carlthome/Proximal Policy Optimization (PPO) of the CartPole problem with PyTorch.ipynb

## Proximal Policy Optimization (PPO) of the CartPole problem with PyTorch.ipynb
{
  "nbformat": 4,
  "nbformat_minor": 0,
  "metadata": {
    "colab": {
      "provenance": [],
      "include_colab_link": true
    },
    "kernelspec": {
      "name": "python3",
      "display_name": "Python 3"
    },
    "language_info": {
      "name": "python"
    },
    "widgets": {
      "application/vnd.jupyter.widget-state+json": {
        "9eb3f337d54041adba415a90fc714218": {
          "model_module": "@jupyter-widgets/controls",
          "model_name": "HBoxModel",
          "model_module_version": "1.5.0",
          "state": {
            "_dom_classes": [],
            "_model_module": "@jupyter-widgets/controls",
            "_model_module_version": "1.5.0",
            "_model_name": "HBoxModel",
            "_view_count": null,
            "_view_module": "@jupyter-widgets/controls",
            "_view_module_version": "1.5.0",
            "_view_name": "HBoxView",
            "box_style": "",
            "children": [
              "IPY_MODEL_f1da27748e4d4a19bde362c1152220bd",
              "IPY_MODEL_4cd8e38a83314f30a1d6d4b2dbb87d1c",
              "IPY_MODEL_d475415af510481eae562066aa800379"
            ],
            "layout": "IPY_MODEL_d01b6ac880054fa6b3fa9d271e651983"
          }
        },
        "f1da27748e4d4a19bde362c1152220bd": {
          "model_module": "@jupyter-widgets/controls",
          "model_name": "HTMLModel",
          "model_module_version": "1.5.0",
          "state": {
            "_dom_classes": [],
            "_model_module": "@jupyter-widgets/controls",
            "_model_module_version": "1.5.0",
            "_model_name": "HTMLModel",
            "_view_count": null,
            "_view_module": "@jupyter-widgets/controls",
            "_view_module_version": "1.5.0",
            "_view_name": "HTMLView",
            "description": "",
            "description_tooltip": null,
            "layout": "IPY_MODEL_44c54eb3e04f44b49e4e25a9887b807f",
            "placeholder": "",
            "style": "IPY_MODEL_336b6ce9b017402b8627f54ccdcf28b1",
            "value": "Training: 100%"
          }
        },
        "4cd8e38a83314f30a1d6d4b2dbb87d1c": {
          "model_module": "@jupyter-widgets/controls",
          "model_name": "FloatProgressModel",
          "model_module_version": "1.5.0",
          "state": {
            "_dom_classes": [],
            "_model_module": "@jupyter-widgets/controls",
            "_model_module_version": "1.5.0",
            "_model_name": "FloatProgressModel",
            "_view_count": null,
            "_view_module": "@jupyter-widgets/controls",
            "_view_module_version": "1.5.0",
            "_view_name": "ProgressView",
            "bar_style": "success",
            "description": "",
            "description_tooltip": null,
            "layout": "IPY_MODEL_ac920b2a8fd14cdc8e1204344a68e712",
            "max": 300,
            "min": 0,
            "orientation": "horizontal",
            "style": "IPY_MODEL_b3213e27c9694d30bd329f9acaee2bba",
            "value": 300
          }
        },
        "d475415af510481eae562066aa800379": {
          "model_module": "@jupyter-widgets/controls",
          "model_name": "HTMLModel",
          "model_module_version": "1.5.0",
          "state": {
            "_dom_classes": [],
            "_model_module": "@jupyter-widgets/controls",
            "_model_module_version": "1.5.0",
            "_model_name": "HTMLModel",
            "_view_count": null,
            "_view_module": "@jupyter-widgets/controls",
            "_view_module_version": "1.5.0",
            "_view_name": "HTMLView",
            "description": "",
            "description_tooltip": null,
            "layout": "IPY_MODEL_c160814f1ed6487aaf89907fd7ff9375",
            "placeholder": "",
            "style": "IPY_MODEL_9bf93268fba940738d13448296a2c8e9",
            "value": " 300/300 [02:09&lt;00:00, 13.62it/s]"
          }
        },
        "d01b6ac880054fa6b3fa9d271e651983": {
