rainbow/01_FrozenLake_DQN.ipynb

{
 "cells": [
  {
   "cell_type": "code",
   "execution_count": 1,
   "id": "c2121cd1",
   "metadata": {},
   "outputs": [],
   "source": [
    "import os\n",
    "from typing import Dict, List, Tuple\n",
    "\n",
    "import gymnasium as gym\n",
    "import matplotlib.pyplot as plt\n",
    "import numpy as np\n",
    "import torch\n",
    "import torch.nn as nn\n",
    "import torch.nn.functional as F\n",
    "import torch.optim as optim\n",
    "from IPython.display import clear_output"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "id": "423245d7",
   "metadata": {},
   "outputs": [],
   "source": []
  },
  {
   "cell_type": "code",
   "execution_count": 100,
   "id": "42b1f7b5",
   "metadata": {},
   "outputs": [],
   "source": [
    "env = gym.make('FrozenLake-v1', render_mode='rgb_array')\n",
    "# env = gym.make('CartPole-v1', render_mode='rgb_array')\n",
    "device = torch.device(\n",
    "    \"cuda\" if torch.cuda.is_available() else \"cpu\"\n",
    ")"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 203,
   "id": "064cb030",
   "metadata": {},
   "outputs": [],
   "source": [
    "class QNetwork(nn.Module):\n",
    "    def __init__(self, in_dim: int, out_dim: int):\n",
    "        super(QNetwork, self).__init__()\n",
    "        '''注意网络的定义为Q(s,a;w),隐变量为w,输入为状态s,输出为当前状态s下所有可能动作A的最大未来收益'''\n",
    "        self.layers = nn.Sequential(\n",
    "            nn.Linear(in_dim, 128),\n",
    "            nn.ReLU(),\n",
    "            nn.Linear(128, 128), \n",
    "            nn.ReLU(), \n",
    "            nn.Linear(128, out_dim)\n",
    "        )\n",
    "#         self.layers.\n",
    "    \n",
    "    def forward(self, x: torch.Tensor) -> torch.Tensor:\n",
    "        '''前向推理实现'''\n",
    "        return self.layers(x)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 322,
   "id": "237b8dce",
   "metadata": {},
   "outputs": [],
   "source": [
    "obs_dim = env.observation_space.n\n",
    "action_dim = env.action_space.n\n",
    "\n",
    "# 构建两个Q函数，其中Q为，Q_target为\n",
    "Q = QNetwork(obs_dim, action_dim).to(device)\n",
    "# Q_target = QNetwork(obs_dim, action_dim).to(device)\n",
    "# Q_target.load_state_dict(Q.state_dict())\n",
    "# Q_target.eval()\n",
    "\n",
    "#\n",
    "optimizer = optim.Adam(Q.parameters(), lr=0.001)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 323,
   "id": "63d4a13b",
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "[0, 0, 0, 1, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0]"
      ]
     },
     "execution_count": 323,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "def convert_state(state):\n",
    "    init_state = torch.tensor([state])\n",
    "    return F.one_hot(init_state, num_classes = obs_dim).numpy()[0].tolist()\n",
    "\n",
    "epsilon = 0.2\n",
    "def select_action(state, explore = True):\n",
    "    # 选择动作，有0.2的概率进行随机探索\n",
    "    if explore and np.random.random() < epsilon:\n",
    "        return env.action_space.sample()\n",
    "    \n",
    "    # 有0.8的概率从当前训练的Q函数中，预测所有可能动作的Q值\n",
    "    cur_state = convert_state(state)\n",
    "    q_value = Q(torch.FloatTensor(cur_state).to(device))\n",
    "#     print(Q.parameters())\n",
    "    action = q_value.argmax().detach().cpu().item()\n",
    "    return action\n",
    "\n",
    "# \n",
    "# state, _ = env.reset()\n",
    "# for i in range(10):\n",
    "#     v = select_action(state)\n",
    "#     print(type(v), v)\n",
    "#     next_state, reward, done, _, _ = env.step(v)\n",
    "\n",
    "convert_state(3)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 324,
   "id": "cea4378d",
   "metadata": {},
   "outputs": [],
   "source": [
    "gamma = 0.98\n",
    "seed = 3\n",
    "\n",
    "scores = []\n",
    "losses = []\n",
    "steps = []\n",
    "#假设要跑100幂\n",
    "epoch = 5000\n",
    "while True:\n",
    "    if epoch <= 0:\n",
    "        break\n",
    "    \n",
    "    epoch -= 1\n",
    "    score = 0\n",
    "    '''使用如下方式时，state的类型是list，否则是一个array'''\n",
    "    state, _ = env.reset()\n",
    "    \n",
    "    step = 0\n",
    "    while True:\n",
    "        '''这里代理选择一个动作来执行'''\n",
    "        action = select_action(state)\n",
    "        '''得到下一个状态,构成(s,a,r,s')'''\n",
    "        next_state, reward, done, truncated, _ = env.step(action)\n",
    "        \n",
    "        score += reward\n",
    "        step += 1\n",
    "        \n",
    "        next_target = convert_state(next_state)\n",
    "        cur_state = convert_state(state)\n",
    "        '''\n",
    "        这里可以计算Q_*(s, A)，以及Q_*(s', A),然后用TD算法，逼近Q_*\n",
    "        '''\n",
    "        index_tensor = torch.tensor([action])\n",
