DDPG for the Crawler environment .ipynb

{
 "cells": [
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "# DDPG for the Crawler environment \n",
    "\n",
    "---\n",
    "\n",
    "This notebook is based on the notebook of the coding exercice [DDPG-pendulum](https://github.com/udacity/deep-reinforcement-learning/tree/master/ddpg-pendulum) of the [Deep Reinforcement Learning Nanodegree](https://www.udacity.com/course/deep-reinforcement-learning-nanodegree--nd893) program.\n",
    "\n",
    "\n",
    "Here, we will train a DDPG agent in the Crawler environment.\n",
    "\n",
    "\n",
    "### 1. Import the Necessary Packages\n",
    "\n",
    "We begin by importing the necessary packages."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 1,
   "metadata": {},
   "outputs": [],
   "source": [
    "from unityagents import UnityEnvironment\n",
    "import torch\n",
    "import numpy as np\n",
    "from collections import deque\n",
    "import matplotlib.pyplot as plt\n",
    "%matplotlib inline\n",
    "\n",
    "from ddpg_agent import Agent"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### 2. Instantiate the Environment and Agent\n",
    "\n",
    "Next, we will start the environment!  **_Before running the code cell below_**, change the `file_name` parameter to match the location of the Unity environment that you can downloaded it from [here](https://github.com/udacity/deep-reinforcement-learning/tree/master/p2_continuous-control).\n",
    "\n",
    "- **Mac**: `\"path/to/Crawler.app\"`\n",
    "- **Windows** (x86): `\"path/to/Crawler_Windows_x86/Crawler.exe\"`\n",
    "- **Windows** (x86_64): `\"path/to/Crawler_Windows_x86_64/Crawler.exe\"`\n",
    "- **Linux** (x86): `\"path/to/Crawler_Linux/Crawler.x86\"`\n",
    "- **Linux** (x86_64): `\"path/to/Crawler_Linux/Crawler.x86_64\"`\n",
    "- **Linux** (x86, headless): `\"path/to/Crawler_Linux_NoVis/Crawler.x86\"`\n",
    "- **Linux** (x86_64, headless): `\"path/to/Crawler_Linux_NoVis/Crawler.x86_64\"`\n",
    "\n",
    "For instance, if you are using a Mac, then you downloaded `Crawler.app`.  If this file is in the same folder as the notebook, then the line below should appear as follows:\n",
    "```\n",
    "env = UnityEnvironment(file_name=\"Crawler.app\")\n",
    "```"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 2,
   "metadata": {},
   "outputs": [
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "INFO:unityagents:\n",
      "'Academy' started successfully!\n",
      "Unity Academy name: Academy\n",
      "        Number of Brains: 1\n",
      "        Number of External Brains : 1\n",
      "        Lesson number : 0\n",
      "        Reset Parameters :\n",
      "\t\t\n",
      "Unity brain name: CrawlerBrain\n",
      "        Number of Visual Observations (per agent): 0\n",
      "        Vector Observation space type: continuous\n",
      "        Vector Observation space size (per agent): 129\n",
      "        Number of stacked Vector Observation: 1\n",
      "        Vector Action space type: continuous\n",
      "        Vector Action space size (per agent): 20\n",
      "        Vector Action descriptions: , , , , , , , , , , , , , , , , , , , \n"
     ]
    }
   ],
   "source": [
    "env = UnityEnvironment(file_name='Crawler_Windows_x86_64/Crawler.exe')"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "Environments contain **_brains_** which are responsible for deciding the actions of their associated agents. Here we check for the first brain available, and set it as the default brain we will be controlling from Python."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 3,
   "metadata": {},
   "outputs": [],
   "source": [
    "# get the default brain\n",
    "brain_name = env.brain_names[0]\n",
    "brain = env.brains[brain_name]"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "#### Examine the State and Action Spaces\n",
    "\n",
    "Run the code cell below to print some information about the environment."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 4,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Number of agents: 12\n",
      "Size of each action: 20\n",
      "There are 12 agents. Each observes a state with length: 129\n",
      "The state for the first agent looks like: [ 0.00000000e+00  0.00000000e+00  0.00000000e+00  2.25000000e+00\n",
      "  1.00000000e+00  0.00000000e+00  1.78813934e-07  0.00000000e+00\n",
      "  1.00000000e+00  0.00000000e+00  0.00000000e+00  0.00000000e+00\n",
      "  0.00000000e+00  0.00000000e+00  0.00000000e+00  0.00000000e+00\n",
      "  0.00000000e+00  0.00000000e+00  0.00000000e+00  0.00000000e+00\n",
      "  0.00000000e+00  0.00000000e+00  0.00000000e+00  0.00000000e+00\n",
      "  6.06093168e-01 -1.42857209e-01 -6.06078804e-01  0.00000000e+00\n",
