Тема_8_ДЗ_DDPG.ipynb

{
 "cells": [
  {
   "cell_type": "code",
   "execution_count": null,
   "id": "80ae5ac8-af3a-4044-b0b4-3a9da2e314b2",
   "metadata": {},
   "outputs": [],
   "source": [
    "import random\n",
    "import numpy as np\n",
    "import gymnasium as gym\n",
    "import torch\n",
    "import torch.nn as nn\n",
    "from torch.nn import functional as F\n",
    "import matplotlib.pyplot as plt\n",
    "from tqdm import tqdm "
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 2,
   "id": "e4690d26-5a43-4177-a0d0-94a2563523c5",
   "metadata": {},
   "outputs": [],
   "source": [
    "env_gym = gym.make('Pendulum-v1', g=9.81)"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "7193d2c2-2beb-4cf1-8ae1-b4902c9f75f0",
   "metadata": {},
   "source": [
    "## Original"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 3,
   "id": "ae781d66-2c8e-482e-9cad-0319e20b5ef2",
   "metadata": {},
   "outputs": [],
   "source": [
    "class QNet(nn.Module):\n",
    "    def __init__(self, hidden_dim=64, n_in=4, n_out=1):\n",
    "        super().__init__()\n",
    "\n",
    "        self.hidden = nn.Linear(n_in, hidden_dim)\n",
    "        self.output = nn.Linear(hidden_dim, n_out)\n",
    "\n",
    "    def forward(self, s, a):\n",
    "        outs = torch.concat((s, a), dim=-1)\n",
    "        outs = self.hidden(outs)\n",
    "        outs = F.relu(outs)\n",
    "        outs = self.output(outs)\n",
    "        return outs\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 4,
   "id": "7983d870-a17c-43ba-b74e-c2592ab6edf2",
   "metadata": {},
   "outputs": [],
   "source": [
    "class PolicyNet(nn.Module):\n",
    "    def __init__(self, hidden_dim=64, n_states=3, n_actions=1):\n",
    "        super().__init__()\n",
    "\n",
    "        self.hidden = nn.Linear(n_states, hidden_dim)\n",
    "        self.output = nn.Linear(hidden_dim, n_actions)\n",
    "\n",
    "    def forward(self, s):\n",
    "        outs = self.hidden(s)\n",
    "        outs = F.relu(outs)\n",
    "        outs = self.output(outs)\n",
    "        outs = torch.tanh(outs)\n",
    "        return outs\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 5,
   "id": "ca10be4b-e2cb-4b25-865f-b33f36b203ec",
   "metadata": {},
   "outputs": [],
   "source": [
    "def optimize(states, actions, rewards, next_states, dones):\n",
    "    # Convert to tensor\n",
    "    states = torch.tensor(np.array(states), dtype=torch.float).to(device)\n",
    "    actions = torch.tensor(np.array(actions), dtype=torch.float).to(device)\n",
    "    rewards = torch.tensor(np.array(rewards), dtype=torch.float).to(device)\n",
    "    rewards = rewards.unsqueeze(dim=1)\n",
    "    next_states = torch.tensor(np.array(next_states), dtype=torch.float).to(device)\n",
    "    dones = torch.tensor(np.array(dones), dtype=torch.float).to(device)\n",
    "    dones = dones.unsqueeze(dim=1)\n",
    "\n",
    "    # Optimize critic loss\n",
    "    opt_q.zero_grad()\n",
    "    q_org = q_origin_model(states, actions)\n",
    "    mu_tgt_next = mu_target_model(next_states)\n",
    "    q_tgt_next = q_target_model(next_states, mu_tgt_next)\n",
    "    q_tgt = rewards + gamma * (1.0 - dones) * q_tgt_next\n",
    "    loss_q = F.mse_loss(\n",
    "        q_org,\n",
    "        q_tgt,\n",
    "        reduction=\"none\")\n",
    "    loss_q.sum().backward()\n",
    "    opt_q.step()\n",
    "\n",
    "    # Optimize actor loss\n",
    "    opt_mu.zero_grad()\n",
    "    mu_org = mu_origin_model(states)\n",
    "    for p in q_origin_model.parameters():\n",
    "        p.requires_grad = False # disable grad in q_origin_model before computation\n",
    "    q_tgt_max = q_origin_model(states, mu_org)\n",
    "    (-q_tgt_max).sum().backward()\n",
    "    opt_mu.step()\n",
    "    for p in q_origin_model.parameters():\n",
    "        p.requires_grad = True # enable grad again"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 6,
