policy/06-TD3-Pendulum.ipynb

{
 "cells": [
  {
   "cell_type": "code",
   "execution_count": 9,
   "id": "06c77d47",
   "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\n",
    "from torch.distributions import Categorical\n",
    "import math\n",
    "import torch.nn as nn\n",
    "import random\n",
    "from tqdm import tqdm"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 10,
   "id": "32b9bc7e",
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "(3,) 3 Box(-2.0, 2.0, (1,), float32) 2.0\n"
     ]
    }
   ],
   "source": [
    "env = gym.make(\"Pendulum-v1\", render_mode=\"rgb_array\")\n",
    "\n",
    "device = torch.device(\n",
    "    \"cuda\" if torch.cuda.is_available() else \"cpu\"\n",
    ")\n",
    "\n",
    "obs_dim = env.observation_space.shape[0]\n",
    "action_info = env.action_space\n",
    "max_action = float(env.action_space.high[0])\n",
    "print(env.observation_space.shape, obs_dim, action_info, max_action)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 11,
   "id": "bd281c63",
   "metadata": {},
   "outputs": [],
   "source": [
    "class ReplayBuffer:\n",
    "    def __init__(self, max_size: int, obs_dim: int, batch_size: int = 32):\n",
    "        self._obs_buf = np.zeros([max_size, obs_dim], dtype=np.float32)\n",
    "        self._act_buf = np.zeros([max_size], dtype=np.float32)\n",
    "        self._reward = np.zeros([max_size], dtype=np.float32)\n",
    "        self._next_obs_buf = np.zeros([max_size, obs_dim], dtype=np.float32)\n",
    "        self._done = np.zeros([max_size], dtype=np.float32)\n",
    "        \n",
    "        self._index = 0\n",
    "        self._size = 0\n",
    "        self._max_size = max_size\n",
    "        self._batch_size = batch_size\n",
    "        \n",
    "    def store(self, obs: np.ndarray, act: np.ndarray, rew: float, next_obs: np.ndarray, done: bool):\n",
    "        self._obs_buf[self._index] = obs\n",
    "        self._act_buf[self._index] = act\n",
    "        self._reward[self._index] = rew\n",
    "        self._next_obs_buf[self._index] = next_obs\n",
    "        self._done[self._index] = done\n",
    "        self._index = (self._index + 1) % self._max_size\n",
    "        self._size = min(self._size + 1, self._max_size)\n",
    "        \n",
    "    def __len__(self) -> int:\n",
    "        return self._size\n",
    "    \n",
    "    def sample(self) -> Dict[str, np.ndarray]:\n",
    "        idxs = np.random.choice(self._size, size=self._batch_size, replace=False)\n",
    "        return dict(obs=self._obs_buf[idxs],\n",
    "                    acts=self._act_buf[idxs],\n",
    "                    rews=self._reward[idxs],\n",
    "                    next_obs=self._next_obs_buf[idxs],\n",
    "                    done=self._done[idxs])"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 12,
   "id": "f9697090",
   "metadata": {},
   "outputs": [],
   "source": [
    "class ActionNetwork(nn.Module):\n",
    "    def __init__(self, in_dim: int, out_dim: int):\n",
    "        \"\"\"Initialization.\"\"\"\n",
    "        super(ActionNetwork, self).__init__()\n",
    "\n",
    "        self.layers = nn.Sequential(\n",
    "            nn.Linear(in_dim, 400, device=device), \n",
    "            nn.ReLU(),\n",
    "            nn.Linear(400, 300, device=device), \n",
    "            nn.ReLU(), \n",
    "            nn.Linear(300, out_dim, device=device)\n",
    "        )\n",
    "\n",
    "    def forward(self, x: torch.Tensor) -> torch.Tensor:\n",
    "        \"\"\"Forward method implementation.\"\"\"\n",
    "        return max_action * torch.tanh(self.layers(x))\n",
    "    \n",
    "class QValueNetwork(nn.Module):\n",
    "    def __init__(self, in_dim: int, out_dim: int):\n",
    "        '''状态价值，用来评估动作的好坏程度'''\n",
    "        super(QValueNetwork, self).__init__()\n",
