tutorial/Tutorial6_DDQN.ipynb

{
   "cells": [
      {
         "cell_type": "markdown",
         "metadata": {},
         "source": [
            "# Tutorial 6\n",
            "We show how to train Double DQN for high frequency trading on Bitcoin."
         ]
      },
      {
         "cell_type": "markdown",
         "metadata": {},
         "source": [
            "## Step 1: Import Packages\n",
            "Modify the system path and load the corresponding packages and functions "
         ]
      },
      {
         "cell_type": "code",
         "execution_count": 1,
         "metadata": {},
         "outputs": [
            {
               "name": "stderr",
               "output_type": "stream",
               "text": [
                  "2023-03-04 19:13:04,200\tINFO services.py:1476 -- View the Ray dashboard at \u001b[1m\u001b[32mhttp://127.0.0.1:8270\u001b[39m\u001b[22m\n",
                  "2023-03-04 19:13:06,268\tINFO worker.py:973 -- Calling ray.init() again after it has already been called.\n"
               ]
            }
         ],
         "source": [
            "import os\n",
            "import sys\n",
            "from pathlib import Path\n",
            "import warnings\n",
            "warnings.filterwarnings(\"ignore\")\n",
            "\n",
            "ROOT = os.path.dirname(os.path.abspath(\".\"))\n",
            "sys.path.append(ROOT)\n",
            "\n",
            "import torch\n",
            "import argparse\n",
            "import os.path as osp\n",
            "from mmcv import Config\n",
            "from trademaster.utils import replace_cfg_vals\n",
            "from trademaster.nets.builder import build_net\n",
            "from trademaster.environments.builder import build_environment\n",
            "from trademaster.datasets.builder import build_dataset\n",
            "from trademaster.agents.builder import build_agent\n",
            "from trademaster.optimizers.builder import build_optimizer\n",
            "from trademaster.losses.builder import build_loss\n",
            "from trademaster.trainers.builder import build_trainer\n",
            "from trademaster.transition.builder import build_transition\n",
            "from trademaster.utils import plot\n",
            "from trademaster.utils import set_seed\n",
            "set_seed(2023)"
         ]
      },
      {
         "cell_type": "markdown",
         "metadata": {},
         "source": [
            "## Step 2: Load Configs\n",
            "Load default config from the file `configs/high_frequency_trading/high_frequency_trading_BTC_dqn_dqn_adam_mse.py`"
         ]
      },
      {
         "cell_type": "code",
         "execution_count": 2,
         "metadata": {},
         "outputs": [],
         "source": [
            "parser = argparse.ArgumentParser(description='Download Alpaca Datasets')\n",
            "parser.add_argument(\"--config\", default=osp.join(ROOT, \"configs\", \"high_frequency_trading\", \"high_frequency_trading_BTC_dqn_dqn_adam_mse.py\"),\n",
            "                    help=\"download datasets config file path\")\n",
            "parser.add_argument(\"--task_name\", type=str, default=\"train\")\n",
            "\n",
            "args, _ = parser.parse_known_args()\n",
            "cfg = Config.fromfile(args.config)\n",
            "task_name = args.task_name\n",
            "cfg = replace_cfg_vals(cfg)"
         ]
      },
      {
         "cell_type": "code",
         "execution_count": 3,
         "metadata": {},
         "outputs": [
            {
               "data": {
                  "text/plain": [