          "model_module": "@jupyter-widgets/base",
          "model_name": "LayoutModel",
          "model_module_version": "1.2.0",
          "state": {
            "_model_module": "@jupyter-widgets/base",
            "_model_module_version": "1.2.0",
            "_model_name": "LayoutModel",
            "_view_count": null,
            "_view_module": "@jupyter-widgets/base",
            "_view_module_version": "1.2.0",
            "_view_name": "LayoutView",
            "align_content": null,
            "align_items": null,
            "align_self": null,
            "border": null,
            "bottom": null,
            "display": null,
            "flex": null,
            "flex_flow": null,
            "grid_area": null,
            "grid_auto_columns": null,
            "grid_auto_flow": null,
            "grid_auto_rows": null,
            "grid_column": null,
            "grid_gap": null,
            "grid_row": null,
            "grid_template_areas": null,
            "grid_template_columns": null,
            "grid_template_rows": null,
            "height": null,
            "justify_content": null,
            "justify_items": null,
            "left": null,
            "margin": null,
            "max_height": null,
            "max_width": null,
            "min_height": null,
            "min_width": null,
            "object_fit": null,
            "object_position": null,
            "order": null,
            "overflow": null,
            "overflow_x": null,
            "overflow_y": null,
            "padding": null,
            "right": null,
            "top": null,
            "visibility": null,
            "width": null
          }
        },
        "44c54eb3e04f44b49e4e25a9887b807f": {
          "model_module": "@jupyter-widgets/base",
          "model_name": "LayoutModel",
          "model_module_version": "1.2.0",
          "state": {
            "_model_module": "@jupyter-widgets/base",
            "_model_module_version": "1.2.0",
            "_model_name": "LayoutModel",
            "_view_count": null,
            "_view_module": "@jupyter-widgets/base",
            "_view_module_version": "1.2.0",
            "_view_name": "LayoutView",
            "align_content": null,
            "align_items": null,
            "align_self": null,
            "border": null,
            "bottom": null,
            "display": null,
            "flex": null,
            "flex_flow": null,
            "grid_area": null,
            "grid_auto_columns": null,
            "grid_auto_flow": null,
            "grid_auto_rows": null,
            "grid_column": null,
            "grid_gap": null,
            "grid_row": null,
            "grid_template_areas": null,
            "grid_template_columns": null,
            "grid_template_rows": null,
            "height": null,
            "justify_content": null,
            "justify_items": null,
            "left": null,
            "margin": null,
            "max_height": null,
            "max_width": null,
            "min_height": null,
            "min_width": null,
            "object_fit": null,
            "object_position": null,
            "order": null,
            "overflow": null,
            "overflow_x": null,
            "overflow_y": null,
            "padding": null,
            "right": null,
            "top": null,
            "visibility": null,
            "width": null
          }
        },
        "336b6ce9b017402b8627f54ccdcf28b1": {
          "model_module": "@jupyter-widgets/controls",
          "model_name": "DescriptionStyleModel",
          "model_module_version": "1.5.0",
          "state": {
            "_model_module": "@jupyter-widgets/controls",
            "_model_module_version": "1.5.0",
            "_model_name": "DescriptionStyleModel",
            "_view_count": null,
            "_view_module": "@jupyter-widgets/base",
            "_view_module_version": "1.2.0",
            "_view_name": "StyleView",
            "description_width": ""
          }
        },
        "ac920b2a8fd14cdc8e1204344a68e712": {