    "        q_value = Q(torch.FloatTensor(cur_state).to(device)).gather(0, index_tensor)\n",
    "        \n",
    "        next_q_value = Q(torch.FloatTensor(next_target).to(device)).max().reshape(1)\n",
    "        mask = 1 - done\n",
    "        target = reward + gamma * next_q_value * mask\n",
    "        \n",
    "#         print(q_value, target, q_value.shape, target.shape)\n",
    "        loss = F.smooth_l1_loss(q_value, target)\n",
    "        optimizer.zero_grad()\n",
    "        loss.backward()\n",
    "        optimizer.step()\n",
    "        \n",
    "        if done:\n",
    "            scores.append(score)\n",
    "            steps.append(step)\n",
    "            break\n",
    "            \n",
    "        state = next_state\n",
    "        \n",
    "        if epsilon < 0.05:\n",
    "            epsilon = 0.05\n",
    "        else:\n",
    "            epsilon *= 0.98\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 331,
   "id": "f7cbd3da",
   "metadata": {},
   "outputs": [
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "/opt/homebrew/anaconda3/lib/python3.11/site-packages/gymnasium/wrappers/record_video.py:94: UserWarning: \u001b[33mWARN: Overwriting existing videos at /Users/webberg/workspace/code/reinforce_learning/rainbow folder (try specifying a different `video_folder` for the `RecordVideo` wrapper if this is not desired)\u001b[0m\n",
      "  logger.warn(\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Moviepy - Building video /Users/webberg/workspace/code/reinforce_learning/rainbow/rl-video-episode-0.mp4.\n",
      "Moviepy - Writing video /Users/webberg/workspace/code/reinforce_learning/rainbow/rl-video-episode-0.mp4\n",
      "\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "                                                                                "
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Moviepy - Done !\n",
      "Moviepy - video ready /Users/webberg/workspace/code/reinforce_learning/rainbow/rl-video-episode-0.mp4\n",
      "score:  1.0 step: 14\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "\r"
     ]
    }
   ],
   "source": [
    "# 观察效果\n",
    "def test():\n",
    "    dvideo_env = gym.wrappers.RecordVideo(env, video_folder='.')\n",
    "    state, _ = dvideo_env.reset(seed=seed)\n",
    "    done = False\n",
    "    score = 0\n",
    "    step = 0\n",
    "    while not done:\n",
    "        action = select_action(state, False)\n",
    "        next_state, reward, done, truncated, _ = dvideo_env.step(action)\n",
    "\n",
    "        state = next_state\n",
    "        score += reward\n",
    "        step +=1\n",
    "\n",
    "    print(\"score: \", score, \"step:\", step)\n",
    "    dvideo_env.close()\n",
    "    \n",
    "test()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 326,
   "id": "acb1416c",
   "metadata": {},
   "outputs": [
    {
     "data": {
      "image/png": "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",
      "text/plain": [
       "<Figure size 2000x500 with 2 Axes>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "# 观察曲线\n",
    "plt.figure(figsize=(20, 5))\n",
    "plt.subplot(131)\n",
    "# plt.title('frame %s. score: %s' % (frame_idx, np.mean(scores[-10:])))\n",
    "plt.title('score')\n",
    "plt.plot(scores)\n",
    "plt.subplot(132)\n",
    "plt.title('steps')\n",
    "plt.plot(steps)\n",
    "# plt.subplot(133)\n",
    "# plt.title('epsilons')\n",
    "# plt.plot(epsilons)\n",
    "plt.show()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "id": "ecf5d907",
   "metadata": {},
   "outputs": [],
   "source": []
  }
 ],
 "metadata": {
  "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": "3.11.5"
  },
  "toc": {
   "base_numbering": 1,
   "nav_menu": {},
   "number_sections": true,
   "sideBar": true,
   "skip_h1_title": false,
   "title_cell": "Table of Contents",
   "title_sidebar": "Contents",
   "toc_cell": false,
   "toc_position": {},
   "toc_section_display": true,
   "toc_window_display": false
  },
  "varInspector": {
   "cols": {
    "lenName": 16,
    "lenType": 16,
    "lenVar": 40
   },
   "kernels_config": {
    "python": {
     "delete_cmd_postfix": "",
     "delete_cmd_prefix": "del ",
     "library": "var_list.py",
     "varRefreshCmd": "print(var_dic_list())"
    },
    "r": {
     "delete_cmd_postfix": ") ",
     "delete_cmd_prefix": "rm(",
     "library": "var_list.r",
     "varRefreshCmd": "cat(var_dic_list()) "
    }
   },
   "types_to_exclude": [
    "module",
    "function",
    "builtin_function_or_method",
    "instance",
    "_Feature"
   ],
   "window_display": false
  }
 },
 "nbformat": 4,
 "nbformat_minor": 5
}