    "        self.layers = nn.Sequential(\n",
    "            nn.Linear(in_dim + out_dim, 400, device=device), \n",
    "            nn.ReLU(),\n",
    "            nn.Linear(400, 300, device=device), \n",
    "            nn.ReLU(), \n",
    "            nn.Linear(300, 1, device=device)\n",
    "        )\n",
    "        \n",
    "    def forward(self, x: torch.Tensor, a: torch.Tensor) -> torch.Tensor:\n",
    "        \"\"\"Forward method implementation.\"\"\"\n",
    "        t = torch.cat([x, a], 1)\n",
    "        return self.layers(t)"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "b3714c9e",
   "metadata": {},
   "source": [
    "### 使用目标网络"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 13,
   "id": "691070da",
   "metadata": {},
   "outputs": [],
   "source": [
    "actor = ActionNetwork(obs_dim, 1)\n",
    "actor_target = ActionNetwork(obs_dim, 1)\n",
    "actor_target.load_state_dict(actor.state_dict())\n",
    "actor_optimizer = optim.Adam(actor.parameters(), lr=0.0001)\n",
    "\n",
    "critic = QValueNetwork(obs_dim, 1)\n",
    "critic_target = QValueNetwork(obs_dim, 1)\n",
    "critic_target.load_state_dict(critic.state_dict())\n",
    "critic_optimizer = optim.Adam(critic.parameters(), lr=0.0001)\n",
    "\n",
    "buffer = ReplayBuffer(2000, obs_dim, 32)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 14,
   "id": "af1e8c0d",
   "metadata": {},
   "outputs": [],
   "source": [
    "def critic_value(state, action):\n",
    "    return critic(state, action)\n",
    "\n",
    "def critic_target_value(state, action):\n",
    "    return critic_target(state, action)\n",
    "\n",
    "def get_action(state):\n",
    "    action = actor(state)\n",
    "    return action\n",
    "\n",
    "def get_action_target(state):\n",
    "    action = actor_target(state)\n",
    "    return action"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 25,
   "id": "d202b2ec",
   "metadata": {},
   "outputs": [
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "  0%|                                                                                         | 0/1000 [00:00<?, ?it/s]C:\\Users\\webbe\\AppData\\Local\\Temp\\ipykernel_9948\\777992856.py:16: DeprecationWarning: Conversion of an array with ndim > 0 to a scalar is deprecated, and will error in future. Ensure you extract a single element from your array before performing this operation. (Deprecated NumPy 1.25.)\n",
      "  self._act_buf[self._index] = act\n",
      "100%|██████████████████████████████████████████████████████████████████████████████| 1000/1000 [08:23<00:00,  1.98it/s]\n"
     ]
    }
   ],
   "source": [
    "gamma = 0.98\n",
    "tau = 0.005\n",
    "exploration_noise = 0.1\n",
    "max_epoch = 1000\n",
    "batch_size = 16\n",
    "\n",
    "steps = []\n",
    "scores = []\n",
    "view_losses = []\n",
    "\n",
    "for i in tqdm(range(max_epoch)):\n",
    "    score = 0\n",
    "    step = 0\n",
    "    cnt = 1\n",
    "    \n",
    "    s, _ = env.reset(seed=3)\n",
    "    while True:\n",
    "        '''让策略网络做预测'''\n",
    "        action = get_action_target(torch.tensor(s).to(device))\n",
    "        action = (action.cpu().data.numpy() + np.random.normal(0, exploration_noise, size=env.action_space.shape[0])).clip(\n",
    "                    env.action_space.low, env.action_space.high)\n",
    "        next_state, reward, terminated, truncated, _ = env.step(np.float32(action))\n",
    "        done = terminated or truncated\n",
    "        \n",
    "        step += 1\n",
    "        score += reward\n",
    "#         print(np.float32(action), reward)\n",
    "        \n",
    "        buffer.store(s, action, reward, next_state, 1 - done)\n",
    "        if done:\n",
    "            steps.append(step)\n",
    "            scores.append(score)\n",
    "            break\n",
    "        \n",
