Atari_DQN.ipynb

{
  "cells": [
    {
      "cell_type": "markdown",
      "id": "56cf3ed9",
      "metadata": {
        "id": "56cf3ed9"
      },
      "source": [
        "<div align=center>\n",
        "\t\t\n",
        "<p></p>\n",
        "<p></p>\n",
        "<font size=5>\n",
        "In the Name of God\n",
        "<font/>\n",
        "<p></p>\n",
        " <br/>\n",
        "    <br/>\n",
        "    <br/>\n",
        "<font color=#FF7500>\n",
        "Sharif University of Technology - Departmenet of Computer Engineering\n",
        "</font>\n",
        "<p></p>\n",
        "<font color=blue>\n",
        "Artifical Intelligence - Dr. Mohammad Hossein Rohban\n",
        "</font>\n",
        "<br/>\n",
        "<br/>\n",
        "Fall 2021\n",
        "\n",
        "</div>\n",
        "\n",
        "<hr/>\n",
        "\t\t<div align=center>\n",
        "\t\t    <font color=red size=6>\n",
        "\t\t\t    <br />\n",
        "Practical Assignment 5-Q2\n",
        "            \t<br/>\n",
        "\t\t\t</font>\n",
        "    <br/>\n",
        "    <br/>\n",
        "<font size=4>\n",
        "\t\t\t<br/><br/>\n",
        "Deadline:  Bahman 17th\n",
        "                <br/><b>\n",
        "              Cheating is Strongly Prohibited\n",
        "                </b><br/><br/>\n",
        "                <font color=red>\n",
        "Please run all the cells.\n",
        "     </font>\n",
        "</font>\n",
        "                <br/>\n",
        "    </div>"
      ]
    },
    {
      "cell_type": "markdown",
      "id": "f00402a1",
      "metadata": {
        "id": "f00402a1"
      },
      "source": [
        "# Personal Data\n"
      ]
    },
    {
      "cell_type": "code",
      "execution_count": 1,
      "id": "10c79f3a",
      "metadata": {
        "id": "10c79f3a"
      },
      "outputs": [],
      "source": [
        "# Set your student number\n",
        "student_number = 97101489\n",
        "Name = 'Amirhosein'\n",
        "Last_Name = 'Javadi'"
      ]
    },
    {
      "cell_type": "code",
      "source": [
        "from google.colab import drive\n",
        "drive.mount('/content/drive')"
      ],
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "60F6OrzPM_lW",
        "outputId": "b65b2a7b-52ee-436d-fbf0-2593cd70af46"
      },
      "id": "60F6OrzPM_lW",
      "execution_count": 2,
      "outputs": [
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "Drive already mounted at /content/drive; to attempt to forcibly remount, call drive.mount(\"/content/drive\", force_remount=True).\n"
          ]
        }
      ]
    },
    {
      "cell_type": "code",
      "source": [
        "!unzip /content/Q2.zip"
      ],
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "_H_EBmYqQoKU",
        "outputId": "36ae2bc1-957e-4b36-833a-d56fddc90f00"
      },
      "id": "_H_EBmYqQoKU",
      "execution_count": 3,
      "outputs": [
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "Archive:  /content/Q2.zip\n",
            "replace .ipynb_checkpoints/Atari_DQN-checkpoint.ipynb? [y]es, [n]o, [A]ll, [N]one, [r]ename: "
          ]
        }
      ]
    },
    {
      "cell_type": "markdown",
      "id": "68003dd5",
      "metadata": {
        "id": "68003dd5"
      },
      "source": [
        "# Rules\n",
        "- You are not allowed to use provided codes that can be found on the internet. \n",
        "- If you want to use a library which is not already imported, you must ask a question on Quera to get the permission of using that.\n",
        "- Do not hesitate to ask questions on Quera, if you have any.\n",
        "- This assignment is due Bahman 17th 23:59:59. you can use up to 1 grace day for this assignment and the hard deadline is Bahman 18th 23:59:59."
      ]
    },
    {
      "cell_type": "markdown",
      "id": "4a58d6af",
      "metadata": {
        "id": "4a58d6af"
      },
      "source": [
        "# Mastering Atari Games using Deep Reinforcement Learning\n",
        "\n",
        "In this assignment, you will learn how to implement a Deep-Q-Network (DQN), and to demonstrate that, you may implement and train a DQN and run it on an Atari game in **OpenAI-Gym** environment. Additionally, you may also want to refer to the [original paper](https://arxiv.org/pdf/1312.5602.pdf) regarding the usage of DQN in Atari networks.\n",
        "\n",
        "For this assignment, you will have two options to choose from:\n",
        "* **Breakout**: A famous Atari game that is a staple for learning deep q networks. This game is already available in the gym library environment. By solving this completely you will receive 100/100 points.\n",
        "* **Shoot-N-Dodge**: This is another Atari game that has been developed for the sole reason of this particular assignment. The game involves a spaceship with guns and a bunch of meteors coming your way! For those of you who are up for the challenge, you may run your DQN on this game and receive **extra points**. By implementing and training a network capable of solving this custom game you will receive 150/100 points with 50 points being extra.\n",
        "\n",
        "In the following, you can read the description of the games and how to start running it. We advise using either `Google-Colab` or `Kaggle` for training the model as DQN models usually take a lot of time to train.\n",
        "\n",
        "Before we start, make sure to install and import the following packages:"
      ]
    },
    {
      "cell_type": "code",
      "execution_count": 4,
      "id": "2d46bde2",
      "metadata": {
        "id": "2d46bde2",
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "outputId": "4080916b-d14c-4db4-b0d0-2becd6601777"
      },
      "outputs": [
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "Requirement already satisfied: gym in /usr/local/lib/python3.7/dist-packages (0.17.3)\n",
            "Requirement already satisfied: cloudpickle<1.7.0,>=1.2.0 in /usr/local/lib/python3.7/dist-packages (from gym) (1.3.0)\n",
            "Requirement already satisfied: numpy>=1.10.4 in /usr/local/lib/python3.7/dist-packages (from gym) (1.19.5)\n",
            "Requirement already satisfied: pyglet<=1.5.0,>=1.4.0 in /usr/local/lib/python3.7/dist-packages (from gym) (1.5.0)\n",
            "Requirement already satisfied: scipy in /usr/local/lib/python3.7/dist-packages (from gym) (1.4.1)\n",
            "Requirement already satisfied: future in /usr/local/lib/python3.7/dist-packages (from pyglet<=1.5.0,>=1.4.0->gym) (0.16.0)\n",
            "Requirement already satisfied: opencv-python in /usr/local/lib/python3.7/dist-packages (4.1.2.30)\n",
            "Requirement already satisfied: numpy>=1.14.5 in /usr/local/lib/python3.7/dist-packages (from opencv-python) (1.19.5)\n",
            "Requirement already satisfied: pillow in /usr/local/lib/python3.7/dist-packages (7.1.2)\n",
            "Requirement already satisfied: numpy in /usr/local/lib/python3.7/dist-packages (1.19.5)\n",
            "Requirement already satisfied: matplotlib in /usr/local/lib/python3.7/dist-packages (3.2.2)\n",
            "Requirement already satisfied: numpy>=1.11 in /usr/local/lib/python3.7/dist-packages (from matplotlib) (1.19.5)\n",
            "Requirement already satisfied: kiwisolver>=1.0.1 in /usr/local/lib/python3.7/dist-packages (from matplotlib) (1.3.2)\n",
            "Requirement already satisfied: python-dateutil>=2.1 in /usr/local/lib/python3.7/dist-packages (from matplotlib) (2.8.2)\n",
            "Requirement already satisfied: cycler>=0.10 in /usr/local/lib/python3.7/dist-packages (from matplotlib) (0.11.0)\n",
            "Requirement already satisfied: pyparsing!=2.0.4,!=2.1.2,!=2.1.6,>=2.0.1 in /usr/local/lib/python3.7/dist-packages (from matplotlib) (3.0.7)\n",
            "Requirement already satisfied: six>=1.5 in /usr/local/lib/python3.7/dist-packages (from python-dateutil>=2.1->matplotlib) (1.15.0)\n",
            "Requirement already satisfied: torch in /usr/local/lib/python3.7/dist-packages (1.10.0+cu111)\n",
            "Requirement already satisfied: typing-extensions in /usr/local/lib/python3.7/dist-packages (from torch) (3.10.0.2)\n",
            "Requirement already satisfied: psutil in /usr/local/lib/python3.7/dist-packages (5.4.8)\n",
            "Requirement already satisfied: array2gif in /usr/local/lib/python3.7/dist-packages (1.0.4)\n",
            "Requirement already satisfied: numpy in /usr/local/lib/python3.7/dist-packages (from array2gif) (1.19.5)\n"
          ]
        }
      ],
      "source": [
        "!pip install gym\n",
        "!pip install opencv-python\n",
        "!pip install pillow\n",
        "!pip install numpy\n",
        "!pip install matplotlib\n",
        "!pip install torch\n",
        "!pip install psutil\n",
        "!pip install array2gif"
      ]
    },
    {
      "cell_type": "code",
      "execution_count": 5,
      "id": "2ea4a50d",
      "metadata": {
        "id": "2ea4a50d"
      },
      "outputs": [],
      "source": [
        "import gym\n",
        "import numpy as np\n",
        "import pandas as pd\n",
        "import matplotlib.pyplot as plt\n",
        "import time\n",
        "from IPython.display import clear_output\n",
        "\n",
        "%matplotlib inline"
      ]
    },
    {
      "cell_type": "code",
      "execution_count": 6,
      "id": "452677bd",
      "metadata": {
        "id": "452677bd",
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "outputId": "aca3087f-9ba0-418d-b7c4-75bcdc074283"
      },
      "outputs": [
        {
          "output_type": "execute_result",
          "data": {
            "text/plain": [
              "device(type='cuda')"
            ]
          },
          "metadata": {},
          "execution_count": 6
        }
      ],
      "source": [
        "# Deep Learning requirements\n",
        "### make sure to use gpu for this assignment\n",
        "import torch\n",
        "import torch.nn as nn\n",
        "device = torch.device('cuda' if torch.cuda.is_available() else 'cpu')\n",
        "device"
      ]
    },
    {
      "cell_type": "markdown",
      "id": "bf742a67",
      "metadata": {
        "id": "bf742a67"
      },
      "source": [
        "As alluded, we will use the `gym` framework to simulate our game."
