
        "%matplotlib inline\n",
        "import pathlib\n",
        "import matplotlib.pyplot as plt\n",
        "import matplotlib.image as mpimg\n",
        "\n",
        "f, axarr = plt.subplots(1,8,dpi=180)\n",
        "\n",
        "i = 0\n",
        "for img in x:\n",
        "    axarr[i].imshow(img)\n",
        "    i += 1"
      ],
      "execution_count": 11,
      "outputs": [
        {
          "output_type": "display_data",
          "data": {
            "image/png": "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\n",
            "text/plain": [
              "<Figure size 1080x720 with 8 Axes>"
            ]
          },
          "metadata": {
            "tags": [],
            "needs_background": "light"
          }
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "QccafXMDTgaj"
      },
      "source": [
        "#use Image-GPT color palette and crop images\n",
        "\n",
        "color_cluster_path = \"/content/image-gpt/clusters/%s/kmeans_centers.npy\"%(model_size)\n",
        "clusters = np.load(color_cluster_path) #get color clusters\n",
        "x_norm = normalize_img(x) #normalize pixels values to -1 to +1\n",
        "\n",
        "samples = color_quantize_np(x_norm,clusters).reshape(x_norm.shape[:-1]) #map pixels to closest color cluster\n",
        "\n",
        "n_px_crop = 8\n",
        "primers = samples.reshape(-1,n_px*n_px)[:,:n_px_crop*n_px] # crop top n_px_crop rows"
      ],
      "execution_count": 12,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "ON0A6H8kUWhW",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 298
        },
        "outputId": "505f4ba6-b5f2-4b23-9e75-fbad16f7d3fe"
      },
      "source": [
        "\n",
        "#visualize samples and crops with Image-GPT color palette. Should look similar to original resized images\n",
        "%matplotlib inline\n",
        "import pathlib\n",
        "import matplotlib.pyplot as plt\n",
        "import matplotlib.image as mpimg\n",
        "\n",
        "samples_img = [np.reshape(np.rint(127.5 * (clusters[s] + 1.0)), [n_px, n_px, 3]).astype(np.uint8) for s in samples] # convert color clusters back to pixels\n",
        "primers_img = [np.reshape(np.rint(127.5 * (clusters[s] + 1.0)), [n_px_crop,n_px, 3]).astype(np.uint8) for s in primers] # convert color clusters back to pixels\n",
        "\n",
        "f, axarr2 = plt.subplots(1,8,dpi=180)\n",
        "i = 0\n",
        "for img in samples_img:\n",
        "    axarr2[i].imshow(img)\n",
        "    i += 1\n",
        "\n",
        "f, axarr = plt.subplots(1,8,dpi=180)\n",
        "i = 0\n",
        "for img in primers_img:\n",
        "    axarr[i].imshow(img)\n",
        "    i += 1"
      ],
      "execution_count": 13,
      "outputs": [
        {
          "output_type": "display_data",
          "data": {
            "image/png": "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\n",
            "text/plain": [
              "<Figure size 1080x720 with 8 Axes>"
            ]
          },
          "metadata": {
            "tags": [],
            "needs_background": "light"
          }
        },
        {
          "output_type": "display_data",
          "data": {
            "image/png": "iVBORw0KGgoAAAANSUhEUgAAA4oAAABrCAYAAAAxfXxCAAAABHNCSVQICAgIfAhkiAAAAAlwSFlzAAAbrgAAG64BjF1z+AAAADh0RVh0U29mdHdhcmUAbWF0cGxvdGxpYiB2ZXJzaW9uMy4yLjIsIGh0dHA6Ly9tYXRwbG90bGliLm9yZy+WH4yJAAAgAElEQVR4nO3dfZAkd33f8ff31w/zsLv3JB0CnfWABMZggvFTCVARZIQMJoUdExs/yDEQApRdsV2Yh7JTqbKd/JPyQ+yyK45dZcup2FV2Ko7LcSrBIXaBHwQJIIiNMQHJCAhId+ike9yHme7f75s/undv5vbhVrqZ22nxeVW1bnamd6an3727+k1P95i7IyIiIiIiIrIpHPQCiIiIiIiIyGLRQFFERERERESmaKAoIiIiIiIiUzRQFBERERERkSkaKIqIiIiIiMgUDRRFRERERERkigaKIiIiIiIiMkUDRREREREREZmigaKIiIiIiIhM0UBRREREREREpmigKCIiIiIiIlM0UBQREREREZEpGiiKiIiIiIjIFA