   "id": "5f7d8fab-afb5-4f7f-9950-a2736f54e6c5",
   "metadata": {},
   "outputs": [],
   "source": [
    "tau = 0.002\n",
    "\n",
    "def update_target():\n",
    "    for var, var_target in zip(q_origin_model.parameters(), q_target_model.parameters()):\n",
    "        var_target.data = tau * var.data + (1.0 - tau) * var_target.data\n",
    "    \n",
    "    for var, var_target in zip(mu_origin_model.parameters(), mu_target_model.parameters()):\n",
    "        var_target.data = tau * var.data + (1.0 - tau) * var_target.data"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 7,
   "id": "663f8ae0-610a-4c77-9cbc-181e9de5fa28",
   "metadata": {},
   "outputs": [],
   "source": [
    "class replayBuffer:\n",
    "    def __init__(self, buffer_size: int):\n",
    "        self.buffer_size = buffer_size\n",
    "        self.buffer = []\n",
    "        self._next_idx = 0\n",
    "\n",
    "    def add(self, item):\n",
    "        if len(self.buffer) > self._next_idx:\n",
    "            self.buffer[self._next_idx] = item\n",
    "        else:\n",
    "            self.buffer.append(item)\n",
    "        if self._next_idx == self.buffer_size - 1:\n",
    "            self._next_idx = 0\n",
    "        else:\n",
    "            self._next_idx = self._next_idx + 1\n",
    "\n",
    "    def sample(self, batch_size):\n",
    "        indices = [random.randint(0, len(self.buffer) - 1) for _ in range(batch_size)]\n",
    "        states   = [self.buffer[i][0] for i in indices]\n",
    "        actions  = [self.buffer[i][1] for i in indices]\n",
    "        rewards  = [self.buffer[i][2] for i in indices]\n",
    "        n_states = [self.buffer[i][3] for i in indices]\n",
    "        dones    = [self.buffer[i][4] for i in indices]\n",
    "        return states, actions, rewards, n_states, dones\n",
    "\n",
    "    def length(self):\n",
    "        return len(self.buffer)\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 16,
   "id": "2626d38a-94fe-40e1-8b77-cb5da2e5f3d4",
   "metadata": {},
   "outputs": [],
   "source": [
    "\"\"\"\n",
    "Ornstein-Uhlenbeck noise implemented by OpenAI\n",
    "Copied from https://github.com/openai/baselines/blob/master/baselines/ddpg/noise.py\n",
    "\"\"\"\n",
    "class OrnsteinUhlenbeckActionNoise:\n",
    "    def __init__(self, mu, sigma, theta=.15, dt=1e-2, x0=None):\n",
    "        self.theta = theta\n",
    "        self.mu = mu\n",
    "        self.sigma = sigma\n",
    "        self.dt = dt\n",
    "        self.x0 = x0\n",
    "        self.reset()\n",
    "\n",
    "    def __call__(self):\n",
    "        x = self.x_prev + self.theta * (self.mu - self.x_prev) * self.dt + \\\n",
    "            self.sigma * np.sqrt(self.dt) * np.random.normal(size=self.mu.shape)\n",
    "        self.x_prev = x\n",
    "        return x\n",
    "\n",
    "    def reset(self):\n",
    "        self.x_prev = self.x0 if self.x0 is not None else np.zeros_like(self.mu)\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 9,
   "id": "cd2a1488-97de-4b01-b824-38cb49f95205",
   "metadata": {},
   "outputs": [],
   "source": [
    "# pick up action with Ornstein-Uhlenbeck noise\n",
    "def pick_sample(s):\n",
    "    with torch.no_grad():\n",
    "        # s = np.array(s)\n",
    "        s_batch = np.expand_dims(s, axis=0)\n",
    "        s_batch = torch.tensor(s_batch, dtype=torch.float).to(device)\n",
    "        action_det = mu_origin_model(s_batch)\n",
    "        action_det = action_det.squeeze(dim=1)\n",
    "        noise = ou_action_noise()\n",
    "        action = action_det.cpu().numpy() + noise\n",
    "        action = np.clip(action, -1.0, 1.0)\n",
    "        return np.array([float(action.item())])"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 10,
   "id": "b82a43e8-86d9-4903-9d65-fb7020c47cf1",
   "metadata": {},
   "outputs": [],
   "source": [
    "# Create objects\n",
    "\n",