      "  0.00000000e+00  0.00000000e+00  0.00000000e+00  0.00000000e+00\n",
      "  0.00000000e+00  0.00000000e+00  0.00000000e+00  0.00000000e+00\n",
      "  0.00000000e+00  0.00000000e+00  1.33339906e+00 -1.42857209e-01\n",
      " -1.33341408e+00  0.00000000e+00  0.00000000e+00  0.00000000e+00\n",
      "  0.00000000e+00  0.00000000e+00  0.00000000e+00  0.00000000e+00\n",
      "  0.00000000e+00  0.00000000e+00  0.00000000e+00  0.00000000e+00\n",
      " -6.06093347e-01 -1.42857209e-01 -6.06078625e-01  0.00000000e+00\n",
      "  0.00000000e+00  0.00000000e+00  0.00000000e+00  0.00000000e+00\n",
      "  0.00000000e+00  0.00000000e+00  0.00000000e+00  0.00000000e+00\n",
      "  0.00000000e+00  0.00000000e+00 -1.33339953e+00 -1.42857209e-01\n",
      " -1.33341372e+00  0.00000000e+00  0.00000000e+00  0.00000000e+00\n",
      "  0.00000000e+00  0.00000000e+00  0.00000000e+00  0.00000000e+00\n",
      "  0.00000000e+00  0.00000000e+00  0.00000000e+00  0.00000000e+00\n",
      " -6.06093168e-01 -1.42857209e-01  6.06078804e-01  0.00000000e+00\n",
      "  0.00000000e+00  0.00000000e+00  0.00000000e+00  0.00000000e+00\n",
      "  0.00000000e+00  0.00000000e+00  0.00000000e+00  0.00000000e+00\n",
      "  0.00000000e+00  0.00000000e+00 -1.33339906e+00 -1.42857209e-01\n",
      "  1.33341408e+00  0.00000000e+00  0.00000000e+00  0.00000000e+00\n",
      "  0.00000000e+00  0.00000000e+00  0.00000000e+00  0.00000000e+00\n",
      "  0.00000000e+00  0.00000000e+00  0.00000000e+00  0.00000000e+00\n",
      "  6.06093347e-01 -1.42857209e-01  6.06078625e-01  0.00000000e+00\n",
      "  0.00000000e+00  0.00000000e+00  0.00000000e+00  0.00000000e+00\n",
      "  0.00000000e+00  0.00000000e+00  0.00000000e+00  0.00000000e+00\n",
      "  0.00000000e+00  0.00000000e+00  1.33339953e+00 -1.42857209e-01\n",
      "  1.33341372e+00  0.00000000e+00  0.00000000e+00  0.00000000e+00\n",
      "  0.00000000e+00]\n"
     ]
    }
   ],
   "source": [
    "# reset the environment\n",
    "env_info = env.reset(train_mode=True)[brain_name]\n",
    "\n",
    "# number of agents\n",
    "num_agents = len(env_info.agents)\n",
    "print('Number of agents:', num_agents)\n",
    "\n",
    "# size of each action\n",
    "action_size = brain.vector_action_space_size\n",
    "print('Size of each action:', action_size)\n",
    "\n",
    "# examine the state space \n",
    "states = env_info.vector_observations\n",
    "state_size = states.shape[1]\n",
    "print('There are {} agents. Each observes a state with length: {}'.format(states.shape[0], state_size))\n",
    "print('The state for the first agent looks like:', states[0])"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 5,
   "metadata": {},
   "outputs": [],
   "source": [
    "# Initialise the agent\n",
    "agent = Agent(state_size, action_size, random_seed=1)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### 3. Train the Agent with DDPG\n",
    "\n",
    "Run the code cell below to train the agent from scratch.  Alternatively, you can skip to the next code cell to load the pre-trained weights from file."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 6,
   "metadata": {},
   "outputs": [],
   "source": [
    "def ddpg(n_episodes=200, max_t=1000):\n",
    "    scores_deque = deque(maxlen=100)\n",
    "    scores = []\n",
    "    max_score = -np.Inf\n",
    "    for i_episode in range(1, n_episodes+1):\n",
    "        env_info = env.reset(train_mode=True)[brain_name]     # reset the environment    \n",
    "        states = env_info.vector_observations                  # get the current state (for each agent)\n",
    "        agent.reset()\n",
    "        score = 0\n",
    "        for t in range(max_t):\n",
    "            actions = agent.act(states)\n",
    "            env_info = env.step(actions)[brain_name]           # send all actions to tne environment\n",
    "            next_states = env_info.vector_observations         # get next state (for each agent)\n",
    "            rewards = env_info.rewards                         # get reward (for each agent)\n",
    "            dones = env_info.local_done                        # see if episode finished\n",
    "            \n",
    "            for state, action, reward, next_state, done in zip(states, actions, rewards, next_states, dones):\n",
    "                agent.step(state, action, reward, next_state, done, t)\n",
    "                \n",
    "            states = next_states\n",
    "            score += np.mean(rewards)                         # update the score (for each agent)\n",
    "            if np.any(dones):\n",
    "                break \n",
    "                \n",
    "        scores_deque.append(score)\n",