      ]
    },
    {
      "cell_type": "markdown",
      "id": "0f7a8ba5",
      "metadata": {
        "id": "0f7a8ba5"
      },
      "source": [
        "## Option 1: Breakout (60 points)\n",
        "\n",
        "The famous game [Breakout](https://en.wikipedia.org/wiki/Breakout_(video_game)) involves a paddle and bricks. It is one of the games in which DQNs have proven to work very good on. \n",
        "\n",
        "<img src=\"figs/nerd.png\" width=400 height=400 />"
      ]
    },
    {
      "cell_type": "code",
      "execution_count": 7,
      "id": "dfb9d1e2",
      "metadata": {
        "id": "dfb9d1e2"
      },
      "outputs": [],
      "source": [
        "# Run this block if you intend to implement the breakout option\n",
        "\n",
        "ENV_NAME = \"BreakoutNoFrameskip-v4\""
      ]
    },
    {
      "cell_type": "markdown",
      "id": "a091fa47",
      "metadata": {
        "id": "a091fa47"
      },
      "source": [
        "## Option 2: Shoot-N-Dodge (60+30 points)"
      ]
    },
    {
      "cell_type": "markdown",
      "id": "5b7eda9f",
      "metadata": {
        "id": "5b7eda9f"
      },
      "source": [
        "In this game you control a space ship that shoots bullets. At each frame, you may either move up, down, or stay still. Every once in a while some meteors colored in red show up at random at the other end of the screen and you should avoid or destroy them. If either one of these meteors hits you, you will die!\n",
        "\n",
        "To destroy the meteors, you have to shoot them three times and each time you shoot them their color changes until they disappear. They change color from red to orange to yellow and then they are gone.\n",
        "\n",
        "<img src=\"figs/shoot-n-dodge.jpg\" width=430 height=420 />"
      ]
    },
    {
      "cell_type": "code",
      "execution_count": 9,
      "id": "e820c808",
      "metadata": {
        "id": "e820c808"
      },
      "outputs": [],
      "source": [
        "# Run this block if you intend to implement the ShootNDodge option\n",
        "\n",
        "ENV_NAME = 'shootndodge-v0'"
      ]
    },
    {
      "cell_type": "markdown",
      "id": "7ea975a6",
      "metadata": {
        "id": "7ea975a6"
      },
      "source": [
        "\n",
        "### Installing the game\n"
      ]
    },
    {
      "cell_type": "markdown",
      "id": "d3799488",
      "metadata": {
        "id": "d3799488"
      },
      "source": [
        "#### Breakout\n",
        "\n",
        "In order to install Breakout on your environment, please execute the following code. "
      ]
    },
    {
      "cell_type": "code",
      "execution_count": 8,
      "id": "fdd4eb07",
      "metadata": {
        "id": "fdd4eb07",
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "outputId": "3a2f3442-5c1f-41da-8165-44838ed1a239"
      },
      "outputs": [
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "Requirement already satisfied: stable-baselines3[extra] in /usr/local/lib/python3.7/dist-packages (1.4.0)\n",
            "Requirement already satisfied: pyvirtualdisplay in /usr/local/lib/python3.7/dist-packages (2.2)\n",
            "Requirement already satisfied: gym[atari] in /usr/local/lib/python3.7/dist-packages (0.17.3)\n",
            "Requirement already satisfied: pyglet in /usr/local/lib/python3.7/dist-packages (1.5.0)\n",
            "Requirement already satisfied: EasyProcess in /usr/local/lib/python3.7/dist-packages (from pyvirtualdisplay) (1.1)\n",
            "Requirement already satisfied: future in /usr/local/lib/python3.7/dist-packages (from pyglet) (0.16.0)\n",
            "Requirement already satisfied: cloudpickle<1.7.0,>=1.2.0 in /usr/local/lib/python3.7/dist-packages (from gym[atari]) (1.3.0)\n",
            "Requirement already satisfied: numpy>=1.10.4 in /usr/local/lib/python3.7/dist-packages (from gym[atari]) (1.19.5)\n",
            "Requirement already satisfied: scipy in /usr/local/lib/python3.7/dist-packages (from gym[atari]) (1.4.1)\n",
            "Requirement already satisfied: Pillow in /usr/local/lib/python3.7/dist-packages (from gym[atari]) (7.1.2)\n",
            "Requirement already satisfied: opencv-python in /usr/local/lib/python3.7/dist-packages (from gym[atari]) (4.1.2.30)\n",
            "Requirement already satisfied: atari-py~=0.2.0 in /usr/local/lib/python3.7/dist-packages (from gym[atari]) (0.2.6)\n",
            "Requirement already satisfied: six in /usr/local/lib/python3.7/dist-packages (from atari-py~=0.2.0->gym[atari]) (1.15.0)\n",
            "Requirement already satisfied: pandas in /usr/local/lib/python3.7/dist-packages (from stable-baselines3[extra]) (1.3.5)\n",
            "Requirement already satisfied: torch>=1.8.1 in /usr/local/lib/python3.7/dist-packages (from stable-baselines3[extra]) (1.10.0+cu111)\n",
            "Requirement already satisfied: matplotlib in /usr/local/lib/python3.7/dist-packages (from stable-baselines3[extra]) (3.2.2)\n",
            "Requirement already satisfied: psutil in /usr/local/lib/python3.7/dist-packages (from stable-baselines3[extra]) (5.4.8)\n",
            "Requirement already satisfied: tensorboard>=2.2.0 in /usr/local/lib/python3.7/dist-packages (from stable-baselines3[extra]) (2.7.0)\n",
            "Requirement already satisfied: google-auth-oauthlib<0.5,>=0.4.1 in /usr/local/lib/python3.7/dist-packages (from tensorboard>=2.2.0->stable-baselines3[extra]) (0.4.6)\n",
            "Requirement already satisfied: protobuf>=3.6.0 in /usr/local/lib/python3.7/dist-packages (from tensorboard>=2.2.0->stable-baselines3[extra]) (3.17.3)\n",
            "Requirement already satisfied: werkzeug>=0.11.15 in /usr/local/lib/python3.7/dist-packages (from tensorboard>=2.2.0->stable-baselines3[extra]) (1.0.1)\n",
            "Requirement already satisfied: setuptools>=41.0.0 in /usr/local/lib/python3.7/dist-packages (from tensorboard>=2.2.0->stable-baselines3[extra]) (57.4.0)\n",
            "Requirement already satisfied: wheel>=0.26 in /usr/local/lib/python3.7/dist-packages (from tensorboard>=2.2.0->stable-baselines3[extra]) (0.37.1)\n",
            "Requirement already satisfied: requests<3,>=2.21.0 in /usr/local/lib/python3.7/dist-packages (from tensorboard>=2.2.0->stable-baselines3[extra]) (2.23.0)\n",
            "Requirement already satisfied: markdown>=2.6.8 in /usr/local/lib/python3.7/dist-packages (from tensorboard>=2.2.0->stable-baselines3[extra]) (3.3.6)\n",
            "Requirement already satisfied: absl-py>=0.4 in /usr/local/lib/python3.7/dist-packages (from tensorboard>=2.2.0->stable-baselines3[extra]) (1.0.0)\n",
            "Requirement already satisfied: tensorboard-plugin-wit>=1.6.0 in /usr/local/lib/python3.7/dist-packages (from tensorboard>=2.2.0->stable-baselines3[extra]) (1.8.1)\n",
            "Requirement already satisfied: google-auth<3,>=1.6.3 in /usr/local/lib/python3.7/dist-packages (from tensorboard>=2.2.0->stable-baselines3[extra]) (1.35.0)\n",
            "Requirement already satisfied: grpcio>=1.24.3 in /usr/local/lib/python3.7/dist-packages (from tensorboard>=2.2.0->stable-baselines3[extra]) (1.43.0)\n",
            "Requirement already satisfied: tensorboard-data-server<0.7.0,>=0.6.0 in /usr/local/lib/python3.7/dist-packages (from tensorboard>=2.2.0->stable-baselines3[extra]) (0.6.1)\n",
            "Requirement already satisfied: pyasn1-modules>=0.2.1 in /usr/local/lib/python3.7/dist-packages (from google-auth<3,>=1.6.3->tensorboard>=2.2.0->stable-baselines3[extra]) (0.2.8)\n",
            "Requirement already satisfied: cachetools<5.0,>=2.0.0 in /usr/local/lib/python3.7/dist-packages (from google-auth<3,>=1.6.3->tensorboard>=2.2.0->stable-baselines3[extra]) (4.2.4)\n",
            "Requirement already satisfied: rsa<5,>=3.1.4 in /usr/local/lib/python3.7/dist-packages (from google-auth<3,>=1.6.3->tensorboard>=2.2.0->stable-baselines3[extra]) (4.8)\n",
            "Requirement already satisfied: requests-oauthlib>=0.7.0 in /usr/local/lib/python3.7/dist-packages (from google-auth-oauthlib<0.5,>=0.4.1->tensorboard>=2.2.0->stable-baselines3[extra]) (1.3.1)\n",
            "Requirement already satisfied: importlib-metadata>=4.4 in /usr/local/lib/python3.7/dist-packages (from markdown>=2.6.8->tensorboard>=2.2.0->stable-baselines3[extra]) (4.10.1)\n",
            "Requirement already satisfied: typing-extensions>=3.6.4 in /usr/local/lib/python3.7/dist-packages (from importlib-metadata>=4.4->markdown>=2.6.8->tensorboard>=2.2.0->stable-baselines3[extra]) (3.10.0.2)\n",
            "Requirement already satisfied: zipp>=0.5 in /usr/local/lib/python3.7/dist-packages (from importlib-metadata>=4.4->markdown>=2.6.8->tensorboard>=2.2.0->stable-baselines3[extra]) (3.7.0)\n",
            "Requirement already satisfied: pyasn1<0.5.0,>=0.4.6 in /usr/local/lib/python3.7/dist-packages (from pyasn1-modules>=0.2.1->google-auth<3,>=1.6.3->tensorboard>=2.2.0->stable-baselines3[extra]) (0.4.8)\n",
            "Requirement already satisfied: chardet<4,>=3.0.2 in /usr/local/lib/python3.7/dist-packages (from requests<3,>=2.21.0->tensorboard>=2.2.0->stable-baselines3[extra]) (3.0.4)\n",
            "Requirement already satisfied: certifi>=2017.4.17 in /usr/local/lib/python3.7/dist-packages (from requests<3,>=2.21.0->tensorboard>=2.2.0->stable-baselines3[extra]) (2021.10.8)\n",
            "Requirement already satisfied: idna<3,>=2.5 in /usr/local/lib/python3.7/dist-packages (from requests<3,>=2.21.0->tensorboard>=2.2.0->stable-baselines3[extra]) (2.10)\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<3,>=2.21.0->tensorboard>=2.2.0->stable-baselines3[extra]) (1.24.3)\n",
            "Requirement already satisfied: oauthlib>=3.0.0 in /usr/local/lib/python3.7/dist-packages (from requests-oauthlib>=0.7.0->google-auth-oauthlib<0.5,>=0.4.1->tensorboard>=2.2.0->stable-baselines3[extra]) (3.2.0)\n",
            "Requirement already satisfied: pyparsing!=2.0.4,!=2.1.2,!=2.1.6,>=2.0.1 in /usr/local/lib/python3.7/dist-packages (from matplotlib->stable-baselines3[extra]) (3.0.7)\n",
            "Requirement already satisfied: kiwisolver>=1.0.1 in /usr/local/lib/python3.7/dist-packages (from matplotlib->stable-baselines3[extra]) (1.3.2)\n",