    "        s = next_state\n",
    "    # 结束一轮游戏\n",
    "    avg_loss = 0\n",
    "#     print(score, L)\n",
    "    for j in range(32):\n",
    "        samples = buffer.sample()\n",
    "        states = torch.FloatTensor(samples['obs']).to(device)\n",
    "        actions = torch.FloatTensor(samples['acts']).reshape(-1, 1).to(device)\n",
    "        rewards = torch.FloatTensor(samples['rews']).reshape(-1, 1).to(device)\n",
    "        next_states = torch.FloatTensor(samples['next_obs']).to(device)\n",
    "        masks = torch.LongTensor(samples['done']).reshape(-1, 1).to(device)\n",
    "#         print(masks.shape)\n",
    "        '''让价值网络做预测'''\n",
    "        q_t = critic_value(states, actions)\n",
    "        next_actions = get_action_target(next_states)\n",
    "        next_q_t = critic_target_value(next_states, next_actions.to(device))\n",
    "        '''计算价值网络的 TD 目标和 TD 误差'''\n",
    "        y_t = rewards + gamma * (next_q_t * masks)\n",
    "#         print('reward_t:', reward_t.shape, 'y_t:', y_t.shape, 'rewards:', rewards, 'result:', (next_q_t * masks).shape)\n",
    "        critic_loss = F.mse_loss(q_t, y_t)\n",
    "        '''更新价值网络'''\n",
    "        critic_optimizer.zero_grad()\n",
    "        critic_loss.backward()\n",
    "        critic_optimizer.step()\n",
    "\n",
    "        new_acts = get_action(states)\n",
    "#         print(new_acts)\n",
    "        actor_loss = -critic_value(states, new_acts).mean()\n",
    "#         print(actor_loss)\n",
    "#         print('state:', states)\n",
    "#         print('act:', new_acts)\n",
    "#         actor_loss = actor_loss.mean()\n",
    "        avg_loss += actor_loss.item()\n",
    "\n",
    "        '''更新策略网络'''\n",
    "        actor_optimizer.zero_grad()\n",
    "        actor_loss.backward()\n",
    "        actor_optimizer.step()\n",
    "\n",
    "        for param, target_param in zip(critic.parameters(), critic_target.parameters()):\n",
    "            target_param.data.copy_(tau * param.data + (1 - tau) * target_param.data)\n",
    "\n",
    "        for param, target_param in zip(actor.parameters(), actor_target.parameters()):\n",
    "            target_param.data.copy_(tau * param.data + (1 - tau) * target_param.data)\n",
    "\n",
    "\n",
    "        cnt += 1\n",
    "    \n",
    "    view_losses.append(avg_loss / cnt)\n",
    "         "
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 26,
   "id": "a5708ce5",
   "metadata": {},
   "outputs": [
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 1440x360 with 3 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "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('actor loss')\n",
    "plt.plot(view_losses)\n",
    "plt.show()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 15,
   "id": "b6fcc157",
   "metadata": {},
   "outputs": [],
   "source": [
    "# 观察效果\n",
    "def test():\n",
    "    dvideo_env = gym.wrappers.RecordVideo(env, video_folder='.')\n",
    "    state, _ = dvideo_env.reset(seed=3)\n",
    "    done = False\n",
    "    score = 0\n",
    "    step = 0\n",
    "    while not done:\n",
    "        action = get_action_target(torch.tensor(state).to(device))\n",
    "#         print(action)\n",
    "        next_state, reward, done, truncated, _ = dvideo_env.step(action.cpu().detach().numpy())\n",
    "\n",
    "        state = next_state\n",
    "        score += reward\n",
    "        step +=1\n",
    "        \n",
    "        if step > 200:\n",
    "            break\n",
    "\n",
    "    print(\"score: \", score, \"step:\", step)\n",
    "    dvideo_env.close()\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "id": "55189e8b",
   "metadata": {},
   "outputs": [],
   "source": [
    "test()"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "2b20858b",
   "metadata": {},
   "source": [