    "        scores.append(score)\n",
    "        mean=np.mean(scores_deque)\n",
    "        \n",
    "        print('\\rEpisode {}\\tAverage Score: {:.2f}\\tScore: {:.2f}'.format(i_episode, mean , score), end=\"\")\n",
    "        if i_episode % 100 == 0:\n",
    "            print('\\rEpisode {}\\tAverage Score: {:.2f}'.format(i_episode, mean))\n",
    "        if  mean>max_score:\n",
    "            max_score=mean\n",
    "            torch.save(agent.actor_local.state_dict(), 'checkpoint_actor_Crawler.pth')\n",
    "            torch.save(agent.critic_local.state_dict(), 'checkpoint_critic_Crawler.pth')\n",
    "            \n",
    "    return scores"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 7,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Episode 100\tAverage Score: 2.65\tScore: 3.95\n",
      "Episode 200\tAverage Score: 4.90\tScore: 8.759\n",
      "Episode 300\tAverage Score: 5.81\tScore: 6.792\n",
      "Episode 400\tAverage Score: 2.67\tScore: 8.131\n",
      "Episode 500\tAverage Score: 4.18\tScore: 4.930\n",
      "Episode 600\tAverage Score: 4.53\tScore: 6.590\n",
      "Episode 700\tAverage Score: 6.59\tScore: 8.185\n",
      "Episode 800\tAverage Score: 9.67\tScore: 5.083\n",
      "Episode 900\tAverage Score: 10.71\tScore: 11.76\n",
      "Episode 1000\tAverage Score: 12.40\tScore: 12.54\n",
      "Episode 1100\tAverage Score: 14.47\tScore: 23.31\n",
      "Episode 1200\tAverage Score: 14.91\tScore: 16.67\n",
      "Episode 1300\tAverage Score: 23.72\tScore: 31.93\n",
      "Episode 1400\tAverage Score: 20.73\tScore: 23.70\n",
      "Episode 1500\tAverage Score: 24.57\tScore: 25.95\n",
      "Episode 1600\tAverage Score: 23.99\tScore: 13.49\n",
      "Episode 1700\tAverage Score: 5.79\tScore: 12.675\n",
      "Episode 1800\tAverage Score: 20.02\tScore: 26.15\n",
      "Episode 1900\tAverage Score: 26.22\tScore: 28.24\n",
      "Episode 2000\tAverage Score: 32.32\tScore: 39.80\n"
     ]
    }
   ],
   "source": [
    "scores = ddpg(2000)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 8,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 640x480 with 1 Axes>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "fig = plt.figure()\n",
    "ax = fig.add_subplot(111)\n",
    "plt.plot(np.arange(1, len(scores)+1), scores)\n",
    "plt.ylabel('Score')\n",
    "plt.xlabel('Episode #')\n",
    "plt.show()"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### 4. Watch the Smart Agent!\n",
    "\n",
    "In the next code cell, we will load the trained weights and test the model."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 9,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Total score (averaged over agents) this episode: 47.86441034134865\n"
     ]
    }
   ],
   "source": [
    "agent.actor_local.load_state_dict(torch.load('checkpoint_actor_Crawler.pth'))\n",
    "agent.critic_local.load_state_dict(torch.load('checkpoint_critic_Crawler.pth'))\n",
    "\n",
    "env_info = env.reset(train_mode=False)[brain_name]     # reset the environment    \n",
    "states = env_info.vector_observations                  # get the current state (for each agent)\n",
    "scores = np.zeros(num_agents)                          # initialize the score (for each agent)\n",
    "while True:\n",
    "    actions = agent.act(states)\n",
    "    env_info = env.step(actions)[brain_name]           # send all actions to tne environment\n",
    "    next_states = env_info.vector_observations         # get next state (for each agent)\n",
    "    rewards = env_info.rewards                         # get reward (for each agent)\n",
    "    dones = env_info.local_done                        # see if episode finished\n",
    "    scores += env_info.rewards                         # update the score (for each agent)\n",
    "    states = next_states                               # roll over states to next time step\n",
    "    if np.any(dones):                                  # exit loop if episode finished\n",
    "        break\n",
    "print('Total score (averaged over agents) this episode: {}'.format(np.mean(scores)))"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "When finished, we can close the environment."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 10,
   "metadata": {},
   "outputs": [],
   "source": [
    "env.close()"
   ]
  }
 ],
 "metadata": {
  "kernelspec": {
   "display_name": "drlnd",
   "language": "python",
   "name": "drlnd"
  },
  "language_info": {
   "codemirror_mode": {
    "name": "ipython",
    "version": 3
   },
   "file_extension": ".py",
   "mimetype": "text/x-python",
   "name": "python",
   "nbconvert_exporter": "python",
   "pygments_lexer": "ipython3",
   "version": "3.10.4"
  }
 },
 "nbformat": 4,
 "nbformat_minor": 2
}