    "device = torch.device(\"cuda\" if torch.cuda.is_available() else \"cpu\")\n",
    "\n",
    "q_origin_model = QNet().to(device)  # Q_phi\n",
    "q_target_model = QNet().to(device)  # Q_phi'\n",
    "_ = q_target_model.requires_grad_(False)  # target model doen't need grad\n",
    "\n",
    "mu_origin_model = PolicyNet().to(device)  # mu_theta\n",
    "mu_target_model = PolicyNet().to(device)  # mu_theta'\n",
    "_ = mu_target_model.requires_grad_(False)  # target model doen't need grad\n",
    "\n",
    "gamma = 0.99\n",
    "opt_q = torch.optim.AdamW(q_origin_model.parameters(), lr=0.0005)\n",
    "opt_mu = torch.optim.AdamW(mu_origin_model.parameters(), lr=0.0005)\n",
    "\n",
    "buffer = replayBuffer(buffer_size=20000)\n",
    "\n",
    "ou_action_noise = OrnsteinUhlenbeckActionNoise(mu=np.zeros(1), sigma=np.ones(1) * 0.05)"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "75e891c7-b42b-4c1d-a464-cd112aa4a978",
   "metadata": {},
   "source": [
    "## Main loop"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 11,
   "id": "8ba7f815-7624-4acd-bbaa-e7c3766c2289",
   "metadata": {
    "scrolled": true
   },
   "outputs": [
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "100%|█████████████████████████████████████████████████████████████████████████████████████████████████████████████████████████████████████████████████████████████████████████████| 1000/1000 [03:41<00:00,  4.51it/s]"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "\n",
      "Done\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "\n"
     ]
    }
   ],
   "source": [
    "batch_size = 250\n",
    "\n",
    "reward_records = []\n",
    "for i in tqdm(range(1000)):\n",
    "    # Run episode till done\n",
    "    s = env_gym.reset()[0]\n",
    "    done = False\n",
    "    cum_reward = 0\n",
    "    while not done:\n",
    "        a = pick_sample(s)\n",
    "        s_next, r, done, truncate, _ = env_gym.step(a)\n",
    "        done = done or truncate\n",
    "        buffer.add([s, a, r, s_next, float(done)])\n",
    "        cum_reward += r\n",
    "    \n",
    "        # Train (optimize parameters)\n",
    "        if buffer.length() >= batch_size:\n",
    "            states, actions, rewards, n_states, dones = buffer.sample(batch_size)\n",
    "            optimize(states, actions, rewards, n_states, dones)\n",
    "            update_target()\n",
    "        s = s_next\n",
    "    \n",
    "    # Output total rewards in episode (max 500)\n",
    "    # print(\"Run episode{} with rewards {}\".format(i, cum_reward), end=\"\\r\")\n",
    "    reward_records.append(cum_reward)\n",
    "\n",
    "    # stop if reward mean > 475.0\n",
    "    if np.average(reward_records[-50:]) > 475.0:\n",
    "        break\n",
    "\n",
    "print(\"\\nDone\")"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 12,
   "id": "63f2d044-6874-40ba-89ee-beb6906dbe8a",
   "metadata": {},
   "outputs": [
    {
     "data": {
      "image/png": "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",
      "text/plain": [
       "<Figure size 640x480 with 1 Axes>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "# Generate recent 50 interval average\n",
    "average_reward = []\n",
    "for idx in range(len(reward_records)):\n",
    "    avg_list = np.empty(shape=(1,), dtype=int)\n",
    "    if idx < 50:\n",
    "        avg_list = reward_records[:idx+1]\n",
    "    else:\n",
    "        avg_list = reward_records[idx-49:idx+1]\n",
    "    average_reward.append(np.average(avg_list))\n",
    "plt.plot(reward_records, label='reward')\n",
    "plt.plot(average_reward, label='average reward')\n",
    "plt.xlabel('N episodes')\n",
    "plt.ylabel('Reward')\n",
    "plt.legend()\n",
    "plt.show()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "id": "dd3a3818-0957-4e4c-bd5e-cd611e733a2e",
   "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": "3.10.4"
  }
 },
 "nbformat": 4,
 "nbformat_minor": 5
}