                     "Config (path: /data/sunshuo/qml/TradeMaster/configs/high_frequency_trading/high_frequency_trading_BTC_dqn_dqn_adam_mse.py): {'data': {'type': 'HighFrequencyTradingDataset', 'data_path': 'data/high_frequency_trading/small_BTC', 'train_path': 'data/high_frequency_trading/small_BTC/train.csv', 'valid_path': 'data/high_frequency_trading/small_BTC/valid.csv', 'test_path': 'data/high_frequency_trading/small_BTC/test.csv', 'test_dynamic_path': 'data/high_frequency_trading/small_BTC/test_labeled_3_300_-0.01_0.01.csv', 'tech_indicator_list': ['imblance_volume_oe', 'sell_spread_oe', 'buy_spread_oe', 'kmid2', 'bid1_size_n', 'ksft2', 'ma_10', 'ksft', 'kmid', 'ask1_size_n', 'trade_diff', 'qtlu_10', 'qtld_10', 'cntd_10', 'beta_10', 'roc_10', 'bid5_size_n', 'rsv_10', 'imxd_10', 'ask5_size_n', 'ma_30', 'max_10', 'qtlu_30', 'imax_10', 'imin_10', 'min_10', 'qtld_30', 'cntn_10', 'rsv_30', 'cntp_10', 'ma_60', 'max_30', 'qtlu_60', 'qtld_60', 'cntd_30', 'roc_30', 'beta_30', 'bid4_size_n', 'rsv_60', 'ask4_size_n', 'imxd_30', 'min_30', 'max_60', 'imax_30', 'imin_30', 'cntd_60', 'roc_60', 'beta_60', 'cntn_30', 'min_60', 'cntp_30', 'bid3_size_n', 'imxd_60', 'ask3_size_n', 'sell_volume_oe', 'imax_60', 'imin_60', 'cntn_60', 'buy_volume_oe', 'cntp_60', 'bid2_size_n', 'kup', 'bid1_size', 'ask1_size', 'std_30', 'ask2_size_n'], 'transcation_cost': 0, 'backward_num_timestamp': 1, 'max_holding_number': 0.01, 'num_action': 11, 'max_punish': 1000000000000.0, 'episode_length': 14400}, 'environment': {'type': 'HighFrequencyTradingEnvironment'}, 'agent': {'type': 'HighFrequencyTradingDDQN', 'auxiliary_coffient': 512, 'reward_scale': 1, 'repeat_times': 1, 'gamma': 0.99, 'batch_size': 64, 'clip_grad_norm': 3.0, 'soft_update_tau': 0, 'state_value_tau': 0.005}, 'trainer': {'type': 'HighFrequencyTradingTrainer', 'epochs': 10, 'work_dir': 'work_dir/high_frequency_trading_BTC_high_frequency_trading_dqn_ddqn_adam_mse', 'seeds': 12345, 'batch_size': 512, 'horizon_len': 512, 'buffer_size': 100000.0, 'num_threads': 8, 'if_remove': False, 'if_discrete': True, 'if_off_policy': True, 'if_keep_save': True, 'if_over_write': False, 'if_save_buffer': False}, 'loss': {'type': 'HFTLoss', 'ada': 1}, 'optimizer': {'type': 'Adam', 'lr': 0.001}, 'act': {'type': 'HFTQNet', 'state_dim': 66, 'action_dim': 11, 'dims': 16, 'explore_rate': 0.01, 'max_punish': 0}, 'cri': None, 'task_name': 'high_frequency_trading', 'dataset_name': 'BTC', 'optimizer_name': 'adam', 'loss_name': 'mse', 'auxiliry_loss_name': 'KLdiv', 'net_name': 'high_frequency_trading_dqn', 'agent_name': 'ddqn', 'work_dir': 'work_dir/high_frequency_trading_BTC_high_frequency_trading_dqn_ddqn_adam_mse', 'batch_size': 512, 'train_environment': {'type': 'HighFrequencyTradingTrainingEnvironment'}}"
                  ]
               },
               "execution_count": 3,
               "metadata": {},
               "output_type": "execute_result"
            }
         ],
         "source": [
            "cfg"
         ]
      },
      {
         "cell_type": "markdown",
         "metadata": {},
         "source": [
            "## Step 3: Build Dataset"
         ]
      },
      {
         "cell_type": "code",
         "execution_count": 4,
         "metadata": {},
         "outputs": [],
         "source": [
            "dataset = build_dataset(cfg)"
         ]
      },
      {
         "cell_type": "markdown",
         "metadata": {},
         "source": [
            "## Step 4: Build Reinforcement Learning Environments\n",
            "Build environments based on cfg and predefined dataset\n",
            "\n"
         ]
      },
      {
         "cell_type": "code",
         "execution_count": 5,
         "metadata": {},
         "outputs": [],
         "source": [
            "valid_environment = build_environment(cfg, default_args=dict(dataset=dataset, task=\"valid\"))\n",
            "test_environment = build_environment(cfg, default_args=dict(dataset=dataset, task=\"test\"))\n",
            "cfg.environment = cfg.train_environment\n",
            "train_environment = build_environment(cfg, default_args=dict(dataset=dataset, task=\"train\"))"
         ]
      },
      {
         "cell_type": "code",
         "execution_count": 6,
         "metadata": {},
         "outputs": [
            {
               "data": {
                  "text/html": [
                     "<div>\n",
                     "<style scoped>\n",
                     "    .dataframe tbody tr th:only-of-type {\n",
                     "        vertical-align: middle;\n",
                     "    }\n",
                     "\n",
                     "    .dataframe tbody tr th {\n",
                     "        vertical-align: top;\n",
                     "    }\n",
                     "\n",
                     "    .dataframe thead th {\n",
                     "        text-align: right;\n",
                     "    }\n",
                     "</style>\n",
                     "<table border=\"1\" class=\"dataframe\">\n",
                     "  <thead>\n",
                     "    <tr style=\"text-align: right;\">\n",
                     "      <th></th>\n",
                     "      <th>bid1_price</th>\n",
                     "      <th>bid1_size</th>\n",
                     "      <th>bid2_price</th>\n",
                     "      <th>bid2_size</th>\n",
                     "      <th>bid3_price</th>\n",
                     "      <th>bid3_size</th>\n",
                     "      <th>bid4_price</th>\n",
                     "      <th>bid4_size</th>\n",
                     "      <th>bid5_price</th>\n",
                     "      <th>bid5_size</th>\n",
                     "      <th>...</th>\n",
                     "      <th>imxd_60</th>\n",
                     "      <th>cntp_10</th>\n",
                     "      <th>cntp_30</th>\n",
                     "      <th>cntp_60</th>\n",
                     "      <th>cntn_10</th>\n",
                     "      <th>cntn_30</th>\n",
                     "      <th>cntn_60</th>\n",
                     "      <th>cntd_10</th>\n",
                     "      <th>cntd_30</th>\n",
                     "      <th>cntd_60</th>\n",
                     "    </tr>\n",
                     "  </thead>\n",
                     "  <tbody>\n",
                     "    <tr>\n",
                     "      <th>0</th>\n",
                     "      <td>23794.3</td>\n",
                     "      <td>0.900828</td>\n",
                     "      <td>23792.9</td>\n",
                     "      <td>0.018588</td>\n",
                     "      <td>23792.8</td>\n",
                     "      <td>0.007388</td>\n",
                     "      <td>23792.1</td>\n",
                     "      <td>0.042018</td>\n",
                     "      <td>23792.0</td>\n",
                     "      <td>0.000300</td>\n",
                     "      <td>...</td>\n",
                     "      <td>1.193895</td>\n",
                     "      <td>0.268138</td>\n",
                     "      <td>0.963686</td>\n",
                     "      <td>0.800684</td>\n",
                     "      <td>-0.418841</td>\n",
                     "      <td>-0.017993</td>\n",
                     "      <td>-0.189727</td>\n",
                     "      <td>0.469470</td>\n",
                     "      <td>0.745885</td>\n",
                     "      <td>0.877378</td>\n",
                     "    </tr>\n",
                     "    <tr>\n",
                     "      <th>1</th>\n",
                     "      <td>23794.3</td>\n",
                     "      <td>0.000878</td>\n",
                     "      <td>23793.2</td>\n",
                     "      <td>0.014988</td>\n",
                     "      <td>23793.1</td>\n",
                     "      <td>0.147063</td>\n",
                     "      <td>23792.8</td>\n",
                     "      <td>0.007388</td>\n",
                     "      <td>23792.0</td>\n",
                     "      <td>0.000300</td>\n",
                     "      <td>...</td>\n",
                     "      <td>1.193895</td>\n",
                     "      <td>0.268138</td>\n",