0URUREREREZIoGiiIiIiIiIjJFA0URERERERGZooGiiIiIiIiITNFAUURERERERKbMdaBoZitm9tNm9gkzu2hm58zsI2b2TjMr5/nYsn/q1B1q1R1q1R1q1R1q1R1q1R1qJbsxd5/PHZvdAnwAuLW9ag3IgF779ceBu939zFwWQPZFnbpDrbpDrbpDrbpDrbpDrbpDrWQvc9mjaGY58F9pNrpHgXvcfQkYAt8LXAC+HvideTy+7I86dYdadYdadYdadYdadYdadYdayRW5+8wn4C2At9NLd7j9+yZuv3sey6BJnZ5Ok1p1Z1Kr7kxq1Z1JrbozqVV3JrXSdKVpXscovrH99/3u/qEdbv894OH28g/OaRnkytSpO9SqO9SqO9SqO9SqO9SqO9RK9jTzYxTNbEizqzoA73H3n9tlvl8Ffgg46e7PmsHjzudgS9nG3e1qvl+trh216g616g616g616o6raaVO15ZadcPV/v6D+Ryj+PyJ+/2bPebbvO2ZZnZsDsshIiIiIiIiT0E+h/u8ceLyl/aYb/K2G4En9rrT/b4C8fa3vglSwrzG3EnR8ZiaG7McshwzowxGHgzMIXcIjmFknmEYeR4oikAIgX7Zp1f2yPKM4eEhveU+ZZ5zeLjMoOjR7xccPTKkLHPWR2Murq0TYyJWThw7daw5fe4Jzlw8TxaM5UGfflFwaGWFm7/qBCsry7hDSs37gat6RDVeJ7mTHDw5yZ3xaExd1YyryLnVEaOqJrNAaRnBjFE9Yr3aoI6Rx8+t8vj5VVJMeO143ayC0L628JmHPsvv/f4fAfDD//T7ufmZ1xMCHDl6iKPXHSbLMgA+/NG/4V//4n37WfVqpVZqpVZqpVZqpVYzb6VOs+1kZnz6wb/bavVjb/1+Tjzr+m2d1GoxWjWczJwy+I4/U5Nef++P77vXlcxjoLgycXltj/kmb1vZda4nqQhQ1zWjjXVSTODgNBtVUTgFkGUZ/f6Asl+S5YH+MCMvjF7RZ6W/TJHlDJcKDq2UZHnGoDek318iC4FeP6co8+Z+PYBDCEZeNPGLmOgVRgwwNsMNQgz0+iVLdb/ZsDZqxmsVRSgIBmWREWOicm82wBgZ1xUpJca1U9dQVTVnnzjL6sU1RlXN2QsbbIxrhoOSY0eWKIuc5Al3B4dYJ6r1SEwO7XVmYBYwYDyuttbZcDDk0KFl8iywsjxkOCix0OwUXl4ZzCqNWqmVWqmVWqmVWqnV3Fqp0/46cVmro0cPceTwoW2d1GoxWjWMsiw4NCx2/Jmal3kMFOfiSu+z3XwFo1ldToyJFOPk7RhOMMiCkecZRVGQl4H+oKAsA4PegMNLK5R5wfJyyZEjPYoiY9BfZtBfIoRAWWTkmZEcxuM27KVlJIRACIa7EYJDMMyNLAvkWUaKiSpG6qqmrhN4s0yJS2egTZ6IMZJSoq6dcbvhrW+MWV1dZzSuuXix2fAMZ7xcErLm+5tnD56aVzzS5qszQMDwdgOcFIKR5zlFnpHnOVkIWxteuHxmtVIrtVIrtVIrtVKra9hKnebXCSDP8mZ9XNapafWkUqnVAf1MsXWPszWPgeKFicvDPeabvO3CrnM9SU34QMgCmNPvlwwGPfI857ojRzl66DBZntHr98jLgiwL9PqBLA+UecmwNyTPMvr9nEE/J8sCWZZDGy3GBIStjWTb41sgb3eX1yliJMyMIisYlgPG44oLo3VWV9cJWcGjjz3BRjWmrhPjcbOxrY3WWR2tkZJjHjACKUZiimRFRhmMZYdeHRkMepR5SWYZG9WY9fURdUxsjKpLy9cuqzcvUYAZvV6xtcx1PaaONZCo6x4xGsENC85oNJ5VGrVSK7VSK7VSK7VSq7m1Uqf9dTKDQf9Sq7WNDaq62tbJDEajatvzVKtr2yoLzYA9mO/6M7U5qKwnBpyzMI+B4iMTl08Af73LfCd2+Z6rkmU5Wd4ECm6sHF3m+PGjDPo9br/5Zm49cYIsy7BgzasWxtS/wbLmcjBCMIxm9++mOjoxNa987LDdEUKgyAuylFHVFdgYw+jlPYpewWpcZ23tcb78+EU2qkjRzzl9ZkCsIuNRRYqJ1WrM+fEIgKW8YJgXzbLlgaJfkDuU/RJ3J88KekWPYIHxaMTjZ1apqsi4jlvL5zHhKeEh4CEAxuGV5a1lvnDhAuNqTF0Her1IVRlZZuQ5PP7EuVml2Uat1GqTWs2OWqnVJrWaHbXqRit12l+nYHDd0UtHfZ1+4gme+Yxj2zpZ5px+/IxaHXCrXmGEAMHSrj9TljkpOVUVtz/ZqzCPgeKnaPbMBuCFwHt3me+F7b8n3X3PE9k8OQYYFgy82WXd6xX0Bj2WV4YcObKy44GfwI6vOOx0++RsfunGPe8jWLOLO1ggRqeqa0bjivWNEVluxCpSj2pSSqxXY9bGY3AnL6FwIwuBkAeybPOHwNr7zTALOLa1gVRVPfVeZxw8Od4eX2xmHD9+DDPD3Tn15dO84HnPxq05CDdFx3BS5nzuC80Y/siRFc6endmO34nnoFZqpVazpVZqpVZq9ZXaSp3208kMnvmM67ZaPXrqNF/7Nbdv6xRwPveFRwE4cniFs+fU6iBahWDkmQGbexy3/0wFHKd5G+sszXyg6O5rZnY/8HLgNcC2z1G05hQ+r26/fN+sl2EnRvNe52IqXrOhxORbZzByr3fdeJzmwNXmIFSoY9M2eSLFGk+Jqq4ZV6M2dMAJYGCZ4YAVhlvzfuy6qllf38BIlHnB0mCIWSCtXmR9PMbNKIuCfn9AngWGyyW9fvNqRRaaV1nW1ivOnh0xGtesboypY2qfT2oPDnZSiqQUsQCWNcuw3BvynNtu4sG/+wIPfvb/8brX3tU8fxIXVy8QglH2Mj76sb8F4Ote+NX82V8+MP9QqJVaqdU8qJVaqdXsqVU3WqnTdKeyl3NsZYXn3n4zn3no8zz42S/wum97xVSnXj+nLDMe+PinAHjRC7+aP79frQ6i1eGVPmWRkZJTx7TtZ6rXz+n1cpzm/mZpXiezeZBmoPjKfZwm9z/MaRm2ybJAUWRTG15KTqoSzSg9UdcV7juvZHenqsdUdU1KMKqdOkKKNePROilFqiqyPhrjDsPhkKWlJcwM8mbjY2Q4iZhqqspYW10j1RWHVw6xcniZoiioY+Ti6gXcnX5RMhgMKYqMI4f7LC0V7UG6JVnIOHX6PF88eZ4LFzdY3xgxjmlro/MYcXdSrJsNz2mWozTKYcHd33IHD/7dF/jUZx7mrT/6L/dcdyduvGGGJa5MrVArtZo5tWKq1YkTN+y71ctf9o3XbPABaqVWajVr6sRUpyPHVrjlphv4zEOf5/9+5nO87cf+1Z7r7+/f+Y3XZKAIarVTq6V+ybiKbIxqYkxsrG1w7vxZzIzlNCQxwGx6L+oszOucqvdPXE7AGeBUO53l0h7g/+XufzqnZdjim/91b3d0NxuRt69MpJSaXbgpkVJzEOrUFC9NdR2p65o61sT60nXNvzVVXVNXzXWbo/pm13G7e32Xc0EZhoWwtcs7s/ay2aUDftv3YYf2FFQOxOTU9ebu69hcu/kebprAzfc032+h+TrLAq/6lju4/rojW8uwtDTg0MoSK8tDBv3e1vV5nvHc22+ZeZedqJVaqdXsqdXOrb72a2679Phmu7Z6zu038cIXPGfmXXaiVmqlVrOlTjt3yvOcv/eC515aBjOGwz4ry8Om1aC/dZtaXXJQrbamLJBnAQvg3g42272sm//O0rz2KG4O8ev2MY7SfG5iADa3vI8Dr53T47eaD+/0KpLGkbpKjOuIxdSM5qsKd4gpkRLEWFFVG3iaPhA0RSMlwz0xrsfUscIskFsPs4zmUza3P3oWoMwhBIipmZoDbwMh5JRlyZHDyywt9SmLklEcUaWK5FCWzUlhy7IgL5yQJcZVBWuRqk5c3BgzrhPnzq5y5swZ1tfHEIysMDIzsmSEFNpd+M1GRzBCP8fyQFnk9PsDvusf3sOv/eZ/AmB1dZ1eWZDcqarmEz9vveVG/sW738ag19v+BGdKrdRKrWZPrfZqVRTNWf9CCKSUdm31E+94C3HGJwjYTq3USq3U6Vp2KinKcqrV2toGZVk0e+Q2W918I+/5kTdTzensp2q131Z9sOYMqTEmUupv7T0synxzNRLrxT+ZzaRHgH8PvB54NlABnwR+F/gVd5/fuaxbnpxURVIeiXWkqpr39p67uMbq+nrzwkU7b6yrrV3Sk2IVqOvmgN8qjYipIg85K71AL28OH/UdtrwsGGXRnLFoVDcbHlj7KkRG2Ss5fHiFw4eH1HVkNB6RkhOBXm8JwyjLjCx3rD3TUVXD6vqYL54+y/nVDcbrIy6eWaWuIr1hSX8waHbZOxTuZMHolT3KvCDhVBaJJHpFTln2OHbsKABLwwHHrz/CyVOPk4fAV524gTvveDGvuedOspDN+RdEQ63USq1mT612b5Xnzf/QXn/dEb71lS/hLz/0cbVSq31Rq260Uqe9/1Zlbavrjh3mW1/5Ev7iQ/+HU19+nCwEbjpxAy+748W8+u6X4dGpR7VaHWCrouw1b41t9yJCM5B19+Z62uM4OzZQdHf/KeCn5vw4W7Z2GzdftQvRbFybI2+Hrd3YEwu6tfv28jMGJffmPto72rzs7fuktzY6t+2vVHhz9dT1W8t4add0CAELqdn9nAXyvJknyzJCCJv739uDcdvd13VNSk6eBQJQljm9siBkgSI5uTvBrLm93f1t3i5Tcqo6bn3eyvLSgH/3S/+c8Tgyrur2QOH2+c/4M1kurQa1Uiu1mjW12n+reuLtRv/4e/8B937Pt6mVWu1IrbrRSp2e2t+qYMYPfO9recPrX71jpypWu74V86lSqyfXalzF5njG9vk0jxmmWu00CL5a8x4oXnOWZ4SYkZPjJCBjbEaO0b4jeFfuRqyt+QDLACFr3iOdh0CRN2cTqr2ZggU81FQ0p6itIyRvvnfzMerkjOom/p6/D83Ii5zloodjeGomDIrCyAsDd2JdTb160rz3v88NR1coQqDo9+gNB80pfcc1adz8sG+e4YmUSHUipZq11XUe8dOcPtN8llFyZzSqMYNemeNO8/7umLDQLN+sqZVaqZVaNQt9MK3Onr8ING9nWl0fEdrjRdRqr4VWK7Va3FbqNL+/VWHixDKzoFZPrlWW5/TywKDIrunP1LwHisfN7AHgeUAGPAp8EPgNd//APB7QskDIMjLLcDcgUNO85zVe4dUQd4ix2fjIndB+fEseAnkocJyMRLZ1CGYiWiIlo47NxpJS+yqFQUpQ1c17nvdcZiDLc4qyT9h8UJ+4EfD2wF0u280+GJQ86/hRhr2CvOhR9vqAsbE2Zn1tTIqR9bVVRqON5pWJmPA6sVFtcH5jxNkLzS/08+cv8o6f+HkePfkYyZ1jRw/xvOfcyite/s284GtuI+SzP++RWqmVWqkVHFyrC6urQNPpx3/yF3j05GN4co4dO8Tzn/dsXvmKO/jq59wKoFaby4xaqdXitlKn+f2tCjM+/6VaPblWCWNl0Oe6Q0uUWUavl1MWOdCcwGZr7/CMf6bmPVAcAt9Ac9bTJZrjFJ8N3GtmvwW8zd339abnfXzMBsDWKNua/+A0G0BKfukUO7RnEto8G5E7vvV97USzO7zZ5dx+tgqGuRHIaHbw+tTjujUHpqYQwKbfM8zEru/2+TS7ry1s3a+1u8JDMELWfO/WbnQDI2A0P6x5llFkGXmWkU2eJcmaDSSm5kDkWEfGdXNArbuThYAVOZkZYOTthj4aVzz8+S+xvDRgYzTm5KnHOXnqcf7s/ge46+XfzNve/F37Wf1qpVZqpVadabXZaTyu+Nznv8TS0oDRRKf3//lH207/iCzb/59LtVIrtZptK3U6+L9VeZap1QG2coysPelQbUZWJ/IsgrP1Flxr552leQ0UHwF+BvgD4NPuPjKzDLijvf5VwJuBVeBHZvnAeTA8C9RFRnPmI6hGkdwiMTbvWQ5m9Hs5wXokT9SxJqWEJaiKhJmTWbM73DCyPCNkAQeyusBihpOIzY5sQrDm4G03ohuZO5iT55d+qOpYkeqauhoTQvP+5CIvKKwkp0dIWfMyijlFP2MwaA4w3hjXjKrYLAc5RqCXGUeGQ8o8Y9gryS57CcTduXBxjVNffoIYE7GuiammLHKOHhm2951hFJy9+Syv//Z7eNGLnssznnGMpV6P3AIPPfxFfv+//E8+8bcP8YG/+AiDfp9ZUyu1Uiu1OshWz735Jr77O17DN3z98zl63QpZHsgxvvD5R/n9P/qTrU69Xslb3/jdaqVWarXgrdRpfn+r3vrGN6jVAbYCYxzHbFTrjKox1ThjvJE3g8PQjJrNAlnowFtP3f19wPsuuy4CHzSzV9MMIL8D+GEz+2V3f3Af97nnEHnzFYytzzHJmlcV3I1YJWKWSPHSiLvMMzLLSSkxqlJzqtkMsqwJl7WhA6EJUBjmEFKOxYSTSMRmw8MoQoYRiM2Yv9nwsuZVD9xJMVLXY2KsCAZ50b66YAUZRbPrugY3JzejX+ZgUKfEqIo0r080j1EGZ6nXJ8+MomgPnp1aV7C+MeLMuQvEiTdbuxm9fo/Dh4cEcgp63HPnS/nmF7+Ak2dPM44V6xtjSsu45cZn8e4ffRO//Ou/y0c//kn+x5/85RW7q5VaqZVadanVq+68g3vufBlnV89x8uxpNsZjCjJuuvGZvOPtP8C/ve8/8sBffYr3/ekH+fbXvFKt1EqtDqiVOh3836pvf82r1OoAWxkZZ1fPcXH9IuNYEcmIFgkWyHs5eZljQGb72/O7X9f8ZDbunszsXTQDxQC8Dvg383u8RIrNZHh7xqJ2F3W7q3lzizYzQhYIDubN57AknDzkW7vcU3TcA8kjyZtd2uZhaxf1TmcccqCOkdG4oqprCJDlgZAbWKI5E1M7I8a4qllfH+EGq6sjLqyNms9+qRyLUHtNShEza84wVkU8Nu+9TingDpkZy4NB88pLAAvNqzJ51rwXPLqTYgVeszEeMxrXVHUkZIbnGVjzgeA/eO/r+OjHP0nyfb2T4KqolVqp1eyp1VNrlYXAvW94LQ/81adwdz7ysU/MuZRaqZVaqdNidNrpb9WHP/ZXarWgrUJozoBqFnZ8XlfjQM566u4Pmdlp4Hrgtrk+VoxEHxEzwzxRFs17g2M04mUfNWIhIy96WGheUYjVGDNjUOQsDUsMI+YQK4gesRiovcJjIm6eTpgEl0VKKXFxfYNz589T1wmyxNJyQW8QSFlNxYjmA0QN3Fg/v8GZ1eZzUp44u8q58+vgTpEZuTUHqpb95rS6dZ2oNirwZuOykGEY/Tzn1htvIASj6Bt52ezWT7VRj6GqK9ZG56nqmosXNzhzZpVYJ1ju01tuXh0p8ozbbn4Wh1aWOH9hdZ6Z1Eqt1EqtFq7VTSdu2Op06rHTaqVWatWxVuo0u79Vj5w6pVYL3GrQK3Cgqrr1OYoHzt3xGEl1jeHtgaRG2GHHePOZJAVYcwahZM3KDhn0y4zmjExGDFCnjLpOeIRETV2PSb5zHKc5qHt1fYS7N1F7gbw0sEikJnlzJqaUIKaKKo2JMfLEmQs8cWYVA/qlUeRGWRYcypcps5IYE9Wowic+S8bMOHroCEdWDpHlgf5SoBwY43HizJkN1tcrxnXk4sY6o2rE6tqYtdUxKSaGZbNb3aD5INAiA5vtgbG7USu1UqvZU6ura2Vbna7F3l+1Uiu1miV1muHfqjm/A0atrr5VSv70GCia2e00exMBHj6IZXgqdvrfOnsK/7PXfMtT/Z/Ep/CDuu2h9v/Yjzz6GOfbz0TqErXqDrXqjq+0VidPneZc2+kZx6978o97gNSqO9SqG77SOsH036rjx489+cc+IF+JrebFfMavEJiZ+R53ak2p/wx8J80JcF/g7p+ewePO/yVEAbjigcpXolbXjlp1h1p1h1p1h1p1x9W0UqdrS6264Wp//0FzMplZu8XMPmxmbzez29qBIWYWzOwlwHtpBokAvz6LQaKIiIiIiIjMzjz2KN7K9NtJR8AFYAXoTVz/W8Db3L2e8eM3JySawSj66WK3daJWi2cRW6nTztSqO3ZaL/r9t5jUqjvUqhsW8W/VXsv1lWwR18k8BooD4C3AS4EXA8eBo8AG8EXgg8B97n7/TB/40uMv3Eo+aHv8klCrBbOIrdRpZ2rVHbv8D61+/y0gteoOteqGRfxbtddyfSVbxHUy84HiQVvElXzQFnWdLOpyHaRFXCeLuEyLYBHXyyIu0yJYxPWyiMu0CBZxvSziMi2CRVwvi7hMB21R18miLtdBWsR1Mo9jFEVERERERKTDNFAUERERERGRKRooioiIiIiIyBQNFEVERERERGSKBooiIiIiIiIy5Wl31lMRERERERG5OtqjKCIiIiIiIlM0UBQREREREZEpGiiKiIiIiIjIFA0URUREREREZIoGiiIiIiIiIjJFA0URERERERGZooGiiIiIiIiITNFAUURERERERKZooCgiIiIiIiJTnhYDRTNbMbOfNrNPmNlFMztnZh8xs3eaWXnQyzdrZnadmb3ZzH7HzP7WzFbNbGRmXzSzPzSz79zje99kZr6P6VVzWna1UquFpFbdoVbdoVbdoVbdoE7d0eVWm/J53vm1YGa3AB8Abm2vWgN6wDe1071mdre7nzmQBZyPk0y32wAq4EQ7fYeZvRf4Lndf2+U+EvDYHo8xmsWCTlIrQK0WmVp1h1p1h1p1h1p1gzp1RydbTXH3zk40K/+vAQceAV7VXh+A7wHOt7f9t4Ne1hk/bwf+N/BDwG0T198K/EZ7uwO/vcP3vqm97XNqpVZqpVZdnNSqO5NadWdSq25M6tSdqYutti3HQa/EqwzwlomV/NIdbv++idvvPujlneHz/pYr3P5rE8/7pstuO6hfEmqlVgs9qVV3JrXqzqRW3ZnUqhuTOnVn6mKry6euH6P4xvbf97v7h3a4/feAh9vLP3htFmn+3P39V5jlNycuf9M8l+VJUKudqdWCUKvuUKvuUKvuUKtuUKfu6GirKZ0dKJrZELiz/fK9O83jzZD8j9svv/VaLNeC2Ji4nB3YUrTUak9q1R1q1R1q1R1q1R1q1Q3q1B0L1WonnR0oAs/n0vL/zR7zbd72TDM7Nt9FWhh3TVz+xC7zHDezB9qzTq2b2WfbszLdtcv8V0OtdnfXxGW1Wmx3TVxWq8V218RltVpsd01cVqvFdtfEZbVaXHdNXFanxXbXxOVFaLVNlweKN05c/tIe803eduOucz1NmNkR4CfbL//C3T+9y6xD4BuAMc128GzgXuD9Znafmc3yjLhqtQO16g616g616g616g616gZ16o4FbbVNlweKKxOXdzul7OW3rew619OAmQXgt4Fn0ezO/mc7zPYI8DPA1wF9dz9GsxHeCfxJO8+bgV+c4aKp1WXUqjvUqjvUqjvUqjvUqhvUqTsWuNV2B3kmnauZgO/n0pmCnrPHfPewx5mWnk4T8CsTz/WfPIXvD8Aftt8fgeeqlVqplVp1ZVKr7kxq1Z1JrboxqVN3pkVttdPU5T2KFyYuD/eYb/K2C7vO1XFm9vNcekXiHe5+35O9D3dPwLvaLwPwuhktnlpNUKvuUKvuUKvuUKvuUKtuUKfuWPBW23R5oPjIxOUTe8w3edsju87VYWb2s8A72y/f5e6/9FTvy90fAk63X952tcvWUquWWnWHWnWHWnWHWnWHWnWDOnVHB1pt0+WB4qeA1F5+4R7zbd520t2fmO8iXXtm9nPAu9sv3+Puv3CQy7MLtUKtukStukOtukOtukOtukGduqMjrbbp7EDR3deA+9svX7PTPGZmwKvbL993LZbrWmp3X2/uen6Pu//cDO7zduD69suH95p3v9RKrbpErbpDrbpDrbpDrbpBnbqjK612dFAHcs5iAt5CcyBnAu7Y4fY3cOlg0bsPenln/Nx/fuK5vXOf32NXuh34Ay4dHPs8tVIrtVKrRZzUqjuTWnVnUqtuTOrUnalrrbY91kGvwKtc+Tnw1+2K+uLmxkWzp/S7gXPtbf/9oJd1xs/7Zyc2unc8ie+7Ffgw8Haa9zPbxPp6CfDHE/f7q2qlVmqlVos4qVV3JrXqzqRW3ZjUqTtTF1ttW5aDXokziHArzS7XzRW2CqxPfP0x4OhBL+cMn+/NE88tAievML3rsnXlE9MG8Fj77+T19wG5WqmVWqnVok1q1Z1JrbozqVU3JnXqztTlVpNTTse5++fM7EU07/19PfBsoAI+Cfwu8CvuPj7ARZy1cNnlG64w//LE5VPAjwAvBV4MHAeO0mx4DwMfBO5z9/svv5NZUCu1WmBq1R1q1R1q1R1q1Q3q1B2dbTVpc1emiIiIiIiICNDhs56KiIiIiIjIfGigKCIiIiIiIlM0UBQREREREZEpGiiKiIiIiIjIFA0URUREREREZIoGiiIiIiIiIjJFA0URERERERGZooGiiIiIiIiITNFAUURERERERKZooCgiIiIiIiJTNFAUERERERGRKRooioiIiIiIyBQNFEVERERERGSKBooiIiIiIrF5zjgAAABOSURBVCIyRQNFERERERERmaKBooiIiIiIiEzRQFFERERERESmaKAoIiIiIiIiUzRQFBERERERkSkaKIqIiIiIiMgUDRRFRERERERkyv8HFQKHsYBBdv0AAAAASUVORK5CYII=\n",
            "text/plain": [
              "<Figure size 1080x720 with 8 Axes>"
            ]
          },
          "metadata": {
            "tags": [],
            "needs_background": "light"
          }
        }
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "UjVa2W5f-6BB"
      },
      "source": [
        "# Functions from run.py\n"
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "L4gn7wHcKWCn"
      },
      "source": [
        "\n",
        "import argparse\n",
        "import json\n",
        "import math\n",
        "import os\n",
        "import random\n",
        "import sys\n",
        "import time\n",
        "sys.path.append('src')\n",