          "model_module": "@jupyter-widgets/base",
          "model_name": "LayoutModel",
          "model_module_version": "1.2.0",
          "state": {
            "_model_module": "@jupyter-widgets/base",
            "_model_module_version": "1.2.0",
            "_model_name": "LayoutModel",
            "_view_count": null,
            "_view_module": "@jupyter-widgets/base",
            "_view_module_version": "1.2.0",
            "_view_name": "LayoutView",
            "align_content": null,
            "align_items": null,
            "align_self": null,
            "border": null,
            "bottom": null,
            "display": null,
            "flex": null,
            "flex_flow": null,
            "grid_area": null,
            "grid_auto_columns": null,
            "grid_auto_flow": null,
            "grid_auto_rows": null,
            "grid_column": null,
            "grid_gap": null,
            "grid_row": null,
            "grid_template_areas": null,
            "grid_template_columns": null,
            "grid_template_rows": null,
            "height": null,
            "justify_content": null,
            "justify_items": null,
            "left": null,
            "margin": null,
            "max_height": null,
            "max_width": null,
            "min_height": null,
            "min_width": null,
            "object_fit": null,
            "object_position": null,
            "order": null,
            "overflow": null,
            "overflow_x": null,
            "overflow_y": null,
            "padding": null,
            "right": null,
            "top": null,
            "visibility": null,
            "width": null
          }
        },
        "b3213e27c9694d30bd329f9acaee2bba": {
          "model_module": "@jupyter-widgets/controls",
          "model_name": "ProgressStyleModel",
          "model_module_version": "1.5.0",
          "state": {
            "_model_module": "@jupyter-widgets/controls",
            "_model_module_version": "1.5.0",
            "_model_name": "ProgressStyleModel",
            "_view_count": null,
            "_view_module": "@jupyter-widgets/base",
            "_view_module_version": "1.2.0",
            "_view_name": "StyleView",
            "bar_color": null,
            "description_width": ""
          }
        },
        "c160814f1ed6487aaf89907fd7ff9375": {
          "model_module": "@jupyter-widgets/base",
          "model_name": "LayoutModel",
          "model_module_version": "1.2.0",
          "state": {
            "_model_module": "@jupyter-widgets/base",
            "_model_module_version": "1.2.0",
            "_model_name": "LayoutModel",
            "_view_count": null,
            "_view_module": "@jupyter-widgets/base",
            "_view_module_version": "1.2.0",
            "_view_name": "LayoutView",
            "align_content": null,
            "align_items": null,
            "align_self": null,
            "border": null,
            "bottom": null,
            "display": null,
            "flex": null,
            "flex_flow": null,
            "grid_area": null,
            "grid_auto_columns": null,
            "grid_auto_flow": null,
            "grid_auto_rows": null,
            "grid_column": null,
            "grid_gap": null,
            "grid_row": null,
            "grid_template_areas": null,
            "grid_template_columns": null,
            "grid_template_rows": null,
            "height": null,
            "justify_content": null,
            "justify_items": null,
            "left": null,
            "margin": null,
            "max_height": null,
            "max_width": null,
            "min_height": null,
            "min_width": null,
            "object_fit": null,
            "object_position": null,
            "order": null,
            "overflow": null,
            "overflow_x": null,
            "overflow_y": null,
            "padding": null,
            "right": null,
            "top": null,
            "visibility": null,
            "width": null
          }
        },
        "9bf93268fba940738d13448296a2c8e9": {