    "### 截断双Q学习"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 16,
   "id": "b53c0598",
   "metadata": {},
   "outputs": [],
   "source": [
    "def generate_critics():\n",
    "    critic = QValueNetwork(obs_dim, 1)\n",
    "    critic_target = QValueNetwork(obs_dim, 1)\n",
    "    critic_target.load_state_dict(critic.state_dict())\n",
    "    critic_optimizer = optim.Adam(critic.parameters(), lr=0.0001)\n",
    "    return critic, critic_target, critic_optimizer"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 45,
   "id": "4536f83b",
   "metadata": {},
   "outputs": [],
   "source": [
    "actor = ActionNetwork(obs_dim, 1)\n",
    "actor_target = ActionNetwork(obs_dim, 1)\n",
    "actor_target.load_state_dict(actor.state_dict())\n",
    "actor_optimizer = optim.Adam(actor.parameters(), lr=0.0001)\n",
    "\n",
    "buffer = ReplayBuffer(2000, obs_dim, 32)\n",
    "critic_1, critic_target_1, critic_optimizer_1 = generate_critics()\n",
    "critic_2, critic_target_2, critic_optimizer_2 = generate_critics()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 46,
   "id": "413f9dc3",
   "metadata": {},
   "outputs": [
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "100%|███████████████████████████████████████| 2000/2000 [04:40<00:00,  7.14it/s]\n"
     ]
    }
   ],
   "source": [
    "gamma = 0.98\n",
    "tau = 0.005\n",
    "exploration_noise = 0.1\n",
    "max_epoch = 2000\n",
    "batch_size = 16\n",
    "\n",
    "steps = []\n",
    "scores = []\n",
    "view_losses = []\n",
    "\n",
    "s, _ = env.reset(seed=3)\n",
    "while True:\n",
    "    '''让策略网络做预测'''\n",
    "    action = get_action_target(torch.tensor(s).to(device))\n",
    "    action = (action.cpu().data.numpy() + np.random.normal(0, exploration_noise, size=env.action_space.shape[0])).clip(\n",
    "                env.action_space.low, env.action_space.high)\n",
    "    next_state, reward, terminated, truncated, _ = env.step(np.float32(action))\n",
    "    done = terminated or truncated\n",
    "\n",
    "    step += 1\n",
    "#         print(np.float32(action), reward)\n",
    "\n",
    "    buffer.store(s, action, reward, next_state, 1 - done)\n",
    "    if done:\n",
    "        steps.append(step)\n",
    "        scores.append(score)\n",
    "        s, _ = env.reset(seed=3)\n",
    "        break\n",
    "\n",
    "    s = next_state\n",
    "    \n",
    "for i in tqdm(range(max_epoch)):\n",
    "    score = 0\n",
    "    step = 0\n",
    "    cnt = 1\n",
    "    actor_cnt = 1\n",
    "    \n",
    "    '''让策略网络做预测'''\n",
    "    action = get_action_target(torch.tensor(s).to(device))\n",
    "    action = (action.cpu().data.numpy() + np.random.normal(0, exploration_noise, size=env.action_space.shape[0])).clip(\n",
    "                env.action_space.low, env.action_space.high)\n",
    "    next_state, reward, terminated, truncated, _ = env.step(np.float32(action))\n",
    "    done = terminated or truncated\n",
    "\n",
    "    step += 1\n",
    "#         print(np.float32(action), reward)\n",
    "\n",
    "    buffer.store(s, action, reward, next_state, 1 - done)\n",
    "    if done:\n",
    "        steps.append(step)\n",
    "        scores.append(score)\n",
    "        s, _ = env.reset(seed=3)\n",
    "    else:\n",
    "        s = next_state\n",
    "        \n",
    "    # 结束一轮游戏\n",
    "    avg_loss = 0\n",
    "#     print(score, L)\n",
    "    for j in range(64):\n",
    "        samples = buffer.sample()\n",
    "        states = torch.FloatTensor(samples['obs']).to(device)\n",
    "        actions = torch.FloatTensor(samples['acts']).reshape(-1, 1).to(device)\n",
    "        rewards = torch.FloatTensor(samples['rews']).reshape(-1, 1).to(device)\n",
    "        next_states = torch.FloatTensor(samples['next_obs']).to(device)\n",