        "\n",
        "import numpy as np\n",
        "import tensorflow as tf\n",
        "\n",
        "from imageio import imwrite\n",
        "from scipy.special import softmax\n",
        "from tensorflow.contrib.training import HParams\n",
        "from tqdm import tqdm\n",
        "\n",
        "from model import model\n",
        "from utils import iter_data, count_parameters\n",
        "\n",
        "\n",
        "def parse_arguments():\n",
        "    parser = argparse.ArgumentParser()\n",
        "\n",
        "    # data and I/O\n",
        "    parser.add_argument(\"--data_path\", type=str, default=\"/root/downloads/imagenet\")\n",
        "    parser.add_argument(\"--ckpt_path\", type=str, default=\"/root/downloads/model.ckpt-1000000\")\n",
        "    parser.add_argument(\"--color_cluster_path\", type=str, default=\"/root/downloads/kmeans_centers.npy\")\n",
        "    parser.add_argument(\"--save_dir\", type=str, default=\"/root/save/\")\n",
        "\n",
        "    # model\n",
        "    parser.add_argument(\"--n_embd\", type=int, default=512)\n",
        "    parser.add_argument(\"--n_head\", type=int, default=8)\n",
        "    parser.add_argument(\"--n_layer\", type=int, default=24)\n",
        "    parser.add_argument(\"--n_px\", type=int, default=32, help=\"image height or width in pixels\")\n",
        "    parser.add_argument(\"--n_vocab\", type=int, default=512, help=\"possible values for each pixel\")\n",
        "\n",
        "    parser.add_argument(\"--bert\", action=\"store_true\", help=\"use the bert objective (defaut: autoregressive)\")\n",
        "    parser.add_argument(\"--bert_mask_prob\", type=float, default=0.15)\n",
        "    parser.add_argument(\"--clf\", action=\"store_true\", help=\"add a learnable classification head\")\n",
        "\n",
        "    # parallelism\n",
        "    parser.add_argument(\"--n_sub_batch\", type=int, default=8, help=\"per-gpu batch size\")\n",
        "    parser.add_argument(\"--n_gpu\", type=int, default=8, help=\"number of gpus to distribute training across\")\n",
        "\n",
        "    # mode\n",
        "    parser.add_argument(\"--eval\", action=\"store_true\", help=\"evaluates the model, requires a checkpoint and dataset\")\n",
        "    parser.add_argument(\"--sample\", action=\"store_true\", help=\"samples from the model, requires a checkpoint and clusters\")\n",
        "\n",
        "    # reproducibility\n",
        "    parser.add_argument(\"--seed\", type=int, default=42, help=\"seed for random, np, tf\")\n",
        "\n",
        "    args = parser.parse_args()\n",
        "    print(\"input args:\\n\", json.dumps(vars(args), indent=4, separators=(\",\", \":\")))\n",
        "    return args\n",
        "\n",
        "\n",
        "def set_seed(seed):\n",
        "    random.seed(seed)\n",
        "    np.random.seed(seed)\n",
        "    tf.set_random_seed(seed)\n",
        "\n",
        "\n",
        "def load_data(data_path):\n",
        "    trX = np.load(f'{data_path}_trX.npy')\n",
        "    trY = np.load(f'{data_path}_trY.npy')\n",
        "    vaX = np.load(f'{data_path}_vaX.npy')\n",
        "    vaY = np.load(f'{data_path}_vaY.npy')\n",
        "    teX = np.load(f'{data_path}_teX.npy')\n",
        "    teY = np.load(f'{data_path}_teY.npy')\n",
        "    return (trX, trY), (vaX, vaY), (teX, teY)\n",
        "\n",
        "\n",
        "def set_hparams(args):\n",
        "    return HParams(\n",
        "        n_ctx=args.n_px*args.n_px,\n",
        "        n_embd=args.n_embd,\n",
        "        n_head=args.n_head,\n",
        "        n_layer=args.n_layer,\n",
        "        n_vocab=args.n_vocab,\n",
        "        bert=args.bert,\n",
        "        bert_mask_prob=args.bert_mask_prob,\n",