          "model_module": "@jupyter-widgets/controls",
          "model_name": "DescriptionStyleModel",
          "model_module_version": "1.5.0",
          "state": {
            "_model_module": "@jupyter-widgets/controls",
            "_model_module_version": "1.5.0",
            "_model_name": "DescriptionStyleModel",
            "_view_count": null,
            "_view_module": "@jupyter-widgets/base",
            "_view_module_version": "1.2.0",
            "_view_name": "StyleView",
            "description_width": ""
          }
        }
      }
    }
  },
  "cells": [
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "view-in-github",
        "colab_type": "text"
      },
      "source": [
        "<a href=\"https://colab.research.google.com/gist/carlthome/542a9899dedcb2c7adf3403df3fe1f14/proximal-policy-optimization-ppo-of-the-cartpole-problem-with-pytorch.ipynb\" target=\"_parent\"><img src=\"https://colab.research.google.com/assets/colab-badge.svg\" alt=\"Open In Colab\"/></a>"
      ]
    },
    {
      "cell_type": "code",
      "execution_count": null,
      "metadata": {
        "id": "rFU7C-g5IN-n"
      },
      "outputs": [],
      "source": [
        "import os\n",
        "\n",
        "import torch\n",
        "import numpy as np"
      ]
    },
    {
      "cell_type": "code",
      "source": [
        "class Memory:\n",
        "    def __init__(self, batch_size):\n",
        "        self.batch_size = batch_size\n",
        "        self.history = {\n",
        "            \"state\": [],\n",
        "            \"probs\": [],\n",
        "            \"values\": [],\n",
        "            \"action\": [],\n",
        "            \"reward\": [],\n",
        "            \"done\": [],\n",
        "        }\n",
        "\n",
        "    def generate_batches(self):\n",
        "        num_states = len(self.history[\"state\"])\n",
        "        batch_start = np.arange(0, num_states, self.batch_size)\n",
        "        indices = np.arange(num_states, dtype=np.int64)\n",
        "        np.random.shuffle(indices)\n",
        "        rows = [indices[i : i + self.batch_size] for i in batch_start]\n",
        "        batches = {k: np.array(v) for k, v in self.history.items()}\n",
        "        batches[\"batches\"] = rows\n",
        "        return batches\n",
        "\n",
        "    def append(self, row):\n",
        "        for k, v in row.items():\n",
        "            self.history[k].append(v)\n",
        "\n",
        "    def clear(self):\n",
        "        for k in self.history:\n",
        "            self.history[k].clear()"
      ],
      "metadata": {
        "id": "9SrnghsTIVyW"
      },
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "code",
      "source": [
        "class Actor(torch.nn.Module):\n",
        "    def __init__(\n",
        "        self,\n",
        "        num_actions,\n",
        "        input_dims,\n",
        "        alpha,\n",
        "        fc1_dims=256,\n",
        "        fc2_dims=256,\n",
        "        chkpt_dir=\"tmp/ppo\",\n",
        "    ):\n",
        "        super().__init__()\n",
        "        self.checkpoint_file = os.path.join(chkpt_dir, \"actor_torch_ppo\")\n",
        "\n",
        "        self.actor = torch.nn.Sequential(\n",
        "            torch.nn.Linear(*input_dims, fc1_dims),\n",
        "            torch.nn.ReLU(),\n",
        "            torch.nn.Linear(fc1_dims, fc2_dims),\n",
        "            torch.nn.ReLU(),\n",
        "            torch.nn.Linear(fc2_dims, num_actions),\n",
        "            torch.nn.Softmax(dim=-1),\n",
        "        )\n",
        "\n",
        "        self.optimizer = torch.optim.Adam(self.parameters(), lr=alpha)\n",
        "        self.device = torch.device(\"cuda:0\" if torch.cuda.is_available() else \"cpu\")\n",
        "        self.to(self.device)\n",
        "\n",
        "    def forward(self, state):\n",
        "        dist = self.actor(state)\n",
        "        dist = torch.distributions.categorical.Categorical(dist)\n",
        "        return dist\n",
        "\n",