            "Requirement already satisfied: cycler>=0.10 in /usr/local/lib/python3.7/dist-packages (from matplotlib->stable-baselines3[extra]) (0.11.0)\n",
            "Requirement already satisfied: python-dateutil>=2.1 in /usr/local/lib/python3.7/dist-packages (from matplotlib->stable-baselines3[extra]) (2.8.2)\n",
            "Requirement already satisfied: pytz>=2017.3 in /usr/local/lib/python3.7/dist-packages (from pandas->stable-baselines3[extra]) (2018.9)\n",
            "Reading package lists... Done\n",
            "Building dependency tree       \n",
            "Reading state information... Done\n",
            "python-opengl is already the newest version (3.1.0+dfsg-1).\n",
            "The following packages were automatically installed and are no longer required:\n",
            "  cuda-command-line-tools-10-0 cuda-command-line-tools-10-1\n",
            "  cuda-command-line-tools-11-0 cuda-compiler-10-0 cuda-compiler-10-1\n",
            "  cuda-compiler-11-0 cuda-cuobjdump-10-0 cuda-cuobjdump-10-1\n",
            "  cuda-cuobjdump-11-0 cuda-cupti-10-0 cuda-cupti-10-1 cuda-cupti-11-0\n",
            "  cuda-cupti-dev-11-0 cuda-documentation-10-0 cuda-documentation-10-1\n",
            "  cuda-documentation-11-0 cuda-documentation-11-1 cuda-gdb-10-0 cuda-gdb-10-1\n",
            "  cuda-gdb-11-0 cuda-gpu-library-advisor-10-0 cuda-gpu-library-advisor-10-1\n",
            "  cuda-libraries-10-0 cuda-libraries-10-1 cuda-libraries-11-0\n",
            "  cuda-memcheck-10-0 cuda-memcheck-10-1 cuda-memcheck-11-0 cuda-nsight-10-0\n",
            "  cuda-nsight-10-1 cuda-nsight-11-0 cuda-nsight-11-1 cuda-nsight-compute-10-0\n",
            "  cuda-nsight-compute-10-1 cuda-nsight-compute-11-0 cuda-nsight-compute-11-1\n",
            "  cuda-nsight-systems-10-1 cuda-nsight-systems-11-0 cuda-nsight-systems-11-1\n",
            "  cuda-nvcc-10-0 cuda-nvcc-10-1 cuda-nvcc-11-0 cuda-nvdisasm-10-0\n",
            "  cuda-nvdisasm-10-1 cuda-nvdisasm-11-0 cuda-nvml-dev-10-0 cuda-nvml-dev-10-1\n",
            "  cuda-nvml-dev-11-0 cuda-nvprof-10-0 cuda-nvprof-10-1 cuda-nvprof-11-0\n",
            "  cuda-nvprune-10-0 cuda-nvprune-10-1 cuda-nvprune-11-0 cuda-nvtx-10-0\n",
            "  cuda-nvtx-10-1 cuda-nvtx-11-0 cuda-nvvp-10-0 cuda-nvvp-10-1 cuda-nvvp-11-0\n",
            "  cuda-nvvp-11-1 cuda-samples-10-0 cuda-samples-10-1 cuda-samples-11-0\n",
            "  cuda-samples-11-1 cuda-sanitizer-11-0 cuda-sanitizer-api-10-1\n",
            "  cuda-toolkit-10-0 cuda-toolkit-10-1 cuda-toolkit-11-0 cuda-toolkit-11-1\n",
            "  cuda-tools-10-0 cuda-tools-10-1 cuda-tools-11-0 cuda-tools-11-1\n",
            "  cuda-visual-tools-10-0 cuda-visual-tools-10-1 cuda-visual-tools-11-0\n",
            "  cuda-visual-tools-11-1 default-jre dkms freeglut3-dev keyboard-configuration\n",
            "  libargon2-0 libcap2 libcryptsetup12 libdevmapper1.02.1 libfontenc1 libidn11\n",
            "  libip4tc0 libjansson4 libnvidia-cfg1-510 libnvidia-common-460\n",
            "  libnvidia-common-510 libnvidia-extra-510 libnvidia-fbc1-510 libnvidia-gl-510\n",
            "  libpam-systemd libpolkit-agent-1-0 libpolkit-backend-1-0\n",
            "  libpolkit-gobject-1-0 libxfont2 libxi-dev libxkbfile1 libxmu-dev\n",
            "  libxmu-headers libxnvctrl0 libxtst6 nsight-compute-2020.2.1\n",
            "  nsight-compute-2022.1.0 nsight-systems-2020.3.2 nsight-systems-2020.3.4\n",
            "  nsight-systems-2021.5.2 nvidia-dkms-510 nvidia-kernel-common-510\n",
            "  nvidia-kernel-source-510 nvidia-modprobe nvidia-settings openjdk-11-jre\n",
            "  policykit-1 policykit-1-gnome python3-xkit screen-resolution-extra systemd\n",
            "  systemd-sysv udev x11-xkb-utils xserver-common xserver-xorg-core-hwe-18.04\n",
            "  xserver-xorg-video-nvidia-510\n",
            "Use 'apt autoremove' to remove them.\n",
            "0 upgraded, 0 newly installed, 0 to remove and 39 not upgraded.\n",
            "Requirement already satisfied: atari-py in /usr/local/lib/python3.7/dist-packages (0.2.6)\n",
            "Requirement already satisfied: numpy in /usr/local/lib/python3.7/dist-packages (from atari-py) (1.19.5)\n",
            "Requirement already satisfied: six in /usr/local/lib/python3.7/dist-packages (from atari-py) (1.15.0)\n",
            "Reading package lists... Done\n",
            "Building dependency tree       \n",
            "Reading state information... Done\n",
            "unrar is already the newest version (1:5.5.8-1).\n",
            "The following packages were automatically installed and are no longer required:\n",
            "  cuda-command-line-tools-10-0 cuda-command-line-tools-10-1\n",
            "  cuda-command-line-tools-11-0 cuda-compiler-10-0 cuda-compiler-10-1\n",
            "  cuda-compiler-11-0 cuda-cuobjdump-10-0 cuda-cuobjdump-10-1\n",
            "  cuda-cuobjdump-11-0 cuda-cupti-10-0 cuda-cupti-10-1 cuda-cupti-11-0\n",
            "  cuda-cupti-dev-11-0 cuda-documentation-10-0 cuda-documentation-10-1\n",
            "  cuda-documentation-11-0 cuda-documentation-11-1 cuda-gdb-10-0 cuda-gdb-10-1\n",
            "  cuda-gdb-11-0 cuda-gpu-library-advisor-10-0 cuda-gpu-library-advisor-10-1\n",
            "  cuda-libraries-10-0 cuda-libraries-10-1 cuda-libraries-11-0\n",
            "  cuda-memcheck-10-0 cuda-memcheck-10-1 cuda-memcheck-11-0 cuda-nsight-10-0\n",
            "  cuda-nsight-10-1 cuda-nsight-11-0 cuda-nsight-11-1 cuda-nsight-compute-10-0\n",
            "  cuda-nsight-compute-10-1 cuda-nsight-compute-11-0 cuda-nsight-compute-11-1\n",
            "  cuda-nsight-systems-10-1 cuda-nsight-systems-11-0 cuda-nsight-systems-11-1\n",
            "  cuda-nvcc-10-0 cuda-nvcc-10-1 cuda-nvcc-11-0 cuda-nvdisasm-10-0\n",
            "  cuda-nvdisasm-10-1 cuda-nvdisasm-11-0 cuda-nvml-dev-10-0 cuda-nvml-dev-10-1\n",
            "  cuda-nvml-dev-11-0 cuda-nvprof-10-0 cuda-nvprof-10-1 cuda-nvprof-11-0\n",
            "  cuda-nvprune-10-0 cuda-nvprune-10-1 cuda-nvprune-11-0 cuda-nvtx-10-0\n",
            "  cuda-nvtx-10-1 cuda-nvtx-11-0 cuda-nvvp-10-0 cuda-nvvp-10-1 cuda-nvvp-11-0\n",
            "  cuda-nvvp-11-1 cuda-samples-10-0 cuda-samples-10-1 cuda-samples-11-0\n",
            "  cuda-samples-11-1 cuda-sanitizer-11-0 cuda-sanitizer-api-10-1\n",
            "  cuda-toolkit-10-0 cuda-toolkit-10-1 cuda-toolkit-11-0 cuda-toolkit-11-1\n",
            "  cuda-tools-10-0 cuda-tools-10-1 cuda-tools-11-0 cuda-tools-11-1\n",
            "  cuda-visual-tools-10-0 cuda-visual-tools-10-1 cuda-visual-tools-11-0\n",
            "  cuda-visual-tools-11-1 default-jre dkms freeglut3-dev keyboard-configuration\n",
            "  libargon2-0 libcap2 libcryptsetup12 libdevmapper1.02.1 libfontenc1 libidn11\n",
            "  libip4tc0 libjansson4 libnvidia-cfg1-510 libnvidia-common-460\n",
            "  libnvidia-common-510 libnvidia-extra-510 libnvidia-fbc1-510 libnvidia-gl-510\n",
            "  libpam-systemd libpolkit-agent-1-0 libpolkit-backend-1-0\n",
            "  libpolkit-gobject-1-0 libxfont2 libxi-dev libxkbfile1 libxmu-dev\n",
            "  libxmu-headers libxnvctrl0 libxtst6 nsight-compute-2020.2.1\n",
            "  nsight-compute-2022.1.0 nsight-systems-2020.3.2 nsight-systems-2020.3.4\n",
            "  nsight-systems-2021.5.2 nvidia-dkms-510 nvidia-kernel-common-510\n",
            "  nvidia-kernel-source-510 nvidia-modprobe nvidia-settings openjdk-11-jre\n",
            "  policykit-1 policykit-1-gnome python3-xkit screen-resolution-extra systemd\n",
            "  systemd-sysv udev x11-xkb-utils xserver-common xserver-xorg-core-hwe-18.04\n",
            "  xserver-xorg-video-nvidia-510\n",
            "Use 'apt autoremove' to remove them.\n",
            "0 upgraded, 0 newly installed, 0 to remove and 39 not upgraded.\n",
            "--2022-02-08 18:52:17--  http://www.atarimania.com/roms/Roms.rar\n",
            "Resolving www.atarimania.com (www.atarimania.com)... 195.154.81.199\n",
            "Connecting to www.atarimania.com (www.atarimania.com)|195.154.81.199|:80... connected.\n",
            "HTTP request sent, awaiting response... 200 OK\n",
            "Length: 11128004 (11M) [application/x-rar-compressed]\n",
            "Saving to: ‘Roms.rar.2’\n",
            "\n",
            "Roms.rar.2          100%[===================>]  10.61M   626KB/s    in 18s     \n",
            "\n",
            "2022-02-08 18:52:35 (604 KB/s) - ‘Roms.rar.2’ saved [11128004/11128004]\n",
            "\n",
            "\n",
            "UNRAR 5.50 freeware      Copyright (c) 1993-2017 Alexander Roshal\n",
            "\n",
            "\n",
            "Extracting from Roms.rar\n",
            "\n",
            "\n",
            "Would you like to replace the existing file HC ROMS.zip\n",
            "11826711 bytes, modified on 2019-12-22 11:24\n",
            "with a new one\n",
            "11826711 bytes, modified on 2019-12-22 11:24\n",
            "\n",
            "[Y]es, [N]o, [A]ll, n[E]ver, [R]ename, [Q]uit \n",
            "User break\n",
            "\n",
            "User break\n",
            "Archive:  ROMS.zip\n",
            "replace ROMS/128 in 1 Game Select ROM (128 in 1) (Unknown) ~.bin? [y]es, [n]o, [A]ll, [N]one, [r]ename: /usr/bin/python3: No module named atari_py.import_roms\n"
          ]
        }
      ],
      "source": [
        "\n",
        "!pip install stable-baselines3[extra] pyvirtualdisplay gym[atari] pyglet \n",
        "!apt-get install python-opengl -y\n",
        "!pip install atari-py\n",
        "!apt-get install unrar\n",
        "\n",
        "!wget http://www.atarimania.com/roms/Roms.rar \n",
        "!unrar x Roms.rar \n",
        "!unzip ROMS.zip\n",
        "!python -m atari_py.import_roms ROMS"
      ]
    },
    {
      "cell_type": "markdown",
      "id": "ba0b763c",
      "metadata": {
        "id": "ba0b763c"
      },
      "source": [
        "#### Shoot N Dodge\n",
        "The game is provided as a python package and can be run as an **OpenAI-Gym** environment. To do so, you will have to go to the setup folder and install the package on your python virtual environment."