    "        masks = torch.LongTensor(samples['done']).reshape(-1, 1).to(device)\n",
    "#         print(masks.shape)\n",
    "        '''让价值网络做预测'''\n",
    "        next_actions = get_action_target(next_states)\n",
    "        next_q_t1 = critic_target_1(next_states, next_actions.to(device))\n",
    "        next_q_t2 = critic_target_2(next_states, next_actions.to(device))\n",
    "        next_q_t = torch.min(next_q_t1, next_q_t2)\n",
    "        '''计算价值网络的 TD 目标和 TD 误差'''\n",
    "        y_t = rewards + gamma * (next_q_t * masks).detach()\n",
    "        \n",
    "        q_t1 = critic_1(states, actions)\n",
    "        q_t2 = critic_2(states, actions)\n",
    "        \n",
    "#         print('reward_t:', reward_t.shape, 'y_t:', y_t.shape, 'rewards:', rewards, 'result:', (next_q_t * masks).shape)\n",
    "        critic_loss_1 = F.mse_loss(q_t1, y_t)\n",
    "        critic_optimizer_1.zero_grad()\n",
    "        critic_loss_1.backward()\n",
    "        critic_optimizer_1.step()\n",
    "        \n",
    "        critic_loss_2 = F.mse_loss(q_t2, y_t)\n",
    "        critic_optimizer_2.zero_grad()\n",
    "        critic_loss_2.backward()\n",
    "        critic_optimizer_2.step()\n",
    "\n",
    "        if cnt % 2 == 0:\n",
    "            new_acts = get_action(states)\n",
    "    #         print(new_acts)\n",
    "            actor_loss = -critic_1(states, new_acts).mean()\n",
    "    #         print(actor_loss)\n",
    "    #         print('state:', states)\n",
    "    #         print('act:', new_acts)\n",
    "    #         actor_loss = actor_loss.mean()\n",
    "            avg_loss += actor_loss.item()\n",
    "\n",
    "            '''更新策略网络'''\n",
    "            actor_optimizer.zero_grad()\n",
    "            actor_loss.backward()\n",
    "            actor_optimizer.step()\n",
    "            actor_cnt += 1\n",
    "            for param, target_param in zip(critic_1.parameters(), critic_target_1.parameters()):\n",
    "                target_param.data.copy_(tau * param.data + (1 - tau) * target_param.data)\n",
    "\n",
    "            for param, target_param in zip(critic_2.parameters(), critic_target_2.parameters()):\n",
    "                target_param.data.copy_(tau * param.data + (1 - tau) * target_param.data)\n",
    "\n",
    "            for param, target_param in zip(actor.parameters(), actor_target.parameters()):\n",
    "                target_param.data.copy_(tau * param.data + (1 - tau) * target_param.data)\n",
    "\n",
    "        cnt += 1\n",
    "    \n",
    "#     view_losses.append(avg_loss / actor_cnt)\n",
    "         "
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 47,
   "id": "0fa238fe",
   "metadata": {},
   "outputs": [
    {
     "data": {
      "image/png": "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",
      "text/plain": [
       "<Figure size 2000x500 with 3 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('actor loss')\n",
    "plt.plot(view_losses)\n",
    "plt.show()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 48,
   "id": "3c2c1444",
   "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/policy 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/policy/rl-video-episode-0.mp4.\n",
      "Moviepy - Writing video /Users/webberg/workspace/code/reinforce_learning/policy/rl-video-episode-0.mp4\n",
      "\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "                                                                                \r"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Moviepy - Done !\n",
      "Moviepy - video ready /Users/webberg/workspace/code/reinforce_learning/policy/rl-video-episode-0.mp4\n",
      "score:  -1351.714561243295 step: 201\n"
     ]
    }
   ],
   "source": [
    "test()"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "c633de32",
   "metadata": {},
   "source": [