        "    def save_checkpoint(self):\n",
        "        torch.save(self.state_dict(), self.checkpoint_file)\n",
        "\n",
        "    def load_checkpoint(self):\n",
        "        self.load_state_dict(torch.load(self.checkpoint_file))"
      ],
      "metadata": {
        "id": "5xhn1E_YJLF4"
      },
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "code",
      "source": [
        "class Critic(torch.nn.Module):\n",
        "    def __init__(\n",
        "        self, input_dims, alpha, fc1_dims=256, fc2_dims=256, chkpt_dir=\"tmp/ppo\"\n",
        "    ):\n",
        "        super().__init__()\n",
        "        self.checkpoint_file = os.path.join(chkpt_dir, \"critic_torch_ppo\")\n",
        "        self.critic = torch.nn.Sequential(\n",
        "            torch.nn.Linear(*input_dims, fc1_dims),\n",
        "            torch.nn.ReLU(),\n",
        "            torch.nn.Linear(fc1_dims, fc2_dims),\n",
        "            torch.nn.Linear(fc2_dims, 1),\n",
        "        )\n",
        "\n",
        "        self.optimizer = torch.optim.Adam(self.parameters(), lr=alpha)\n",
        "        self.device = torch.device(\"cuda:0\" if torch.cuda.is_available() else \"cpu\")\n",
        "        self.to(self.device)\n",
        "\n",
        "    def forward(self, state):\n",
        "        value = self.critic(state)\n",
        "        return value\n",
        "\n",
        "    def save_checkpoint(self):\n",
        "        torch.save(self.state_dict(), self.checkpoint_file)\n",
        "\n",
        "    def load_checkpoint(self):\n",
        "        self.load_state_dict(torch.load(self.checkpoint_file))"
      ],
      "metadata": {
        "id": "L-zogfxjKBXD"
      },
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "code",
      "source": [
        "class Agent:\n",
        "    def __init__(\n",
        "        self,\n",
        "        num_actions,\n",
        "        input_dims,\n",
        "        gamma=0.99,\n",
        "        alpha=1e-3,\n",
        "        gae_lambda=0.95,\n",
        "        policy_clip=0.2,\n",
        "        batch_size=64,\n",
        "        N=2048,\n",
        "        num_epochs=10,\n",
        "    ):\n",
        "        self.gamma = gamma\n",
        "        self.policy_clip = policy_clip\n",
        "        self.num_epochs = num_epochs\n",
        "        self.gae_lambda = gae_lambda\n",
        "\n",
        "        self.actor = Actor(num_actions, input_dims, alpha)\n",
        "        self.critic = Critic(input_dims, alpha)\n",
        "        self.memory = Memory(batch_size)\n",
        "\n",
        "    def remember(self, state, action, probs, values, reward, done):\n",
        "        row = {\n",
        "            \"state\": state,\n",
        "            \"probs\": probs,\n",
        "            \"values\": values,\n",
        "            \"action\": action,\n",
        "            \"reward\": reward,\n",
        "            \"done\": done,\n",
        "        }\n",
        "        self.memory.append(row)\n",
        "\n",
        "    def save_models(self):\n",
        "        self.actor.save_checkpoint()\n",
        "        self.critic.save_checkpoint()\n",
        "\n",
        "    def load_models(self):\n",
        "        self.actor.load_checkpoint()\n",
        "        self.critic.load_checkpoint()\n",
        "\n",
        "    def choose_action(self, observation):\n",
        "        state = torch.tensor(np.asarray([observation]), dtype=torch.float).to(self.actor.device)\n",
        "\n",
        "        dist = self.actor(state)\n",
        "        value = self.critic(state)\n",
        "        action = dist.sample()\n",
        "\n",
        "        probs = torch.squeeze(dist.log_prob(action)).item()\n",
        "        action = torch.squeeze(action).item()\n",
        "        value = torch.squeeze(value).item()\n",
        "\n",
        "        return action, probs, value\n",
        "\n",
        "    def learn(self):\n",
        "        for _ in range(self.num_epochs):\n",
        "\n",