      ]
    },
    {
      "cell_type": "code",
      "execution_count": 9,
      "id": "5a59e170",
      "metadata": {
        "id": "5a59e170",
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "outputId": "e37bd12e-2bca-4378-935c-c00bf52fa895"
      },
      "outputs": [
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "Obtaining file:///content/gym-shootndodge\n",
            "Requirement already satisfied: gym in /usr/local/lib/python3.7/dist-packages (from gym-shootndodge==1.2) (0.17.3)\n",
            "Requirement already satisfied: opencv-python in /usr/local/lib/python3.7/dist-packages (from gym-shootndodge==1.2) (4.1.2.30)\n",
            "Requirement already satisfied: pillow in /usr/local/lib/python3.7/dist-packages (from gym-shootndodge==1.2) (7.1.2)\n",
            "Requirement already satisfied: numpy in /usr/local/lib/python3.7/dist-packages (from gym-shootndodge==1.2) (1.19.5)\n",
            "Requirement already satisfied: cloudpickle<1.7.0,>=1.2.0 in /usr/local/lib/python3.7/dist-packages (from gym->gym-shootndodge==1.2) (1.3.0)\n",
            "Requirement already satisfied: scipy in /usr/local/lib/python3.7/dist-packages (from gym->gym-shootndodge==1.2) (1.4.1)\n",
            "Requirement already satisfied: pyglet<=1.5.0,>=1.4.0 in /usr/local/lib/python3.7/dist-packages (from gym->gym-shootndodge==1.2) (1.5.0)\n",
            "Requirement already satisfied: future in /usr/local/lib/python3.7/dist-packages (from pyglet<=1.5.0,>=1.4.0->gym->gym-shootndodge==1.2) (0.16.0)\n",
            "Installing collected packages: gym-shootndodge\n",
            "  Attempting uninstall: gym-shootndodge\n",
            "    Found existing installation: gym-shootndodge 1.2\n",
            "    Can't uninstall 'gym-shootndodge'. No files were found to uninstall.\n",
            "  Running setup.py develop for gym-shootndodge\n",
            "Successfully installed gym-shootndodge-1.2\n"
          ]
        }
      ],
      "source": [
        "# NOTE:\n",
        "# .... You should be at the root directory of the repository in order to run this\n",
        "\n",
        "!pip install -e gym-shootndodge/.\n",
        "# !pip install -e /content/drive/MyDrive/AI/5/Q2/gym-shootndodge/.\n",
        "# ------------------------------------------------------------------------------------\n",
        "\n",
        "# Install from git address (useful for Google Colab)\n",
        "\n",
        "# !pip install -e git+https://github.com/sut-ai/F2021-P5/tree/master/practical/questions/gym-shootndodge/."
      ]
    },
    {
      "cell_type": "markdown",
      "id": "dff4d113",
      "metadata": {
        "id": "dff4d113"
      },
      "source": [
        "## Seeing the game\n",
        "\n",
        "Without further ado, we shall start see how your game works. To do so, run the code below:"
      ]
    },
    {
      "cell_type": "code",
      "execution_count": 13,
      "id": "8ef502e0",
      "metadata": {
        "id": "8ef502e0",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 269
        },
        "outputId": "4b33d557-ecb9-4ccb-b693-bce375392e3d"
      },
      "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"
          }
        }
      ],
      "source": [
        "env = gym.make(ENV_NAME)\n",
        "obs = env.reset()\n",
        "plt.imshow(env.render(mode='rgb_array'), aspect='auto')\n",
        "plt.show()"
      ]
    },
    {
      "cell_type": "markdown",
      "id": "dcd300e9",
      "metadata": {
        "id": "dcd300e9"
      },
      "source": [
        "Run the game in multiple frames to see what it looks like:"
      ]
    },
    {
      "cell_type": "code",
      "execution_count": 14,
      "id": "f8dcb334",
      "metadata": {
        "id": "f8dcb334",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 269
        },
        "outputId": "b0e306f4-d696-4328-9c2c-907ade64ce37"
      },
      "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"
          }
        }
      ],
      "source": [
        "env = gym.make(ENV_NAME)\n",
        "env.seed(2022)\n",
        "obs = env.reset()\n",
        "\n",
        "a = env.action_space.sample()\n",
        "for i in range(5000):\n",
        "    if (i + 1) % 40 == 0:\n",
        "        a = env.action_space.sample()\n",
        "    observation, reward, done, info = env.step(a)\n",
        "    if (i + 1) % 5 == 0:\n",
        "        clear_output(True)\n",
        "        screen = env.render(mode='rgb_array')\n",
        "        plt.imshow(screen, aspect='auto')\n",
        "        plt.show()\n",
        "    if done:\n",
        "        break"
      ]
    },
    {
      "cell_type": "code",
      "source": [
        "# extra\n",
        "print(env.observation_space.shape)\n",
        "print(np.shape(obs))\n",
        "import cv2\n",
        "\n",
        "img = screen[55:195,10:150] # cropping\n",
        "img = cv2.cvtColor(img, cv2.COLOR_BGR2GRAY) # Gray scale\n",
        "img = cv2.resize(img, (84, 84), interpolation = cv2.INTER_AREA)\n",
        "img = np.float32(img)/255\n",
        "plt.imshow(img*255, cmap='gray', aspect='auto')\n",
        "img = np.expand_dims(img, axis=2)\n",
        "print(np.shape(img))\n",
        "\n",
        "def observation(self, img):\n",
        "  \"\"\"what happens to each observation\"\"\"\n",
        "  img = img[55:195,10:150] # cropping\n",
        "  img = cv2.cvtColor(img, cv2.COLOR_BGR2GRAY) # Gray scale\n",
        "  img = cv2.resize(img, (self.img_size[0], self.img_size[1]), interpolation = cv2.INTER_AREA) # resizing\n",
        "  img = np.float32(img)/255\n",
        "  img = np.expand_dims(img, axis=2)\n",
        "        # Here's what you need to do:\n",
        "        #  * crop image, remove irrelevant parts\n",
        "        #  * resize image to self.img_size\n",
        "        #     (use imresize from any library you want,\n",
        "        #      e.g. opencv, skimage, PIL, keras)\n",
        "        #  * cast image to grayscale (in case of breakout)\n",
        "        #  * convert image pixels to (0,1) range, float32 type\n",
        "\n",
        "        # TODO: complete observation descaling\n",
        "  processed_img = img\n",
        "  return processed_img\n",
        "\n",
        "def observation(self, img):\n",
        "      \"\"\"what happens to each observation\"\"\"\n",
        "      img = img[55:195,10:150] # cropping\n",
        "      img = cv2.cvtColor(img, cv2.COLOR_BGR2GRAY) # Gray scale\n",
        "      img = cv2.resize(img, (self.img_size[0], self.img_size[1]), interpolation = cv2.INTER_AREA) # resizing\n",
        "      img = np.float32(img)/255\n",
        "      img = np.expand_dims(img, axis=2)\n",
        "      processed_img = img\n",
        "      return processed_img"
      ],
      "metadata": {
        "id": "n3jre0A4fw_H",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 320
        },
        "outputId": "56458b9f-d43c-4385-8283-2995e2cf52b8"
      },
      "id": "n3jre0A4fw_H",
      "execution_count": 15,
      "outputs": [
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "(210, 160, 3)\n",
            "(210, 160, 3)\n",
            "(84, 84, 1)\n"
          ]
        },
        {
          "output_type": "display_data",
          "data": {
            "image/png": "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\n",
            "text/plain": [
              "<Figure size 432x288 with 1 Axes>"
            ]
          },
          "metadata": {
            "needs_background": "light"
          }
        }
      ]
    },
    {
      "cell_type": "markdown",
      "id": "c83cb315",
      "metadata": {
        "id": "c83cb315"
      },
      "source": [
        "Feel free to hard code any policies before moving on to the next steps."
      ]
    },
    {
      "cell_type": "markdown",
      "id": "320da2f7",
      "metadata": {
        "id": "320da2f7"
      },
      "source": [
        "## Environment Wrappers\n",
        "\n",
        "In this section, we will do some preprocessings on the environment. Basically our work consists of wrapping our environment around three wrappers. You should implement each of these wrappers in the `src/wrappers.py` file provided to you in the assignment folder. \n",
        "* Reducing image input size: The Atari game given as input has high resolution which is definitely not needed for the model and it can learn efficiently without all of the additional data. These additional pixels create overhead. We can thus save a lot of time by preprocessing game image, including:\n",
        "    - Resizing to a smaller shape, (84 x 84 recommended for the Breakout and 50 x 50 recommended for ShootNDodge)\n",
        "    - Converting to grayscale (Not recommended for ShootNDodge)\n",
        "    - Cropping irrelevant image parts (top, bottom and edges)\n",
        "    - Also please keep one dimension for channel even if the image is grayscale so that final shape would be (H x W x C)\n",
        "* Reward clipping: There might be the case that many of the RL problems have unstable rewards. In case the rewards are not continuous with respect to the state space, then the local optimization methods used for learning the network will fall short. To address this issue we create a wrapper around our original environment - Note that this is not a necessity and in case you feel comfortable with the rewards given in the original environment, then your wrapper should simply return the previous values as if nothing has heppened! Feel free to accumulate more knowledge about reward clipping where we reduce the range of the rewards to [-1, 1]. In case you observe increasing gradient norms in the learning process, then it would be good to consider this method.\n",
        "* Remove Blinking: Since we resize the image, we might come accross blinking whereas in one frame a certain thing is visible, but at the split-frame afterward, it fades and then returns again. To remove this issue we introduce a `MaxAndSkipEnv` in which we skip one frame and each frame's intensity will be equal to the maximum of the previous frame and itself. A preliminary code is implemented in the `wrappers.py` file.\n",
        "* Episodic life: In case of *Breakout* losing takes 5 lives. To increase the chance of the model learning, we may end the game whenever one single life is lost.\n",
        "* Starting the game: In case of *Breakout* one should take a particular action to start the game. For example, they might need to press the `space` key in order for the game to start rolling. A wrapper is implemented to forcefully enter the game start action in the beginning. Indeed, in case this wrapper is not implemented, the agent can learn to take that action to gain higher reward; however, this will make the training time even longer."
      ]
    },
    {
      "cell_type": "markdown",
      "id": "2e2c2ddc",
      "metadata": {
        "id": "2e2c2ddc"
      },
      "source": [
        "After implementing or checking the wrappers in `src/wrappers.py` we can culminate all of the wrappers in the following class. A baseline is implemented for this wrapper but feel free to change it at will."
      ]
    },
    {
      "cell_type": "code",
      "execution_count": 10,
      "id": "ae0e3340",
      "metadata": {
        "id": "ae0e3340"
      },
      "outputs": [],
      "source": [
        "from src.wrappers import PreprocessAtariObs, MaxAndSkipEnv, EpisodicLifeEnv, FireResetEnv, ClipRewardEnv, PreprocessAtariObs\n",
        "\n",
        "# def PrimaryAtariWrap(env, game, clip_rewards=True):\n",
        "def PrimaryAtariWrap(env, clip_rewards=True): \n",
        "    name = env.unwrapped.spec.id\n",
        "    # print(name)\n",
        "    # This wrapper holds the same action for <skip> frames and outputs\n",
        "    # the maximal pixel value of 2 last frames (to handle blinking\n",
        "    # in some envs)\n",
        "    env = MaxAndSkipEnv(env, skip=4)\n",
        "\n",
        "    # This wrapper sends done=True when each life is lost\n",
        "    # (not all the 5 lives that are givern by the game rules).\n",
        "    # It should make easier for the agent to understand that losing is bad.\n",
        "    if name == \"BreakoutNoFrameskip-v4\":\n",
        "        EpisodicLifeEnv(env)\n",
        "\n",
        "    # This wrapper laucnhes the ball when an episode starts.\n",
        "    # Without it the agent has to learn this action, too.\n",
        "    # Actually it can but learning would take longer.\n",
        "    if name == \"BreakoutNoFrameskip-v4\":\n",
        "        env = FireResetEnv(env)\n",
        "\n",
        "    # This wrapper squeezes the rewards to range [-1, 1]\n",
        "    if clip_rewards:\n",
        "        env = ClipRewardEnv(env)\n",
        "\n",
        "    # This wrapper is yours :)\n",
        "    env = PreprocessAtariObs(env, True)\n",
        "    # print(env)\n",
        "    return env"
      ]
    },
    {
      "cell_type": "markdown",
      "id": "f285b7a6",
      "metadata": {
        "id": "f285b7a6"
      },
      "source": [
        "\n",
        "## Frame Buffer\n",
        "\n",
        "It is reasonable to assume that only one frame of the game does not hold information about the game dynamics. Hence, each observation should in fact contain multiple frames. There is a rule of thumb in these Atari games to consider buffering the last 4 frames of the game. In that case, each observation will be of form (H x W x C*F) where F represents the number of frames saved.\n",
        "\n",
        "To implement this, you may check the `src/framebuffer.py` file or change/implement your own frame buffer. With all of that done, we can run the following function to have an environment that has rich observations and reasonable rewards that provide the features necessary to train our DQN model.\n"
      ]
    },
    {
      "cell_type": "code",
      "execution_count": 11,
      "id": "78f9d5fa",
      "metadata": {
        "id": "78f9d5fa",
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "outputId": "c94985d7-5f9f-4ef7-8fcc-4faa41a03001"
      },
      "outputs": [
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "4\n",
            "(84, 84, 4)\n",
            "Box(0.0, 1.0, (84, 84, 4), float32)\n"
          ]
        }
      ],
      "source": [
        "from src.framebuffer import FrameBuffer\n",
        "\n",
        "def make_env(clip_rewards=False, seed=None):\n",
        "    env = gym.make(ENV_NAME)  # create raw env\n",
        "    if seed is not None:\n",
        "        env.seed(seed)\n",
        "    env = PrimaryAtariWrap(env, clip_rewards)\n",
        "    env = FrameBuffer(env, n_frames=4)\n",
        "    return env\n",
        "\n",
        "env = make_env()\n",
        "env.reset()\n",
        "n_actions = env.action_space.n\n",
        "print(n_actions)\n",
        "state_shape = env.observation_space.shape\n",
        "print(state_shape)\n",
        "\n",
        "obs = env.observation_space\n",
        "print(obs)"
      ]
    },
    {
      "cell_type": "markdown",
      "id": "4d24af5e",
      "metadata": {
        "id": "4d24af5e"
      },
      "source": [
        "### The DQN architecture\n",
        "\n",
        "We now need to build a neural network that can map images to state q-values. This network will be called on every agent's step so it better not be resnet-152 unless you have an array of GPUs. Instead, you can use strided convolutions with a small number of features to save time and memory.\n",
        "\n",
        "You can build any architecture you want, but for reference, here's something that will more or less work:\n",
        "\n",
        "<img src=\"figs/dqn_arch.png\" width=400 height=400/>\n",
        "\n",
        "**IMPORTANT NOTE!** The pytorch library takes inputs as tensors of size `[batch_size, channel_size, h, w]`; however, for visualization reasons, we have used numpy arrays of size `[h, w, channel_size]` up until now! Make sure to do the needed transformation before giving it convolutional layers."