        "        clf=args.clf,\n",
        "    )\n",
        "\n",
        "\n",
        "def create_model(x, y, n_gpu, hparams):\n",
        "    gen_logits = []\n",
        "    gen_loss = []\n",
        "    clf_loss = []\n",
        "    tot_loss = []\n",
        "    accuracy = []\n",
        "\n",
        "    trainable_params = None\n",
        "    for i in range(n_gpu):\n",
        "        with tf.device(\"/gpu:%d\" % i):\n",
        "            results = model(hparams, x[i], y[i], reuse=(i != 0))\n",
        "\n",
        "            gen_logits.append(results[\"gen_logits\"])\n",
        "            gen_loss.append(results[\"gen_loss\"])\n",
        "            clf_loss.append(results[\"clf_loss\"])\n",
        "\n",
        "            if hparams.clf:\n",
        "                tot_loss.append(results[\"gen_loss\"] + results[\"clf_loss\"])\n",
        "            else:\n",
        "                tot_loss.append(results[\"gen_loss\"])\n",
        "\n",
        "            accuracy.append(results[\"accuracy\"])\n",
        "\n",
        "            if i == 0:\n",
        "                trainable_params = tf.trainable_variables()\n",
        "                print(\"trainable parameters:\", count_parameters())\n",
        "\n",
        "    return trainable_params, gen_logits, gen_loss, clf_loss, tot_loss, accuracy\n",
        "\n",
        "\n",
        "def reduce_mean(gen_loss, clf_loss, tot_loss, accuracy, n_gpu):\n",
        "    with tf.device(\"/gpu:0\"):\n",
        "        for i in range(1, n_gpu):\n",
        "            gen_loss[0] += gen_loss[i]\n",
        "            clf_loss[0] += clf_loss[i]\n",
        "            tot_loss[0] += tot_loss[i]\n",
        "            accuracy[0] += accuracy[i]\n",
        "        gen_loss[0] /= n_gpu\n",
        "        clf_loss[0] /= n_gpu\n",
        "        tot_loss[0] /= n_gpu\n",
        "        accuracy[0] /= n_gpu\n",
        "\n",
        "\n",
        "def evaluate(sess, evX, evY, X, Y, gen_loss, clf_loss, accuracy, n_batch, desc, permute=False):\n",
        "    metrics = []\n",
        "    for xmb, ymb in iter_data(evX, evY, n_batch=n_batch, truncate=True, verbose=True):\n",
        "        metrics.append(sess.run([gen_loss[0], clf_loss[0], accuracy[0]], {X: xmb, Y: ymb}))\n",
        "    eval_gen_loss, eval_clf_loss, eval_accuracy = [np.mean(m) for m in zip(*metrics)]\n",
        "    print(f\"{desc} gen: {eval_gen_loss:.4f} clf: {eval_clf_loss:.4f} acc: {eval_accuracy:.2f}\")\n",
        "\n",
        "\n",
        "# naive sampler without caching\n",
        "def sample(sess, X, gen_logits, n_sub_batch, n_gpu, n_px, n_vocab, clusters, save_dir,primers=None):\n",
        "  \n",
        "    samples = np.zeros([n_gpu * n_sub_batch, n_px * n_px], dtype=np.int32)\n",
        "\n",
        "    if primers is None:\n",
        "      N_cond_px = 0\n",
        "    else:\n",
        "      N_cond_px = primers.shape[1]\n",
        "      samples[:,:N_cond_px] = primers\n",
        "      \n",
        "\n",
        "    print('Conditioning on %d out of %d pixels'%(N_cond_px,n_px*n_px))\n",
        "    for i in tqdm(range(N_cond_px,n_px * n_px), ncols=80, leave=False):\n",
        "        np_gen_logits = sess.run(gen_logits, {X: samples})\n",
        "        for j in range(n_gpu):\n",
        "            p = softmax(np_gen_logits[j][:, i, :], axis=-1)  # logits to probas\n",
        "            for k in range(n_sub_batch):\n",
        "                c = np.random.choice(n_vocab, p=p[k])  # choose based on probas\n",
        "                samples[j * n_sub_batch + k, i] = c\n",
        "    \n",
        "    # dequantize\n",
        "    samples = [np.reshape(np.rint(127.5 * (clusters[s] + 1.0)), [32, 32, 3]).astype(np.uint8) for s in samples]\n",