        "            batches = self.memory.generate_batches()\n",
        "            state_history = batches[\"state\"]\n",
        "            old_probs_history = batches[\"probs\"]\n",
        "            values_history = batches[\"values\"]\n",
        "            action_history = batches[\"action\"]\n",
        "            reward_history = batches[\"reward\"]\n",
        "            dones_history = batches[\"done\"]\n",
        "            batches = batches[\"batches\"]\n",
        "\n",
        "            values = values_history\n",
        "            advantage = np.zeros(len(reward_history), dtype=np.float32)\n",
        "            for t in range(len(reward_history) - 1):\n",
        "                discount = 1.0\n",
        "                a_t = 0\n",
        "                for k in range(t, len(reward_history) - 1):\n",
        "                    a_t += discount * (\n",
        "                        reward_history[k]\n",
        "                        + self.gamma * values[k + 1] * (1 - int(dones_history[k]))\n",
        "                        - values[k]\n",
        "                    )\n",
        "                    discount *= self.gamma * self.gae_lambda\n",
        "                advantage[t] = a_t\n",
        "\n",
        "            advantage = torch.tensor(advantage).to(self.actor.device)\n",
        "\n",
        "            values = torch.tensor(values).to(self.actor.device)\n",
        "\n",
        "            for batch in batches:\n",
        "                states = torch.tensor(state_history[batch], dtype=torch.float).to(self.actor.device)\n",
        "                old_probs = torch.tensor(old_probs_history[batch]).to(self.actor.device)\n",
        "                actions = torch.tensor(action_history[batch]).to(self.actor.device)\n",
        "\n",
        "                dist = self.actor(states)\n",
        "                critic_value = self.critic(states)\n",
        "\n",
        "                critic_value = torch.squeeze(critic_value)\n",
        "\n",
        "                new_probs = dist.log_prob(actions)\n",
        "                prob_ratio = new_probs.exp() / old_probs.exp()\n",
        "\n",
        "                weighted_probs = advantage[batch] * prob_ratio\n",
        "                weighted_clipped_probs = (\n",
        "                    torch.clamp(prob_ratio, 1 - self.policy_clip, 1 + self.policy_clip)\n",
        "                    * advantage[batch]\n",
        "                )\n",
        "\n",
        "                actor_loss = -torch.min(weighted_probs, weighted_clipped_probs).mean()\n",
        "\n",
        "                returns = advantage[batch] + values[batch]\n",
        "                critic_loss = (returns - critic_value) ** 2\n",
        "                critic_loss = critic_loss.mean()\n",
        "\n",
        "                total_loss = actor_loss + 0.5 * critic_loss\n",
        "\n",
        "                self.actor.optimizer.zero_grad()\n",
        "                self.critic.optimizer.zero_grad()\n",
        "\n",
        "                total_loss.backward()\n",
        "\n",
        "                self.actor.optimizer.step()\n",
        "                self.critic.optimizer.step()\n",
        "\n",
        "        self.memory.clear()"
      ],
      "metadata": {
        "id": "oaawO7INLMbc"
      },
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "markdown",
      "source": [
        "# Experiments\n"
      ],
      "metadata": {
        "id": "vhZwr2rhPiXS"
      }
    },
    {
      "cell_type": "code",
      "source": [
        "!mkdir -p tmp/ppo"
      ],
      "metadata": {
        "id": "jIkeuMqQRQNn"
      },
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "code",
      "source": [
        "import gym\n",
        "import numpy as np\n",
        "import matplotlib.pyplot as plt\n",
        "import tqdm.auto as tqdm\n",
        "\n",
        "env = gym.make(\"CartPole-v1\", new_step_api=True)\n",
        "env.action_space.seed(42)\n",
        "steps_per_update = 20\n",
        "batch_size = 5\n",