      ]
    },
    {
      "cell_type": "code",
      "execution_count": 12,
      "id": "4722d940",
      "metadata": {
        "id": "4722d940"
      },
      "outputs": [],
      "source": [
        "class DQNAgent(nn.Module):\n",
        "    def __init__(self, env, epsilon=0):\n",
        "\n",
        "        super().__init__()\n",
        "        self.epsilon = epsilon\n",
        "        self.state_shape = (-1, \n",
        "                            env.observation_space.shape[2],\n",
        "                            env.observation_space.shape[0],\n",
        "                            env.observation_space.shape[1])\n",
        "        self.n_actions = env.action_space.n\n",
        "\n",
        "        # TODO\n",
        "        # Define your network body here. Please make sure agent is fully contained here\n",
        "        # nn.Flatten() can be useful\n",
        "        #<YOUR CODE>\n",
        "        #######################################################\n",
        "        self.layers = nn.Sequential(\n",
        "            nn.Conv2d(4, 16, 3, stride=2),   # B 16  41  41  \n",
        "            nn.ReLU(),\n",
        "            nn.Conv2d(16, 32, 3, stride=2),  # B 32  20  20  \n",
        "            nn.ReLU(),\n",
        "            nn.Conv2d(32, 64, 3, stride=2),  # B 64  9  9\n",
        "            nn.ReLU(),\n",
        "        )                \n",
        "        self.fc = nn.Sequential(\n",
        "            nn.Flatten(),                    # B 2048\n",
        "            nn.Linear(9*9*64, 256),         # B 256\n",
        "            nn.ReLU(),\n",
        "            nn.Linear(256, self.n_actions)\n",
        "        )\n",
        "        #######################################################\n",
        "\n",
        "    def forward(self, state_t):\n",
        "        \"\"\"\n",
        "        takes agent's observation (tensor), returns qvalues (tensor)\n",
        "        :param state_t: a batch of 4-frame buffers, shape = [batch_size, h, w, 4 * c]\n",
        "        \"\"\"\n",
        "        # TODO\n",
        "        #######################################################\n",
        "        state_t = torch.reshape(state_t, self.state_shape)#.to(device)\n",
        "        state_t = self.layers(state_t)#.to(device)\n",
        "        qvalues = self.fc(state_t)#.to(device)\n",
        "        #######################################################\n",
        "        # Use your network to compute qvalues for given state\n",
        "        # qvalues = None\n",
        "\n",
        "        assert qvalues.requires_grad, \"qvalues must be a torch tensor with grad\"\n",
        "        assert len(qvalues.shape) == 2 and qvalues.shape[0] == state_t.shape[0] and qvalues.shape[1] == self.n_actions\n",
        "\n",
        "        return qvalues\n",
        "\n",
        "    def get_qvalues(self, states):\n",
        "        \"\"\"\n",
        "        like forward, but works on numpy arrays, not tensors\n",
        "        \"\"\"\n",
        "        model_device = next(self.parameters()).device\n",
        "        states = torch.tensor(states, device=model_device, dtype=torch.float)\n",
        "        qvalues = self.forward(states)\n",
        "        return qvalues.data.cpu().numpy()\n",
        "        \n",
        "\n",
        "    def sample_actions(self, qvalues, epsilon):\n",
        "        \"\"\"pick actions given qvalues. Uses epsilon-greedy exploration strategy. \"\"\"\n",
        "        epsilon = self.epsilon\n",
        "        batch_size, n_actions = qvalues.shape\n",
        "\n",
        "        random_actions = np.random.choice(n_actions, size=batch_size)\n",
        "        best_actions = qvalues.argmax(axis=-1)\n",
        "\n",
        "        should_explore = np.random.choice(\n",
        "            [0, 1], batch_size, p=[1-epsilon, epsilon])\n",
        "        return np.where(should_explore, random_actions, best_actions)"
      ]
    },
    {
      "cell_type": "markdown",
      "id": "e22c3d5e",
      "metadata": {
        "id": "e22c3d5e"
      },
      "source": [
        "Run the agent to see if it encounters any errors:"
      ]
    },
    {
      "cell_type": "code",
      "execution_count": 13,
      "id": "436ac702",
      "metadata": {
        "id": "436ac702"
      },
      "outputs": [],
      "source": [
        "def evaluate(env, agent, n_games=1, greedy=False, t_max=10000, visualize=False):\n",
        "    \"\"\" Plays n_games full games. If greedy, picks actions as argmax(qvalues). Returns mean reward. \"\"\"\n",
        "    rewards = []\n",
        "    for g in range(n_games):\n",
        "        s = env.reset()\n",
        "        total_reward = 0\n",
        "        for t in range(t_max):\n",
        "            qvalues = agent.get_qvalues(torch.tensor([s], dtype=torch.float32))\n",
        "            a = agent.sample_actions(qvalues, epsilon=greedy)[0]\n",
        "            next_s, r, done, _ = env.step(a)\n",
        "            total_reward += r\n",
        "            s = next_s\n",
        "            if done:\n",
        "                break\n",
        "        print(f'reward of game {g+1} is {total_reward}')\n",
        "        rewards.append(total_reward)\n",
        "    # TODO: your code here!\n",
        "    return np.mean(rewards)"
      ]
    },
    {
      "cell_type": "code",
      "execution_count": 14,
      "id": "e1d5a291",
      "metadata": {
        "id": "e1d5a291",
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "outputId": "639cec93-3169-48d2-8bc8-34cc0e11ed05"
      },
      "outputs": [
        {
          "output_type": "stream",
          "name": "stderr",
          "text": [
            "/usr/local/lib/python3.7/dist-packages/ipykernel_launcher.py:8: UserWarning: Creating a tensor from a list of numpy.ndarrays is extremely slow. Please consider converting the list to a single numpy.ndarray with numpy.array() before converting to a tensor. (Triggered internally at  ../torch/csrc/utils/tensor_new.cpp:201.)\n",
            "  \n",
            "/usr/local/lib/python3.7/dist-packages/ipykernel_launcher.py:57: UserWarning: To copy construct from a tensor, it is recommended to use sourceTensor.clone().detach() or sourceTensor.clone().detach().requires_grad_(True), rather than torch.tensor(sourceTensor).\n"
          ]
        },
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "reward of game 1 is 0.0\n",
            "reward of game 2 is 0.0\n",
            "reward of game 3 is 0.0\n",
            "reward of game 4 is 0.0\n",
            "reward of game 5 is 0.0\n",
            "reward of game 6 is 0.0\n",
            "reward of game 7 is 0.0\n",
            "reward of game 8 is 0.0\n",
            "reward of game 9 is 0.0\n",
            "reward of game 10 is 0.0\n",
            "reward of game 11 is 0.0\n",
            "reward of game 12 is 0.0\n",
            "reward of game 13 is 0.0\n",
            "reward of game 14 is 0.0\n",
            "reward of game 15 is 0.0\n"
          ]
        },
        {
          "output_type": "execute_result",
          "data": {
            "text/plain": [
              "0.0"
            ]
          },
          "metadata": {},
          "execution_count": 14
        }
      ],
      "source": [
        "# Run the code below to see if the network can run a policy\n",
        "agent = DQNAgent(env)\n",
        "evaluate(env, agent, n_games=15, t_max=10)"
      ]
    },
    {
      "cell_type": "markdown",
      "id": "5f7a38fc",
      "metadata": {
        "id": "5f7a38fc"
      },
      "source": [
        "### Experience Replay\n",
        "\n",
        "The common practice in DQN literature consists of using an experience replay buffer in which tuples of form $(s, a, r, s')$ are stored. In each epoch, a random batch of this buffer will be given to the network for training. \n",
        "\n",
        "In this section, we will implement this buffer; the interface is fairly simple:\n",
        "* `exp_replay.add(obs, act, rw, next_obs, done)` - saves (s,a,r,s',done) tuple into the buffer\n",
        "* `exp_replay.sample(batch_size)` - returns observations, actions, rewards, next_observations and is_done for batch_size random samples.\n",
        "* `len(exp_replay)` - returns number of elements stored in replay buffer."