                     "      <td>0.963686</td>\n",
                     "      <td>0.800684</td>\n",
                     "      <td>-1.027695</td>\n",
                     "      <td>-0.017993</td>\n",
                     "      <td>-0.358007</td>\n",
                     "      <td>0.891105</td>\n",
                     "      <td>0.745885</td>\n",
                     "      <td>1.031828</td>\n",
                     "    </tr>\n",
                     "    <tr>\n",
                     "      <th>2</th>\n",
                     "      <td>23794.3</td>\n",
                     "      <td>0.000878</td>\n",
                     "      <td>23793.2</td>\n",
                     "      <td>0.014988</td>\n",
                     "      <td>23792.8</td>\n",
                     "      <td>0.007388</td>\n",
                     "      <td>23792.0</td>\n",
                     "      <td>0.000300</td>\n",
                     "      <td>23791.8</td>\n",
                     "      <td>0.147071</td>\n",
                     "      <td>...</td>\n",
                     "      <td>1.139902</td>\n",
                     "      <td>0.268138</td>\n",
                     "      <td>1.257798</td>\n",
                     "      <td>0.976534</td>\n",
                     "      <td>-1.027695</td>\n",
                     "      <td>-0.017993</td>\n",
                     "      <td>-0.526286</td>\n",
                     "      <td>0.891105</td>\n",
                     "      <td>0.969180</td>\n",
                     "      <td>1.340727</td>\n",
                     "    </tr>\n",
                     "    <tr>\n",
                     "      <th>3</th>\n",
                     "      <td>23794.3</td>\n",
                     "      <td>0.000878</td>\n",
                     "      <td>23793.2</td>\n",
                     "      <td>0.014988</td>\n",
                     "      <td>23792.8</td>\n",
                     "      <td>0.007388</td>\n",
                     "      <td>23792.0</td>\n",
                     "      <td>0.000300</td>\n",
                     "      <td>23791.8</td>\n",
                     "      <td>0.147071</td>\n",
                     "      <td>...</td>\n",
                     "      <td>1.139902</td>\n",
                     "      <td>0.268138</td>\n",
                     "      <td>1.257798</td>\n",
                     "      <td>0.800684</td>\n",
                     "      <td>-1.027695</td>\n",
                     "      <td>-0.017993</td>\n",
                     "      <td>-0.526286</td>\n",
                     "      <td>0.891105</td>\n",
                     "      <td>0.969180</td>\n",
                     "      <td>1.186278</td>\n",
                     "    </tr>\n",
                     "    <tr>\n",
                     "      <th>4</th>\n",
                     "      <td>23794.3</td>\n",
                     "      <td>0.000878</td>\n",
                     "      <td>23793.2</td>\n",
                     "      <td>0.014988</td>\n",
                     "      <td>23792.8</td>\n",
                     "      <td>0.007388</td>\n",
                     "      <td>23792.0</td>\n",
                     "      <td>0.000300</td>\n",
                     "      <td>23791.8</td>\n",
                     "      <td>0.147071</td>\n",
                     "      <td>...</td>\n",
                     "      <td>1.112905</td>\n",
                     "      <td>0.268138</td>\n",
                     "      <td>0.963686</td>\n",
                     "      <td>0.800684</td>\n",
                     "      <td>-1.027695</td>\n",
                     "      <td>-0.017993</td>\n",
                     "      <td>-0.694566</td>\n",
                     "      <td>0.891105</td>\n",
                     "      <td>0.745885</td>\n",
                     "      <td>1.340727</td>\n",
                     "    </tr>\n",
                     "  </tbody>\n",