    "### 向目标策略网络中加入噪声"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 54,
   "id": "460a1ee9",
   "metadata": {},
   "outputs": [],
   "source": [
    "def add_noise(action, exploration_noise = 0.1, size = env.action_space.shape[0], low = env.action_space.low, high = env.action_space.high):\n",
    "    action = action.detach().numpy() + np.random.normal(0, exploration_noise, size=size)\n",
    "    return action.clip(low, high)\n",
    "\n",
    "def gen_noise(action, policy_noise = 0.2, clip = 0.5):\n",
    "    noise = torch.ones_like(action).data.normal_(0, policy_noise).to(device)\n",
    "    noise = noise.clamp(-clip, clip)\n",
    "    return noise"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 64,
   "id": "58349fe2",
   "metadata": {},
   "outputs": [],
   "source": [
    "actor = ActionNetwork(obs_dim, 1)\n",
    "actor_target = ActionNetwork(obs_dim, 1)\n",
    "actor_target.load_state_dict(actor.state_dict())\n",
    "actor_optimizer = optim.Adam(actor.parameters(), lr=0.0001)\n",
    "\n",
    "max_buffer_size = 10000\n",
    "buffer = ReplayBuffer(max_buffer_size, obs_dim, 64)\n",
    "critic_1, critic_target_1, critic_optimizer_1 = generate_critics()\n",
    "critic_2, critic_target_2, critic_optimizer_2 = generate_critics()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 66,
   "id": "bf8f6609",
   "metadata": {},
   "outputs": [
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "100%|█████████████████████████████████████| 10000/10000 [12:39<00:00, 13.17it/s]\n"
     ]
    }
   ],
   "source": [
    "gamma = 0.98\n",
    "tau = 0.005\n",
    "exploration_noise = 0.1\n",
    "max_epoch = 10000\n",
    "batch_size = 16\n",
    "\n",
    "steps = []\n",
    "scores = []\n",
    "view_losses = []\n",
    "\n",
    "s, _ = env.reset(seed=3)\n",
    "while len(buffer) < max_buffer_size:\n",
    "    '''让策略网络做预测'''\n",
    "    action = get_action_target(torch.tensor(s).to(device))\n",
    "    action = add_noise(action)\n",
    "    next_state, reward, terminated, truncated, _ = env.step(np.float32(action))\n",
    "    done = terminated or truncated\n",
    "\n",
    "    step += 1\n",
    "#         print(np.float32(action), reward)\n",
    "\n",
    "    buffer.store(s, action, reward, next_state, 1 - done)\n",
    "    if done:\n",
    "        steps.append(step)\n",
    "        scores.append(score)\n",
    "        s, _ = env.reset(seed=31)\n",
    "        continue\n",
    "\n",
    "    s = next_state\n",
    "    \n",
    "for i in tqdm(range(max_epoch)):\n",
    "    score = 0\n",
    "    step = 0\n",
    "    cnt = 1\n",
    "    actor_cnt = 1\n",
    "    \n",
    "    '''让策略网络做预测'''\n",
    "    action = get_action_target(torch.tensor(s).to(device))\n",
    "    action = add_noise(action)\n",
    "    next_state, reward, terminated, truncated, _ = env.step(np.float32(action))\n",
    "    done = terminated or truncated\n",
    "\n",
    "    step += 1\n",
    "#         print(np.float32(action), reward)\n",
    "\n",
    "    buffer.store(s, action, reward, next_state, 1 - done)\n",
    "    if done:\n",
    "        steps.append(step)\n",
    "        scores.append(score)\n",
    "        s, _ = env.reset(seed=3)\n",
    "    else:\n",
    "        s = next_state\n",
    "        \n",
    "    # 结束一轮游戏\n",
    "    avg_loss = 0\n",
    "#     print(score, L)\n",
    "    for j in range(32):\n",
    "        samples = buffer.sample()\n",
    "        states = torch.FloatTensor(samples['obs']).to(device)\n",
    "        actions = torch.FloatTensor(samples['acts']).reshape(-1, 1).to(device)\n",
    "        rewards = torch.FloatTensor(samples['rews']).reshape(-1, 1).to(device)\n",
    "        next_states = torch.FloatTensor(samples['next_obs']).to(device)\n",
    "        masks = torch.LongTensor(samples['done']).reshape(-1, 1).to(device)\n",