        "num_epochs = 4\n",
        "alpha = 1e-3\n",
        "\n",
        "agent = Agent(\n",
        "    num_actions=env.action_space.n,\n",
        "    batch_size=batch_size,\n",
        "    alpha=alpha,\n",
        "    num_epochs=num_epochs,\n",
        "    input_dims=env.observation_space.shape,\n",
        ")\n",
        "\n",
        "num_episodes = 300\n",
        "score = 0\n",
        "score_history = []\n",
        "best_score = env.reward_range[0]\n",
        "num_updates = 0\n",
        "num_steps = 0\n",
        "\n",
        "records = []\n",
        "for episode in tqdm.trange(num_episodes, desc=\"Training\"):\n",
        "    observation = env.reset()\n",
        "    terminated = False\n",
        "    truncated = False\n",
        "    score = 0\n",
        "    while not terminated and not truncated:\n",
        "        action, prob, val = agent.choose_action(observation)\n",
        "\n",
        "        new_observation, reward, terminated, truncated, info = env.step(action)\n",
        "        num_steps += 1\n",
        "        score += reward\n",
        "\n",
        "        done = truncated or terminated\n",
        "        agent.remember(observation, action, prob, val, reward, done)\n",
        "\n",
        "        if num_steps % steps_per_update == 0:\n",
        "            agent.learn()\n",
        "            num_updates += 1\n",
        "\n",
        "        observation = new_observation\n",
        "\n",
        "    score_history.append(score)\n",
        "\n",
        "    avg_score = np.mean(score_history[-100:])\n",
        "    if avg_score > best_score:\n",
        "        best_score = avg_score\n",
        "        agent.save_models()\n",
        "\n",
        "    record = {\n",
        "        \"Episode\": episode,\n",
        "        \"Score\": score,\n",
        "        \"Average Score\": avg_score,\n",
        "        \"Best Score\": best_score,\n",
        "        \"Step\": num_steps,\n",
        "        \"Updates\": num_updates,\n",
        "    }\n",
        "    records.append(record)"
      ],
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 49,
          "referenced_widgets": [
            "9eb3f337d54041adba415a90fc714218",
            "f1da27748e4d4a19bde362c1152220bd",
            "4cd8e38a83314f30a1d6d4b2dbb87d1c",
            "d475415af510481eae562066aa800379",
            "d01b6ac880054fa6b3fa9d271e651983",
            "44c54eb3e04f44b49e4e25a9887b807f",
            "336b6ce9b017402b8627f54ccdcf28b1",
            "ac920b2a8fd14cdc8e1204344a68e712",
            "b3213e27c9694d30bd329f9acaee2bba",
            "c160814f1ed6487aaf89907fd7ff9375",
            "9bf93268fba940738d13448296a2c8e9"
          ]
        },
        "collapsed": true,
        "id": "nKPFUrdiPbgD",
        "outputId": "388dcc48-6b33-41d0-f33f-5132e8044e58"
      },
      "execution_count": null,
      "outputs": [
        {
          "output_type": "display_data",
          "data": {
            "text/plain": [
              "Training:   0%|          | 0/300 [00:00<?, ?it/s]"
            ],
            "application/vnd.jupyter.widget-view+json": {
              "version_major": 2,
              "version_minor": 0,
              "model_id": "9eb3f337d54041adba415a90fc714218"
            }
          },
          "metadata": {}
        }
      ]
    },
    {
      "cell_type": "code",
      "source": [
        "import pandas as pd\n",
        "\n",
        "pd.DataFrame(records).set_index(\"Episode\").plot(subplots=True, figsize=(5, 10));"
      ],
      "metadata": {
        "id": "qDVSZfVfUXwH",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 850
        },
        "outputId": "0c00ebd6-3972-43e2-e583-5236c60823c1"
      },
      "execution_count": null,
      "outputs": [
        {
          "output_type": "display_data",
          "data": {
            "text/plain": [
              "<Figure size 500x1000 with 5 Axes>"
            ],
            "image/png": "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\n"
          },
          "metadata": {}
        }
      ]
    }
  ]
}