ARC.ipynb

{
  "nbformat": 4,
  "nbformat_minor": 0,
  "metadata": {
    "colab": {
      "name": "ARC.ipynb",
      "provenance": [],
      "collapsed_sections": [],
      "include_colab_link": true
    },
    "kernelspec": {
      "name": "python3",
      "display_name": "Python 3"
    },
    "accelerator": "GPU"
  },
  "cells": [
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "view-in-github",
        "colab_type": "text"
      },
      "source": [
        "<a href=\"https://colab.research.google.com/github/tnfru/abstraction_and_reasoning/blob/master/ARC.ipynb\" target=\"_parent\"><img src=\"https://colab.research.google.com/assets/colab-badge.svg\" alt=\"Open In Colab\"/></a>"
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "rTty1fNJbePh",
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "outputId": "2100ab9b-e0ba-4eef-dd21-089a5da00fae"
      },
      "source": [
        "import sys\n",
        "if 'google.colab' in sys.modules:\n",
        "  %tensorflow_version 2.x\n",
        "\n",
        "  !wget https://raw.githubusercontent.com/tnfru/abstraction_and_reasoning/master/data_utils.py -P utils -nc\n",
        "\n",
        "  !wget https://raw.githubusercontent.com/tnfru/abstraction_and_reasoning/master/cellular_automata.py -P utils -nc\n",
        "  sys.path.append('utils')\n",
        "\n",
        "  from google_drive_downloader import GoogleDriveDownloader as gdd"
      ],
      "execution_count": 11,
      "outputs": [
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "File ‘utils/data_utils.py’ already there; not retrieving.\n",
            "\n",
            "File ‘utils/cellular_automata.py’ already there; not retrieving.\n",
            "\n"
          ]
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "grWM-1SYcqeb"
      },
      "source": [
        "import os, json\n",
        "import numpy as np\n",
        "from matplotlib import pyplot as plt\n",
        "\n",
        "import tensorflow as tf\n",
        "from tensorflow.keras.layers import Conv2D, BatchNormalization, Activation, Input, Dropout\n",
        "from tensorflow.keras.activations import softmax\n",
        "from tensorflow.keras import Sequential\n",
        "\n",
        "import data_utils\n",
        "from data_utils import plot_task, plot_prediction, inp2img\n",
        "import cellular_automata"
      ],
      "execution_count": 2,
      "outputs": []
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "NCEATf256UID"
      },
      "source": [
        "### Set strategy"
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "vd6ZfNAv6SdY",
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "outputId": "f73adab6-8b5c-49ca-e625-d68e6ee4299c"
      },
      "source": [
        "gpus = tf.config.experimental.list_logical_devices(\"GPU\")\n",
        "\n",
        "if len(gpus) > 0:\n",
        "  strategy = tf.distribute.MirroredStrategy(gpus)\n",
        "  print('Running on ', len(gpus), ' GPU(s) ')\n",
        "else:\n",
        "  strategy = tf.distribute.get_strategy()\n",
        "  print('Running on CPU')"
      ],
      "execution_count": 3,
      "outputs": [
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "INFO:tensorflow:Using MirroredStrategy with devices ('/job:localhost/replica:0/task:0/device:GPU:0',)\n",
            "Running on  1  GPU(s) \n"
          ]
        }
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "8o7bNcH-m1ZT"
      },
      "source": [
        "## Read Data"
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "7Mz6lUS9ens3",
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "outputId": "b2996c6b-fc84-4cd6-d958-2426e7d9e6b0"
      },
      "source": [
        "if 'google.colab' in sys.modules:\n",
        "  gdd.download_file_from_google_drive(file_id='1FUk30rXuym1RygfaDq4g-Uj-CkkG63IE',\n",
        "                                      dest_path='./data/arc.zip',\n",
        "                                      unzip=True)\n",
        "train_tasks, val_tasks, test_tasks = data_utils.get_task_sets()"
      ],
      "execution_count": 4,
      "outputs": [
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "Downloading 1FUk30rXuym1RygfaDq4g-Uj-CkkG63IE into ./data/arc.zip... Done.\n",
            "Unzipping...Done.\n"
          ]
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "2zuMRMgPHbuS",
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "outputId": "3bb78ea7-2afa-4304-cfcb-8ceb14402afa"
      },
      "source": [
        "print(\"Number of equal input and output shapes in train:\", data_utils.count_equal_dims(train_tasks), \"of\", len(train_tasks))\n",
        "print(\"Number of equal input and output shapes in val:\", data_utils.count_equal_dims(val_tasks), \"of\", len(val_tasks))\n",
        "print(\"Number of equal input and output shapes in test:\", data_utils.count_equal_dims(test_tasks, isTrain=False), \"of\", len(test_tasks))"
      ],