    "#         print(masks.shape)\n",
    "        '''让价值网络做预测'''\n",
    "        noise = gen_noise(actions)\n",
    "        next_actions = get_action_target(next_states) + noise\n",
    "        next_q_t1 = critic_target_1(next_states, next_actions.to(device))\n",
    "        next_q_t2 = critic_target_2(next_states, next_actions.to(device))\n",
    "        next_q_t = torch.min(next_q_t1, next_q_t2)\n",
    "        '''计算价值网络的 TD 目标和 TD 误差'''\n",
    "        y_t = rewards + gamma * (next_q_t * masks).detach()\n",
    "        \n",
    "        q_t1 = critic_1(states, actions)\n",
    "        q_t2 = critic_2(states, actions)\n",
    "        \n",
    "#         print('reward_t:', reward_t.shape, 'y_t:', y_t.shape, 'rewards:', rewards, 'result:', (next_q_t * masks).shape)\n",
    "        critic_loss_1 = F.mse_loss(q_t1, y_t)\n",
    "        critic_optimizer_1.zero_grad()\n",
    "        critic_loss_1.backward()\n",
    "        critic_optimizer_1.step()\n",
    "        \n",
    "        critic_loss_2 = F.mse_loss(q_t2, y_t)\n",
    "        critic_optimizer_2.zero_grad()\n",
    "        critic_loss_2.backward()\n",
    "        critic_optimizer_2.step()\n",
    "\n",
    "        if cnt % 2 == 0:\n",
    "            new_acts = get_action(states)\n",
    "    #         print(new_acts)\n",
    "            actor_loss = -critic_1(states, new_acts).mean()\n",
    "    #         print(actor_loss)\n",
    "    #         print('state:', states)\n",
    "    #         print('act:', new_acts)\n",
    "    #         actor_loss = actor_loss.mean()\n",
    "            avg_loss += actor_loss.item()\n",
    "\n",
    "            '''更新策略网络'''\n",
    "            actor_optimizer.zero_grad()\n",
    "            actor_loss.backward()\n",
    "            actor_optimizer.step()\n",
    "            actor_cnt += 1\n",
    "            for param, target_param in zip(critic_1.parameters(), critic_target_1.parameters()):\n",
    "                target_param.data.copy_(tau * param.data + (1 - tau) * target_param.data)\n",
    "\n",
    "            for param, target_param in zip(critic_2.parameters(), critic_target_2.parameters()):\n",
    "                target_param.data.copy_(tau * param.data + (1 - tau) * target_param.data)\n",
    "\n",
    "            for param, target_param in zip(actor.parameters(), actor_target.parameters()):\n",
    "                target_param.data.copy_(tau * param.data + (1 - tau) * target_param.data)\n",
    "\n",
    "        cnt += 1\n",
    "    \n",
    "    view_losses.append(avg_loss / actor_cnt)\n",
    "         "
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 67,
   "id": "f02a67eb",
   "metadata": {},
   "outputs": [
    {
     "data": {
      "image/png": "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",
      "text/plain": [
       "<Figure size 2000x500 with 3 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('actor loss')\n",
    "plt.plot(view_losses)\n",
    "plt.show()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 68,
   "id": "d9de334b",
   "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/policy 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/policy/rl-video-episode-0.mp4.\n",
      "Moviepy - Writing video /Users/webberg/workspace/code/reinforce_learning/policy/rl-video-episode-0.mp4\n",
      "\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "                                                                                \r"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Moviepy - Done !\n",
      "Moviepy - video ready /Users/webberg/workspace/code/reinforce_learning/policy/rl-video-episode-0.mp4\n",
      "score:  -957.1748237814347 step: 201\n"
     ]
    }
   ],
   "source": [
    "test()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "id": "c44d92ac",
   "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
}