                     "</table>\n",
                     "<p>5 rows × 96 columns</p>\n",
                     "</div>"
                  ],
                  "text/plain": [
                     "   bid1_price  bid1_size  bid2_price  bid2_size  bid3_price  bid3_size  \\\n",
                     "0     23794.3   0.900828     23792.9   0.018588     23792.8   0.007388   \n",
                     "1     23794.3   0.000878     23793.2   0.014988     23793.1   0.147063   \n",
                     "2     23794.3   0.000878     23793.2   0.014988     23792.8   0.007388   \n",
                     "3     23794.3   0.000878     23793.2   0.014988     23792.8   0.007388   \n",
                     "4     23794.3   0.000878     23793.2   0.014988     23792.8   0.007388   \n",
                     "\n",
                     "   bid4_price  bid4_size  bid5_price  bid5_size  ...   imxd_60   cntp_10  \\\n",
                     "0     23792.1   0.042018     23792.0   0.000300  ...  1.193895  0.268138   \n",
                     "1     23792.8   0.007388     23792.0   0.000300  ...  1.193895  0.268138   \n",
                     "2     23792.0   0.000300     23791.8   0.147071  ...  1.139902  0.268138   \n",
                     "3     23792.0   0.000300     23791.8   0.147071  ...  1.139902  0.268138   \n",
                     "4     23792.0   0.000300     23791.8   0.147071  ...  1.112905  0.268138   \n",
                     "\n",
                     "    cntp_30   cntp_60   cntn_10   cntn_30   cntn_60   cntd_10   cntd_30  \\\n",
                     "0  0.963686  0.800684 -0.418841 -0.017993 -0.189727  0.469470  0.745885   \n",
                     "1  0.963686  0.800684 -1.027695 -0.017993 -0.358007  0.891105  0.745885   \n",
                     "2  1.257798  0.976534 -1.027695 -0.017993 -0.526286  0.891105  0.969180   \n",
                     "3  1.257798  0.800684 -1.027695 -0.017993 -0.526286  0.891105  0.969180   \n",
                     "4  0.963686  0.800684 -1.027695 -0.017993 -0.694566  0.891105  0.745885   \n",
                     "\n",
                     "    cntd_60  \n",
                     "0  0.877378  \n",
                     "1  1.031828  \n",
                     "2  1.340727  \n",
                     "3  1.186278  \n",
                     "4  1.340727  \n",
                     "\n",
                     "[5 rows x 96 columns]"
                  ]
               },
               "execution_count": 6,
               "metadata": {},
               "output_type": "execute_result"
            }
         ],
         "source": [
            "train_environment.df.head()"
         ]
      },
      {
         "cell_type": "markdown",
         "metadata": {},
         "source": [
            "## Step 5: Build Net \n",
            "Update information about the state and action dimension. Create networks and optimizer for Double DQN\n"
         ]
      },
      {
         "cell_type": "code",
         "execution_count": 7,
         "metadata": {},
         "outputs": [],
         "source": [
            "action_dim = train_environment.action_dim\n",
            "state_dim = train_environment.state_dim\n",
            "\n",
            "cfg.act.update(dict(action_dim=action_dim, state_dim=state_dim))\n",
            "act = build_net(cfg.act)\n",
            "act_optimizer = build_optimizer(cfg, default_args=dict(params=act.parameters()))\n",
            "if cfg.cri:\n",
            "    cfg.cri.update(dict(action_dim=action_dim, state_dim=state_dim))\n",
            "    cri = build_net(cfg.cri)\n",
            "    cri_optimizer = build_optimizer(cfg, default_args=dict(params=cri.parameters()))\n",
            "else:\n",
            "    cri = None\n",
            "    cri_optimizer = None"
         ]
      },
      {
         "cell_type": "markdown",
         "metadata": {},
         "source": [
            "## Step 6: Build Loss"
         ]
      },
      {