      "execution_count": 7,
      "outputs": [
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "Number of equal input and output shapes in train: 262 of 400\n",
            "Number of equal input and output shapes in val: 270 of 400\n",
            "Number of equal input and output shapes in test: 65 of 100\n"
          ]
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "-HptRnmngvmY"
      },
      "source": [
        "BENCHMARK_TASK_IDS = [1, 347, 26, 32]"
      ],
      "execution_count": 8,
      "outputs": []
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "DAipsbk0v67l"
      },
      "source": [
        "# Cellular Automata\n",
        "\n",
        "For every problem where the input shape is equal to the output shape we can use a Cellular Automata for the transformation.\n",
        "\n",
        "I want to thank [arseny-n](kaggle.com/arsenynerinovsky/cellular-automata-as-a-language-for-reasoning ) for introducing cellular automata to this problem set.\n",
        "\n",
        "The CA model I use here is an implementation of [teddykokers PyTorch CA](https://www.kaggle.com/teddykoker/training-cellular-automata-part-i-game-of-life) in tensorflow. I really like the the architecture teddy suggested, so I took the opportunity to practise translating a model from torch to tf.\n"
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "OGqdo8tpzgbD"
      },
      "source": [
        "## Vizualize results"
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "BJFh1vCMODYQ"
      },
      "source": [
        "def get_model():\n",
        "  inputs = Input(shape=(None, None, 10))\n",
        "  x = Conv2D(128, 3, padding='same')(inputs)\n",
        "  x = Activation('relu')(x)\n",
        "  x = Dropout(0.1)(x)\n",
        "  x = Conv2D(10, 1)(x)\n",
        "  outputs = softmax(x, axis=3)\n",
        "\n",
        "  return tf.keras.Model(inputs=inputs, outputs=outputs)"
      ],
      "execution_count": 9,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "Nvwgj9sTODYR",
        "outputId": "c2e83fa4-bc01-4f2e-e606-c7fcb8fcea9e",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 1000
        }
      },
      "source": [
        "cellular_automata.train_and_plot(train_tasks, [1], model_fn=get_model)"
      ],
      "execution_count": 10,
      "outputs": [
        {
          "output_type": "display_data",
          "data": {
            "image/png": "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\n",
            "text/plain": [
              "<Figure size 432x288 with 1 Axes>"
            ]
          },
          "metadata": {
            "needs_background": "light"
          }
        },
        {
          "output_type": "display_data",
          "data": {
            "image/png": "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\n",
            "text/plain": [
              "<Figure size 432x2304 with 3 Axes>"
            ]
          },
          "metadata": {
            "needs_background": "light"
          }
        },
        {
          "output_type": "display_data",
          "data": {
            "image/png": "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\n",
            "text/plain": [
              "<Figure size 432x2304 with 3 Axes>"
            ]
          },
          "metadata": {
            "needs_background": "light"
          }
        },
        {
          "output_type": "display_data",
          "data": {
            "image/png": "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\n",
            "text/plain": [
              "<Figure size 432x2304 with 3 Axes>"
            ]
          },
          "metadata": {
            "needs_background": "light"
          }
        },
        {
          "output_type": "display_data",
          "data": {
            "image/png": "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\n",
            "text/plain": [
              "<Figure size 432x2304 with 3 Axes>"
            ]
          },
          "metadata": {
            "needs_background": "light"
          }
        },
        {
          "output_type": "display_data",
          "data": {
            "image/png": "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\n",
            "text/plain": [
              "<Figure size 432x2304 with 3 Axes>"
            ]
          },
          "metadata": {
            "needs_background": "light"
          }
        },
        {
          "output_type": "display_data",
          "data": {
            "image/png": "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\n",
            "text/plain": [
              "<Figure size 432x2304 with 3 Axes>"
            ]
          },
          "metadata": {
            "needs_background": "light"
          }
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "QUBXux6hOYFb"
      },
      "source": [
        ""
      ],
      "execution_count": null,
      "outputs": []
    }
  ]
}