        "\n",
        "    # write to png\n",
        "    for i in range(n_gpu * n_sub_batch):\n",
        "        imwrite(f\"{args.save_dir}/seed_{args.seed}_sample_{i}.png\", samples[i])\n",
        "\n",
        "\n",
        "def main(args,primers=None):\n",
        "    tf.reset_default_graph()\n",
        "    set_seed(args.seed)\n",
        "\n",
        "    n_batch = args.n_sub_batch * args.n_gpu\n",
        "\n",
        "    if args.sample:\n",
        "        n_class = 1000\n",
        "        print(\"Skipping dataset requirement for sampling.\")\n",
        "    else:\n",
        "        if args.data_path.endswith(\"cifar10\"):\n",
        "            n_class = 10\n",
        "        elif args.data_path.endswith(\"imagenet\"):\n",
        "            n_class = 1000\n",
        "        else:\n",
        "            raise ValueError(\"Dataset not supported.\")\n",
        "\n",
        "    X = tf.placeholder(tf.int32, [n_batch, args.n_px * args.n_px])\n",
        "    Y = tf.placeholder(tf.float32, [n_batch, n_class])\n",
        "\n",
        "    x = tf.split(X, args.n_gpu, 0)\n",
        "    y = tf.split(Y, args.n_gpu, 0)\n",
        "\n",
        "    hparams = set_hparams(args)\n",
        "    trainable_params, gen_logits, gen_loss, clf_loss, tot_loss, accuracy = create_model(x, y, args.n_gpu, hparams)\n",
        "    reduce_mean(gen_loss, clf_loss, tot_loss, accuracy, args.n_gpu)\n",
        "\n",
        "    saver = tf.train.Saver(var_list=[tp for tp in trainable_params if not 'clf' in tp.name])\n",
        "    with tf.Session(config=tf.ConfigProto(allow_soft_placement=True, log_device_placement=False)) as sess:\n",
        "        sess.run(tf.global_variables_initializer())\n",
        "\n",
        "        saver.restore(sess, args.ckpt_path)\n",
        "\n",
        "        if args.eval:\n",
        "            (trX, trY), (vaX, vaY), (teX, teY) = load_data(args.data_path)\n",
        "            evaluate(sess, trX[:len(vaX)], trY[:len(vaY)], X, Y, gen_loss, clf_loss, accuracy, n_batch, \"train\")\n",
        "            evaluate(sess, vaX, vaY, X, Y, gen_loss, clf_loss, accuracy, n_batch, \"valid\")\n",
        "            evaluate(sess, teX, teY, X, Y, gen_loss, clf_loss, accuracy, n_batch, \"test\")\n",
        "\n",
        "        if args.sample:\n",
        "            if not os.path.exists(args.save_dir):\n",
        "                os.makedirs(args.save_dir)\n",
        "            clusters = np.load(args.color_cluster_path)\n",
        "            sample(sess, X, gen_logits, args.n_sub_batch, args.n_gpu, args.n_px, args.n_vocab, clusters, args.save_dir,primers=primers)\n",
        "\n"
      ],
      "execution_count": 14,
      "outputs": []
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "70aBx7tw_Ukc"
      },
      "source": [
        "# Sample"
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "9PfdQPv4TmK7"
      },
      "source": [
        "#set model hyperparameters\n",
        "MODELS={\"l\":(1536,16,48),\"m\":(1024,8,36),\"s\":(512,8,24) }\n",
        "n_embd,n_head,n_layer=MODELS[model_size]"
      ],
      "execution_count": 15,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "AwuR8Wq7KgnL",
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "outputId": "08e78e98-669c-416b-8c9c-cec96025a5c7"
      },
      "source": [
        "\n",
        "sys.argv=\"\"\"src/run.py  --sample --n_embd %d --n_head %d --n_layer %d\n",
        "--ckpt_path /content/image-gpt/models/%s/model.ckpt-1000000 --color_cluster_path /content/image-gpt/clusters/%s/kmeans_centers.npy \n",
        "--data_path /content/image-gpt/datasets/s/imagenet_notused --save_dir /content/image-gpt/output \n",