      ]
    },
    {
      "cell_type": "code",
      "execution_count": 15,
      "id": "c6bf37bd",
      "metadata": {
        "id": "c6bf37bd"
      },
      "outputs": [],
      "source": [
        "# This code is shamelessly stolen from\n",
        "# https://github.com/openai/baselines/blob/master/baselines/deepq/replay_buffer.py\n",
        "import numpy as np\n",
        "import random\n",
        "\n",
        "\n",
        "class ReplayBuffer:\n",
        "    def __init__(self, size):\n",
        "        \"\"\"Create Replay buffer.\n",
        "        Parameters\n",
        "        ----------\n",
        "        size: int\n",
        "            Max number of transitions to store in the buffer. When the buffer\n",
        "            overflows the old memories are dropped.\n",
        "        \"\"\"\n",
        "        self._storage = []\n",
        "        self._maxsize = size\n",
        "        self._next_idx = 0\n",
        "\n",
        "    def __len__(self):\n",
        "        return len(self._storage)\n",
        "\n",
        "    def add(self, obs_t, action, reward, obs_tp1, done):\n",
        "        # fill the data cyclically, if the list is not yet complete append it\n",
        "        data = (obs_t, action, reward, obs_tp1, done)\n",
        "\n",
        "        if self._next_idx >= len(self._storage):\n",
        "            self._storage.append(data)\n",
        "        else:\n",
        "            self._storage[self._next_idx] = data\n",
        "        self._next_idx = (self._next_idx + 1) % self._maxsize\n",
        "\n",
        "    def _encode_sample(self, idxes):\n",
        "        obses_t, actions, rewards, obses_tp1, dones = [], [], [], [], []\n",
        "        for i in idxes:\n",
        "            data = self._storage[i]\n",
        "            obs_t, action, reward, obs_tp1, done = data\n",
        "            obses_t.append(np.array(obs_t, copy=False))\n",
        "            actions.append(np.array(action, copy=False))\n",
        "            rewards.append(reward)\n",
        "            obses_tp1.append(np.array(obs_tp1, copy=False))\n",
        "            dones.append(done)\n",
        "        return (\n",
        "            np.array(obses_t),\n",
        "            np.array(actions),\n",
        "            np.array(rewards),\n",
        "            np.array(obses_tp1),\n",
        "            np.array(dones)\n",
        "        )\n",
        "\n",
        "    def sample(self, batch_size):\n",
        "        \"\"\"Sample a batch of experiences.\n",
        "        Parameters\n",
        "        ----------\n",
        "        batch_size: int\n",
        "            How many transitions to sample.\n",
        "        Returns\n",
        "        -------\n",
        "        obs_batch: np.array\n",
        "            batch of observations\n",
        "        act_batch: np.array\n",
        "            batch of actions executed given obs_batch\n",
        "        rew_batch: np.array\n",
        "            rewards received as results of executing act_batch\n",
        "        next_obs_batch: np.array\n",
        "            next set of observations seen after executing act_batch\n",
        "        done_mask: np.array\n",
        "            done_mask[i] = 1 if executing act_batch[i] resulted in\n",
        "            the end of an episode and 0 otherwise.\n",
        "        \"\"\"\n",
        "        idxes = [\n",
        "            random.randint(0, len(self._storage) - 1)\n",
        "            for _ in range(batch_size)\n",
        "        ]\n",
        "        return self._encode_sample(idxes)"
      ]
    },
    {
      "cell_type": "markdown",
      "id": "1d7262eb",
      "metadata": {
        "id": "1d7262eb"
      },
      "source": [
        "Implement the function `play_and_record` in which we play the game for a number of steps and add the steps into an experience replay buffer."
      ]
    },
    {
      "cell_type": "code",
      "execution_count": 16,
      "id": "0cae0ae1",
      "metadata": {
        "id": "0cae0ae1"
      },
      "outputs": [],
      "source": [
        "def play_and_record(initial_state, agent, env, exp_replay, n_steps=1):\n",
        "    \"\"\"\n",
        "    Play the game for exactly n steps, record every (s,a,r,s', done) to replay buffer. \n",
        "    Whenever game ends, add record with done=True and reset the game.\n",
        "    It is guaranteed that env has done=False when passed to this function.\n",
        "\n",
        "    PLEASE DO NOT RESET ENV UNLESS IT IS \"DONE\"\n",
        "\n",
        "    :returns: return sum of rewards over time and the state in which the env stays\n",
        "    \"\"\"\n",
        "    s = initial_state\n",
        "    sum_rewards = 0\n",
        "\n",
        "    # Play the game for n_steps as per instructions above\n",
        "    for i in range(n_steps):\n",
        "        # TODO: implement this section\n",
        "        ##################################################################\n",
        "        qvalues = agent.get_qvalues(torch.tensor([s], dtype=torch.float32))\n",
        "        a = agent.sample_actions(qvalues, epsilon=1)[0]\n",
        "        next_s, r, done, _ = env.step(a)  \n",
        "        exp_replay.add(s, a, r, next_s, False)\n",
        "        if done == True:\n",
        "          s = env.reset()\n",
        "        else:\n",
        "          s = next_s\n",
        "        sum_rewards += r\n",
        "        ##################################################################\n",
        "    return sum_rewards, s"
      ]
    },
    {
      "cell_type": "markdown",
      "id": "20af7e57",
      "metadata": {
        "id": "20af7e57"
      },
      "source": [
        "Run some simple sanity checks."
      ]
    },
    {
      "cell_type": "code",
      "execution_count": 17,
      "id": "88cafe9f",
      "metadata": {
        "id": "88cafe9f",
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "outputId": "ee33a8df-c6b7-41ee-8735-9799ae7c4cbf"
      },
      "outputs": [
        {
          "output_type": "stream",
          "name": "stderr",
          "text": [
            "/usr/local/lib/python3.7/dist-packages/ipykernel_launcher.py:57: UserWarning: To copy construct from a tensor, it is recommended to use sourceTensor.clone().detach() or sourceTensor.clone().detach().requires_grad_(True), rather than torch.tensor(sourceTensor).\n"
          ]
        },
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "Well done!\n"
          ]
        }
      ],
      "source": [
        "exp_replay = ReplayBuffer(10)\n",
        "\n",
        "for _ in range(30):\n",
        "    exp_replay.add(env.reset(), env.action_space.sample(),\n",
        "                   1.0, env.reset(), done=False)\n",
        "\n",
        "obs_batch, act_batch, reward_batch, next_obs_batch, is_done_batch = exp_replay.sample(\n",
        "    5)\n",
        "\n",
        "assert len(exp_replay) == 10, \"experience replay size should be 10 because that's what maximum capacity is\"\n",
        "\n",
        "exp_replay = ReplayBuffer(2000)\n",
        "\n",
        "state = env.reset()\n",
        "play_and_record(state, agent, env, exp_replay, n_steps=1000)\n",
        "\n",
        "# if you're using your own experience replay buffer, some of those tests may need correction.\n",
        "# just make sure you know what your code does\n",
        "assert len(exp_replay) == 1000, \"play_and_record should have added exactly 1000 steps, \"\\\n",
        "                                 \"but instead added %i\" % len(exp_replay)\n",
        "is_dones = list(zip(*exp_replay._storage))[-1]\n",
        "\n",
        "for _ in range(100):\n",
        "    obs_batch, act_batch, reward_batch, next_obs_batch, is_done_batch = exp_replay.sample(\n",
        "        10)\n",
        "    assert obs_batch.shape == next_obs_batch.shape == (10,) + state_shape\n",
        "    assert act_batch.shape == (\n",
        "        10,), \"actions batch should have shape (10,) but is instead %s\" % str(act_batch.shape)\n",
        "    assert reward_batch.shape == (\n",
        "        10,), \"rewards batch should have shape (10,) but is instead %s\" % str(reward_batch.shape)\n",
        "    assert is_done_batch.shape == (\n",
        "        10,), \"is_done batch should have shape (10,) but is instead %s\" % str(is_done_batch.shape)\n",
        "    assert [int(i) in (0, 1)\n",
        "            for i in is_dones], \"is_done should be strictly True or False\"\n",
        "    assert [\n",
        "        0 <= a < n_actions for a in act_batch], \"actions should be within [0, n_actions)\"\n",
        "\n",
        "print(\"Well done!\")"
      ]
    },
    {
      "cell_type": "markdown",
      "id": "baf20465",
      "metadata": {
        "id": "baf20465"
      },
      "source": [
        "### Target Networks\n",
        "\n",
        "DQN is basically a policy iteration method in which we both approximate the $Q$ values and improve them as we go. This might cause any errors in estimating to propagate and create unwanted difficulties. To address this issue we use **Target Networks**. Recall the main Q-Learning framework:\n",
        "\n",
        "$$Q(s, a; \\Theta) := Q(s, a; \\Theta) + \\eta \\big(R(s, a) + \\gamma \\times \\max_{a'} \\hat{Q}(s, a'; \\Theta')  - Q(s, a; \\Theta)\\big)$$\n",
        "\n",
        "In this framework we approximate the $\\hat{Q}(s, a'; \\Theta')$ values using a network parameterized by $\\Theta'$ and this is the actual target network. After certain frequencies we load the current network into the target network."
      ]
    },
    {
      "cell_type": "code",
      "execution_count": 18,
      "id": "e48d22e2",
      "metadata": {
        "id": "e48d22e2",
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "outputId": "69bdceba-2697-4456-93e9-c6c8d08c3997"
      },
      "outputs": [
        {
          "output_type": "execute_result",
          "data": {
            "text/plain": [
              "<All keys matched successfully>"
            ]
          },
          "metadata": {},
          "execution_count": 18
        }
      ],
      "source": [
        "target_network = DQNAgent(env, epsilon=0.5).to(device)\n",
        "# This is how you can load weights from agent into target network\n",
        "target_network.load_state_dict(agent.state_dict())"
      ]
    },
    {
      "cell_type": "markdown",
      "id": "0ba70316",
      "metadata": {
        "id": "0ba70316"
      },
      "source": [
        "## Q-Learning\n",
        "Here we will finally implement the Q-Learning procedure. Here's the process:\n",
        "\n",
        "Compute Q-learning TD error:\n",
        "\n",
        "$$ L = { 1 \\over N} \\sum_i [ Q_{\\theta}(s,a) - Q_{reference}(s,a) ] ^2 $$\n",
        "\n",
        "With Q-reference defined as\n",
        "\n",
        "$$ Q_{reference}(s,a) = r(s,a) + \\gamma \\cdot max_{a'} Q_{target}(s', a') $$\n",
        "\n",
        "Where\n",
        "* $Q_{target}(s',a')$ denotes q-value of next state and next action predicted by __target_network__\n",
        "* $s, a, r, s'$ are current state, action, reward and next state respectively\n",
        "* $\\gamma$ is a discount factor defined two cells above.\n",
        "\n",
        "\n",
        "__Note 1:__ there's an example input below. Feel free to experiment with it before you write the function.\n",
        "\n",
        "__Note 2:__ compute_td_loss is a source of 99% of bugs in this homework. If reward doesn't improve, it often helps to go through it line by line [with a rubber duck](https://rubberduckdebugging.com/)."
      ]
    },
    {
      "cell_type": "code",
      "execution_count": 19,
      "id": "8858a4f5",
      "metadata": {
        "id": "8858a4f5"
      },
      "outputs": [],
      "source": [
        "def compute_td_loss(states, actions, rewards, next_states, is_done,\n",
        "                    agent, target_network,\n",
        "                    gamma=0.99,\n",
        "                    check_shapes=False,\n",
        "                    device=device):\n",
        "    \"\"\" Compute td loss using torch operations only. Use the formulae above. \"\"\"\n",
        "    states = torch.tensor(states, device=device, dtype=torch.float)    # shape: [batch_size, *state_shape]\n",
        "    \n",
        "    # for some torch reason should not make actions a tensor\n",
        "    actions = torch.tensor(actions, device=device, dtype=torch.long)    # shape: [batch_size]\n",
        "    rewards = torch.tensor(rewards, device=device, dtype=torch.float)  # shape: [batch_size]\n",
        "    # shape: [batch_size, *state_shape]\n",
        "    next_states = torch.tensor(next_states, device=device, dtype=torch.float)\n",
        "    is_done = torch.tensor(is_done.astype('float32'),device=device,dtype=torch.float)  # shape: [batch_size]\n",
        "    is_not_done = 1 - is_done\n",
        "    agent = agent.to(device)\n",
        "    # get q-values for all actions in current states\n",
        "    predicted_qvalues = agent(states)\n",
        "\n",
        "    # compute q-values for all actions in next states\n",
        "    predicted_next_qvalues = target_network(next_states)\n",
        "    \n",
        "    # select q-values for chosen actions\n",
        "    predicted_qvalues_for_actions = predicted_qvalues[range(\n",
        "        len(actions)), actions]\n",
        "\n",
        "    # compute V*(next_states) using predicted next q-values\n",
        "    # TODO\n",
        "    ##################################################################\n",
        "    next_state_values = predicted_next_qvalues.max(axis=-1)[0] \n",
        "    #<YOUR CODE>\n",
        "    ##################################################################\n",
        "\n",
        "    assert next_state_values.dim(\n",
        "    ) == 1 and next_state_values.shape[0] == states.shape[0], \"must predict one value per state\"\n",
        "\n",
        "    # compute \"target q-values\" for loss - it's what's inside square parentheses in the above formula.\n",
        "    # at the last state use the simplified formula: Q(s,a) = r(s,a) since s' doesn't exist\n",
        "    # you can multiply next state values by is_not_done to achieve this.\n",
        "    # TODO\n",
        "    ##################################################################\n",
        "    target_qvalues_for_actions = rewards + gamma * next_state_values * (1-is_done)\n",
        "    # target_qvalues_for_actions = torch.where(is_done , rewards, target_qvalues_for_actions)\n",
        "     #<YOUR CODE>\n",
        "    ##################################################################\n",
        "    # mean squared error loss to minimize\n",
        "    loss = torch.mean((predicted_qvalues_for_actions -\n",
        "                       target_qvalues_for_actions.detach()) ** 2)\n",
        "\n",
        "    if check_shapes:\n",
        "        assert predicted_next_qvalues.data.dim(\n",
        "        ) == 2, \"make sure you predicted q-values for all actions in next state\"\n",
        "        assert next_state_values.data.dim(\n",
        "        ) == 1, \"make sure you computed V(s') as maximum over just the actions axis and not all axes\"\n",
        "        assert target_qvalues_for_actions.data.dim(\n",
        "        ) == 1, \"there's something wrong with target q-values, they must be a vector\"\n",
        "\n",
        "    return loss"
      ]
    },
    {
      "cell_type": "markdown",
      "id": "02fa41ef",
      "metadata": {
        "id": "02fa41ef"
      },
      "source": [
        "Sanity checks!"
      ]
    },
    {
      "cell_type": "code",
      "execution_count": 20,
      "id": "36770c20",
      "metadata": {
        "id": "36770c20"
      },
      "outputs": [],
      "source": [
        "obs_batch, act_batch, reward_batch, next_obs_batch, is_done_batch = exp_replay.sample(10)\n",
        "\n",
        "loss = compute_td_loss(obs_batch, act_batch, reward_batch, next_obs_batch, is_done_batch,\n",
        "                       agent, target_network,\n",
        "                       gamma=0.99, check_shapes=True, device=device)\n",
        "loss.backward()\n",
        "\n",
        "assert loss.requires_grad and tuple(loss.data.size()) == (\n",
        "    ), \"you must return scalar loss - mean over batch\"\n",
        "assert np.any(next(agent.parameters()).grad.data.cpu().numpy() !=\n",
        "              0), \"loss must be differentiable w.r.t. network weights\"\n",
        "assert np.all(next(target_network.parameters()).grad is None), \"target network should not have grads\""
      ]
    },
    {
      "cell_type": "markdown",
      "id": "1b4dd741",
      "metadata": {
        "id": "1b4dd741"
      },
      "source": [
        "## Main Learning Loop"
      ]
    },
    {
      "cell_type": "code",
      "execution_count": 21,
      "id": "d61fead8",
      "metadata": {
        "id": "d61fead8",
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "outputId": "d3af232a-592e-44d7-f3ac-a967ccca1d46"
      },
      "outputs": [
        {
          "output_type": "execute_result",
          "data": {
            "text/plain": [
              "<torch._C.Generator at 0x7f7a39d55230>"
            ]
          },
          "metadata": {},
          "execution_count": 21
        }
      ],
      "source": [
        "seed = 13790110 #<your favourite random seed>\n",
        "random.seed(seed)\n",
        "np.random.seed(seed)\n",
        "torch.manual_seed(seed)"
      ]
    },
    {
      "cell_type": "code",
      "execution_count": 22,
      "id": "4bd7dbbb",
      "metadata": {
        "id": "4bd7dbbb",
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "outputId": "b04e76d4-5ad0-4306-afd6-9be7cba17e02"
      },
      "outputs": [
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "(84, 84, 4) 4\n"
          ]
        },
        {
          "output_type": "execute_result",
          "data": {
            "text/plain": [
              "<All keys matched successfully>"
            ]
          },
          "metadata": {},
          "execution_count": 22
        }
      ],
      "source": [
        "env = make_env(seed)\n",
        "state_shape = env.observation_space.shape\n",
        "n_actions = env.action_space.n\n",
        "state = env.reset()\n",
        "print(state_shape, n_actions)\n",
        "\n",
        "# agent = DQNAgent(state_shape, n_actions, epsilon=1).to(device)\n",
        "agent = DQNAgent(env, epsilon=1).to(device)\n",
        "#target_network = DQNAgent(state_shape, n_actions).to(device)\n",
        "target_network = DQNAgent(env, epsilon=1).to(device)\n",
        "target_network.load_state_dict(agent.state_dict())"
      ]
    },
    {
      "cell_type": "markdown",
      "id": "e9e1bfd9",
      "metadata": {
        "id": "e9e1bfd9"
      },
      "source": [
        "The following are the hyperparameters for the learning process. Feel free to change them as you will; although some default values are assigned to them currently.\n",
        "\n",
        "**IMPORTANT NOTE!** Since the training takes time, you may (and should!) save your model with some frequency."
      ]
    },
    {
      "cell_type": "code",
      "execution_count": 23,
      "id": "dfd35d03",
      "metadata": {
        "id": "dfd35d03"
      },
      "outputs": [],
      "source": [
        "exp_replay = ReplayBuffer(70000)\n",
        "timesteps_per_epoch = 1\n",
        "batch_size = 64\n",
        "total_steps = 3 * 10**6\n",
        "decay_steps = 10**6\n",
        "\n",
        "opt = torch.optim.Adam(agent.parameters(), lr=1e-5)\n",
        "\n",
        "init_epsilon = 1 #0.30096  #0.42945  0.49933  0.622 # 1\n",
        "final_epsilon = 0.1\n",
        "\n",
        "loss_freq = 50\n",
        "refresh_target_network_freq = 5000\n",
        "eval_freq = 5000\n",
        "\n",
        "max_grad_norm = 50\n",
        "\n",
        "n_lives = 5\n",
        "\n",
        "save_freq = 10000"
      ]
    },
    {
      "cell_type": "code",
      "execution_count": 37,
      "id": "70083ca5",
      "metadata": {
        "id": "70083ca5"
      },
      "outputs": [],
      "source": [
        "# DO NOT RERUN THIS TO MAINTAIN THE HISTORY!\n",
        "import pickle\n",
        "\n",
        "mean_rw_history = []\n",
        "td_loss_history = []\n",
        "grad_norm_history = []\n",
        "initial_state_v_history = []\n",
        "step = 0\n",
        "\n",
        "with open('/content/drive/MyDrive/Colab Notebooks/Save_Model_Hw5/mean_rw_history', \"rb\") as fp:\n",
        "  mean_rw_history = pickle.load(fp)\n",
        "with open('/content/drive/MyDrive/Colab Notebooks/Save_Model_Hw5/td_loss_history', \"rb\") as fp:\n",
        "  td_loss_history = pickle.load(fp)\n",
        "with open('/content/drive/MyDrive/Colab Notebooks/Save_Model_Hw5/grad_norm_history', \"rb\") as fp:\n",
        "  grad_norm_history = pickle.load(fp)\n",
        "with open('/content/drive/MyDrive/Colab Notebooks/Save_Model_Hw5/initial_state_v_history', \"rb\") as fp:\n",
        "  initial_state_v_history = pickle.load(fp)\n",
        "\n",
        "agent = torch.load('/content/drive/MyDrive/Colab Notebooks/Save_Model_Hw5/temporary_saved_agent.pth')"
      ]
    },
    {
      "cell_type": "code",
      "execution_count": 38,
      "id": "5a665c30",
      "metadata": {
        "id": "5a665c30",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 989
        },
        "outputId": "c1f2f966-b855-4e88-dc9e-3b6af5331827"
      },
      "outputs": [
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "buffer size = 1, epsilon = 0.30096\n"
          ]
        },
        {
          "output_type": "display_data",
          "data": {
            "image/png": "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\n",
            "text/plain": [
              "<Figure size 1152x648 with 4 Axes>"
            ]
          },
          "metadata": {
            "needs_background": "light"
          }
        },
        {
          "output_type": "stream",
          "name": "stderr",
          "text": [
            "  0%|          | 1/3000001 [01:41<84852:19:21, 101.82s/it]/usr/local/lib/python3.7/dist-packages/ipykernel_launcher.py:57: UserWarning: To copy construct from a tensor, it is recommended to use sourceTensor.clone().detach() or sourceTensor.clone().detach().requires_grad_(True), rather than torch.tensor(sourceTensor).\n",
            "  0%|          | 316/3000001 [01:55<303:45:11,  2.74it/s]\n"
          ]
        },
        {
          "output_type": "error",
          "ename": "KeyboardInterrupt",
          "evalue": "ignored",
          "traceback": [
            "\u001b[0;31m---------------------------------------------------------------------------\u001b[0m",
            "\u001b[0;31mKeyboardInterrupt\u001b[0m                         Traceback (most recent call last)",
            "\u001b[0;32m<ipython-input-38-c6f69841e830>\u001b[0m in \u001b[0;36m<module>\u001b[0;34m()\u001b[0m\n\u001b[1;32m     21\u001b[0m     \u001b[0;31m##################################################################\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[1;32m     22\u001b[0m     \u001b[0;31m# TODO: sample batch_size of data from experience replay>\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[0;32m---> 23\u001b[0;31m     \u001b[0msampleFromExperienceReplay\u001b[0m \u001b[0;34m=\u001b[0m \u001b[0mexp_replay\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0msample\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0mbatch_size\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[0m\u001b[1;32m     24\u001b[0m     \u001b[0mcurrent_states\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0mactions\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0mrewards\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0mnext_state\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0mis_done_s\u001b[0m \u001b[0;34m=\u001b[0m \u001b[0msampleFromExperienceReplay\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[1;32m     25\u001b[0m     \u001b[0mopt\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0mzero_grad\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n",
            "\u001b[0;32m<ipython-input-15-4427737db677>\u001b[0m in \u001b[0;36msample\u001b[0;34m(self, batch_size)\u001b[0m\n\u001b[1;32m     73\u001b[0m             \u001b[0;32mfor\u001b[0m \u001b[0m_\u001b[0m \u001b[0;32min\u001b[0m \u001b[0mrange\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0mbatch_size\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[1;32m     74\u001b[0m         ]\n\u001b[0;32m---> 75\u001b[0;31m         \u001b[0;32mreturn\u001b[0m \u001b[0mself\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0m_encode_sample\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0midxes\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[0m",
            "\u001b[0;32m<ipython-input-15-4427737db677>\u001b[0m in \u001b[0;36m_encode_sample\u001b[0;34m(self, idxes)\u001b[0m\n\u001b[1;32m     45\u001b[0m             \u001b[0mnp\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0marray\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0mactions\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m,\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[1;32m     46\u001b[0m             \u001b[0mnp\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0marray\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0mrewards\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m,\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[0;32m---> 47\u001b[0;31m             \u001b[0mnp\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0marray\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0mobses_tp1\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m,\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[0m\u001b[1;32m     48\u001b[0m             \u001b[0mnp\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0marray\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0mdones\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[1;32m     49\u001b[0m         )\n",
            "\u001b[0;31mKeyboardInterrupt\u001b[0m: "
          ]
        }
      ],
      "source": [
        "import utils\n",
        "from tqdm import trange \n",
        "\n",
        "state = env.reset()\n",
        "for step in trange(step, total_steps + 1):\n",
        "    if not utils.is_enough_ram():\n",
        "        print('less that 100 Mb RAM available, freezing')\n",
        "        print('make sure everythin is ok and make KeyboardInterrupt to continue')\n",
        "        try:\n",
        "            while True:\n",
        "                pass\n",
        "        except KeyboardInterrupt:\n",
        "            pass\n",
        "\n",
        "    agent.epsilon = utils.linear_decay(init_epsilon, final_epsilon, step, decay_steps)\n",
        "\n",
        "    # play\n",
        "    _, state = play_and_record(state, agent, env, exp_replay, timesteps_per_epoch)\n",
        "\n",
        "    # train\n",
        "    ##################################################################\n",
        "    # TODO: sample batch_size of data from experience replay>\n",
        "    sampleFromExperienceReplay = exp_replay.sample(batch_size)\n",
        "    current_states, actions, rewards, next_state, is_done_s = sampleFromExperienceReplay\n",