        "--n_gpu %d --n_px %d --n_sub_batch %d --seed 42\"\"\"%(n_embd,n_head,n_layer,model_size,model_size,n_gpu,n_px,n_sub_batch)\n",
        "\n",
        "sys.argv=sys.argv.split()\n",
        "\n",
        "args = parse_arguments()\n",
        "\n",
        "main(args,primers=primers) #conditional generation\n",
        "#main(args) #unconditional generation"
      ],
      "execution_count": 16,
      "outputs": [
        {
          "output_type": "stream",
          "text": [
            "input args:\n",
            " {\n",
            "    \"data_path\":\"/content/image-gpt/datasets/s/imagenet_notused\",\n",
            "    \"ckpt_path\":\"/content/image-gpt/models/s/model.ckpt-1000000\",\n",
            "    \"color_cluster_path\":\"/content/image-gpt/clusters/s/kmeans_centers.npy\",\n",
            "    \"save_dir\":\"/content/image-gpt/output\",\n",
            "    \"n_embd\":512,\n",
            "    \"n_head\":8,\n",
            "    \"n_layer\":24,\n",
            "    \"n_px\":32,\n",
            "    \"n_vocab\":512,\n",
            "    \"bert\":false,\n",
            "    \"bert_mask_prob\":0.15,\n",
            "    \"clf\":false,\n",
            "    \"n_sub_batch\":8,\n",
            "    \"n_gpu\":1,\n",
            "    \"eval\":false,\n",
            "    \"sample\":true,\n",
            "    \"seed\":42\n",
            "}\n",
            "Skipping dataset requirement for sampling.\n",
            "WARNING:tensorflow:From src/model.py:156: The name tf.variable_scope is deprecated. Please use tf.compat.v1.variable_scope instead.\n",
            "\n",
            "WARNING:tensorflow:From src/model.py:164: The name tf.get_variable is deprecated. Please use tf.compat.v1.get_variable instead.\n",
            "\n",
            "WARNING:tensorflow:From src/model.py:37: The name tf.rsqrt is deprecated. Please use tf.math.rsqrt instead.\n",
            "\n",
            "trainable parameters: 76571648\n",
            "INFO:tensorflow:Restoring parameters from /content/image-gpt/models/s/model.ckpt-1000000\n"
          ],
          "name": "stdout"
        },
        {
          "output_type": "stream",
          "text": [
            "\r  0%|                                                   | 0/768 [00:00<?, ?it/s]"
          ],
          "name": "stderr"
        },
        {
          "output_type": "stream",
          "text": [
            "Conditioning on 256 out of 1024 pixels\n"
          ],
          "name": "stdout"
        },
        {
          "output_type": "stream",
          "text": [
            ""
          ],
          "name": "stderr"
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "4s522c8lGqy5",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 191
        },
        "outputId": "6c6f3c9f-74d6-44e5-ed95-57075a8e6695"
      },
      "source": [
        "#visualize output samples\n",
        "%matplotlib inline\n",
        "import pathlib\n",
        "import matplotlib.pyplot as plt\n",
        "import matplotlib.image as mpimg\n",
        "import glob\n",
        "\n",
        "samples = pathlib.Path('/content/image-gpt/output').glob('*.png')\n",
        "\n",
        "f, axarr = plt.subplots(1,len(glob.glob('/content/image-gpt/output/*.png')),dpi=180)\n",
        "\n",
        "i = 0\n",
        "for image in samples:\n",
        "    axarr[i].imshow(mpimg.imread(image))\n",
        "    i += 1"
      ],
      "execution_count": 17,
      "outputs": [
        {
          "output_type": "display_data",
          "data": {
            "image/png": "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\n",
            "text/plain": [
              "<Figure size 1080x720 with 8 Axes>"
            ]
          },
          "metadata": {
            "tags": [],
            "needs_background": "light"
          }
        }
      ]
    }
  ]
}