        "    opt.zero_grad()\n",
        "    loss = compute_td_loss(current_states, actions, rewards, next_state, is_done_s,\n",
        "                           agent, target_network)\n",
        "    # TODO: compute TD loss\n",
        "    ##################################################################\n",
        "    loss.backward()\n",
        "    grad_norm = nn.utils.clip_grad_norm_(agent.parameters(), max_grad_norm)\n",
        "    opt.step()\n",
        "    opt.zero_grad()\n",
        "    \n",
        "    if (step + 1) % save_freq == 0:\n",
        "        torch.save(agent, 'saved-models/temporary_saved_agent.pth')\n",
        "        torch.save(agent, '/content/drive/MyDrive/Colab Notebooks/Save_Model_Hw5/temporary_saved_agent.pth')\n",
        "        with open('saved-models/mean_rw_history', \"wb\") as fp:\n",
        "          pickle.dump(mean_rw_history, fp)\n",
        "        with open('/content/drive/MyDrive/Colab Notebooks/Save_Model_Hw5/mean_rw_history', \"wb\") as fp:\n",
        "          pickle.dump(mean_rw_history, fp)\n",
        "        with open('saved-models/td_loss_history', \"wb\") as fp:\n",
        "          pickle.dump(td_loss_history, fp)\n",
        "        with open('/content/drive/MyDrive/Colab Notebooks/Save_Model_Hw5/td_loss_history', \"wb\") as fp:\n",
        "          pickle.dump(td_loss_history, fp)\n",
        "        with open('saved-models/grad_norm_history', \"wb\") as fp:\n",
        "          pickle.dump(grad_norm_history, fp)\n",
        "        with open('/content/drive/MyDrive/Colab Notebooks/Save_Model_Hw5/grad_norm_history', \"wb\") as fp:\n",
        "          pickle.dump(grad_norm_history, fp)\n",
        "        with open('saved-models/initial_state_v_history', \"wb\") as fp:\n",
        "          pickle.dump(initial_state_v_history, fp)\n",
        "        with open('/content/drive/MyDrive/Colab Notebooks/Save_Model_Hw5/initial_state_v_history', \"wb\") as fp:\n",
        "          pickle.dump(initial_state_v_history, fp)\n",
        "        \n",
        "    if step % loss_freq == 0:\n",
        "        td_loss_history.append(loss.data.cpu().item())\n",
        "        grad_norm_history.append(grad_norm)\n",
        "\n",
        "    if step % refresh_target_network_freq == 0:\n",
        "        # Load agent weights into target_network\n",
        "        # TODO\n",
        "        #<YOUR CODE>\n",
        "        ##################################################################\n",
        "        target_network.load_state_dict(agent.state_dict())\n",
        "        ##################################################################\n",
        "\n",
        "    if step % eval_freq == 0:\n",
        "        mean_rw_history.append(evaluate(\n",
        "            make_env(clip_rewards=True, seed=step), agent, n_games=3 * n_lives, greedy=True)\n",
        "        )\n",
        "        initial_state_q_values = agent.get_qvalues(\n",
        "            [make_env(seed=step).reset()]\n",
        "        )\n",
        "        initial_state_v_history.append(np.max(initial_state_q_values))\n",
        "\n",
        "        clear_output(True)\n",
        "        print(\"buffer size = %i, epsilon = %.5f\" %\n",
        "              (len(exp_replay), agent.epsilon))\n",
        "\n",
        "        plt.figure(figsize=[16, 9])\n",
        "\n",
        "        plt.subplot(2, 2, 1)\n",
        "        plt.title(\"Mean reward per life\")\n",
        "        plt.plot(mean_rw_history)\n",
        "        plt.grid()\n",
        "\n",
        "        assert not np.isnan(td_loss_history[-1])\n",
        "        plt.subplot(2, 2, 2)\n",
        "        plt.title(\"TD loss history (smoothened)\")\n",
        "        plt.plot(utils.smoothen(td_loss_history))\n",
        "        plt.grid()\n",
        "\n",
        "        plt.subplot(2, 2, 3)\n",
        "        plt.title(\"Initial state V\")\n",
        "        plt.plot(initial_state_v_history)\n",
        "        plt.grid()\n",
        "\n",
        "        plt.subplot(2, 2, 4)\n",
        "        plt.title(\"Grad norm history (smoothened)\")\n",
        "        plt.plot(utils.smoothen(grad_norm_history))\n",
        "        plt.grid()\n",
        "\n",
        "        plt.show()"
      ]
    },
    {
      "cell_type": "code",
      "source": [
        "import utils\n",
        "from tqdm import trange \n",
        "\n",
        "plt.figure(figsize=[16, 9])\n",
        "\n",
        "plt.subplot(2, 2, 1)\n",
        "plt.title(\"Mean reward per life\")\n",
        "plt.plot(mean_rw_history)\n",
        "plt.grid()\n",
        "\n",
        "assert not np.isnan(td_loss_history[-1])\n",
        "plt.subplot(2, 2, 2)\n",
        "plt.title(\"TD loss history (smoothened)\")\n",
        "plt.plot(utils.smoothen(td_loss_history))\n",
        "plt.grid()\n",
        "\n",
        "plt.subplot(2, 2, 3)\n",
        "plt.title(\"Initial state V\")\n",
        "plt.plot(initial_state_v_history)\n",
        "plt.grid()\n",
        "\n",
        "plt.subplot(2, 2, 4)\n",
        "plt.title(\"Grad norm history (smoothened)\")\n",
        "plt.plot(utils.smoothen(grad_norm_history))\n",
        "plt.grid()\n",
        "\n",
        "plt.show()"
      ],
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 553
        },
        "id": "6dPMp2Eor0s9",
        "outputId": "2ea6e6f7-fbce-4230-8719-422d2905ffe3"
      },
      "id": "6dPMp2Eor0s9",
      "execution_count": 39,
      "outputs": [
        {
          "output_type": "display_data",
          "data": {
            "image/png": "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\n",
            "text/plain": [
              "<Figure size 1152x648 with 4 Axes>"
            ]
          },
          "metadata": {
            "needs_background": "light"
          }
        }
      ]
    },
    {
      "cell_type": "markdown",
      "id": "447b8d8d",
      "metadata": {
        "id": "447b8d8d"
      },
      "source": [
        "## How to interpret plots:\n",
        "\n",
        "This aint no supervised learning so don't expect anything to improve monotonously. \n",
        "* **TD loss** is the MSE between agent's current Q-values and target Q-values. It may slowly increase or decrease, it's ok. The \"not ok\" behavior includes going NaN or stayng at exactly zero before agent has perfect performance.\n",
        "* **grad norm** just shows the intensivity of training. Not ok is growing to values of about 100 (or maybe even 50) though it depends on network architecture.\n",
        "* **mean reward** is the expected sum of r(s,a) agent gets over the full game session. It will oscillate, but on average it should get higher over time (after a few thousand iterations...). \n",
        " * In basic q-learning implementation it takes about 40k steps to \"warm up\" agent before it starts to get better.\n",
        "* **Initial state V** is the expected discounted reward for episode in the oppinion of the agent. It should behave more smoothly than **mean reward**. It should get higher over time but sometimes can experience drawdowns because of the agaent's overestimates.\n",
        "* **buffer size** - this one is simple. It should go up and cap at max size.\n",
        "* **epsilon** - agent's willingness to explore. If you see that agent's already at 0.01 epsilon before it's average reward is above 0 - it means you need to increase epsilon. Set it back to some 0.2 - 0.5 and decrease the pace at which it goes down.\n",
        "* Smoothing of plots is done with a gaussian kernel\n",
        "\n",
        "At first your agent will lose quickly. Then it will learn to suck less and at least hit the ball a few times before it loses. Finally it will learn to actually score points.\n",
        "\n",
        "**Training will take time.** A lot of it actually. Probably you will not see any improvment during first **150k** time steps (note that by default in this notebook agent is evaluated every 5000 time steps)."
      ]
    },
    {
      "cell_type": "markdown",
      "id": "23c703d0",
      "metadata": {
        "id": "23c703d0"
      },
      "source": [
        "## See how the model works and Assignment Submission\n",
        "\n",
        "To submit the assignment, you can submit a zip file containing the filled up files and notebook. You should make sure to have the following items checked:\n",
        "* Each of the blocks provided should run without error.\n",
        "* Run the training and save your model in `saved-models/<STD_ID>_DQN_<ENV_NAME>.pth` format and upload it alongside the assignment.\n",
        "* The history of your plots **from the first epoch until the end**. \n",
        "* You should also save your output as a gif `<STD_ID>_DQN_<ENV_NAME>.gif` and upload it alongside your files. This gif file contains the performance of your model during a complete round of the chosen game."
      ]
    },
    {
      "cell_type": "code",
      "execution_count": null,
      "id": "847f5884",
      "metadata": {
        "id": "847f5884"
      },
      "outputs": [],
      "source": [
        "# Note that you can change the arguments and parameteres of these following functions as you want,\n",
        "# but your functions must work in the described procedure.\n",
        "\n",
        "from array2gif import write_gif\n",
        "\n",
        "def load_model(PATH):\n",
        "    # This function takes the path of saved model as the input, and outputs the loaded model\n",
        "    Final_agent = torch.load(PATH)\n",
        "    # agent = torch.load('/content/drive/MyDrive/Colab Notebooks/Save_Model_Hw5/temporary_saved_agent.pth')\n",
        "    return Final_agent\n",
        "\n",
        "def generate_gif(model, PATH):\n",
        "    # This funcition evaluate a model during a complete round of the chosen game,\n",
        "    # and saves the gif containing the whole steps of the game in <PATH>\n",
        "    counter = 0\n",
        "    s = env.reset()\n",
        "    total_reward = 0\n",
        "    t_max=10000\n",
        "    final_video = []\n",
        "    while True:\n",
        "        counter += 1\n",
        "        qvalues = model.get_qvalues(torch.tensor([s], dtype=torch.float32))\n",
        "        a = model.sample_actions(qvalues, epsilon=0)[0]\n",
        "        next_s, r, done, _ = env.step(a)\n",
        "        total_reward += r\n",
        "        s = next_s\n",
        "        if done:\n",
        "            break\n",
        "        \n",
        "        if counter % 2 == 0:\n",
        "              screen = env.render(mode='rgb_array')\n",
        "              final_video.append(screen)\n",
        "\n",
        "    print(total_reward)\n",
        "    if total_reward>9:\n",
        "        write_gif(final_video, '97101489_DQN_BreakoutNoFrameskip-v4.gif', fps=60)\n",
        "\n",
        "\n",
        "generate_gif(load_model('/content/drive/MyDrive/Colab Notebooks/Save_Model_Hw5/temporary_saved_agent.pth'), 'final_performance.gif')\n"
      ]
    }
  ],
  "metadata": {
    "colab": {
      "collapsed_sections": [
        "68003dd5",
        "447b8d8d"
      ],
      "name": "Atari_DQN.ipynb",
      "provenance": []
    },
    "kernelspec": {
      "display_name": "Python 3 (ipykernel)",
      "language": "python",
      "name": "python3"
    },
    "language_info": {
      "codemirror_mode": {
        "name": "ipython",
        "version": 3
      },
      "file_extension": ".py",
      "mimetype": "text/x-python",
      "name": "python",
      "nbconvert_exporter": "python",
      "pygments_lexer": "ipython3",
      "version": "3.8.10"
    },
    "accelerator": "GPU"
  },
  "nbformat": 4,
  "nbformat_minor": 5
}