         "cell_type": "code",
         "execution_count": 8,
         "metadata": {},
         "outputs": [],
         "source": [
            "criterion = build_loss(cfg)"
         ]
      },
      {
         "cell_type": "markdown",
         "metadata": {},
         "source": [
            "## Step 7: Build Agent"
         ]
      },
      {
         "cell_type": "code",
         "execution_count": 9,
         "metadata": {},
         "outputs": [],
         "source": [
            "device = torch.device('cuda' if torch.cuda.is_available() else 'cpu')\n",
            "agent = build_agent(cfg,default_args=dict(action_dim=action_dim,state_dim=state_dim,act=act,cri=cri,act_optimizer=act_optimizer,cri_optimizer=cri_optimizer,criterion=criterion,device=device))\n"
         ]
      },
      {
         "cell_type": "markdown",
         "metadata": {},
         "source": [
            "## Step 8: Build Trainer\n",
            "Build trainer from config and create work directionary to save the result, model and config"
         ]
      },
      {
         "cell_type": "code",
         "execution_count": 10,
         "metadata": {},
         "outputs": [
            {
               "name": "stdout",
               "output_type": "stream",
               "text": [
                  "| Arguments Keep work_dir: /data/sunshuo/qml/TradeMaster/work_dir/high_frequency_trading_BTC_high_frequency_trading_dqn_ddqn_adam_mse\n"
               ]
            }
         ],
         "source": [
            "\n",
            "trainer = build_trainer(cfg, default_args=dict(train_environment=train_environment,valid_environment=valid_environment,test_environment=test_environment,agent=agent,device=device))\n",
            "\n",
            "cfg.dump(osp.join(ROOT, cfg.work_dir, osp.basename(args.config)))"
         ]
      },
      {
         "cell_type": "markdown",
         "metadata": {},
         "source": [
            "## Step 9: RL Agent Training\n",
            "Train the Double DQN agent based on the config and save results in workdir"
         ]
      },
      {
         "cell_type": "code",
         "execution_count": 11,
         "metadata": {},
         "outputs": [
            {
               "name": "stdout",
               "output_type": "stream",
               "text": [
                  "Train Episode: [1/10]\n",
                  "+---------------+--------------+------------+--------------+\n",
                  "| Profit Margin | Sharp Ratio  | Volatility | Max Drawdown |\n",
                  "+---------------+--------------+------------+--------------+\n",
                  "|  -19.710863%  | -2792.392715 | 0.003066%  |  19.713213%  |\n",
                  "+---------------+--------------+------------+--------------+\n",
                  "+---------------+--------------+------------+--------------+\n",
                  "| Profit Margin | Sharp Ratio  | Volatility | Max Drawdown |\n",
                  "+---------------+--------------+------------+--------------+\n",
                  "|  -17.955225%  | -2142.013586 | 0.003603%  |  17.979934%  |\n",
                  "+---------------+--------------+------------+--------------+\n",
                  "Valid Episode: [1/10]\n",
                  "+--------------+------------+--------------+\n",
                  "| Total Return | Volatility | Max Drawdown |\n",
                  "+--------------+------------+--------------+\n",
                  "| -45.740756%  | 0.005121%  |  45.740756%  |\n",
                  "+--------------+------------+--------------+\n",
                  "Train Episode: [2/10]\n",
                  "+---------------+--------------+------------+--------------+\n",
                  "| Profit Margin | Sharp Ratio  | Volatility | Max Drawdown |\n",
                  "+---------------+--------------+------------+--------------+\n",
                  "|  -35.120112%  | -1486.029697 | 0.011352%  |  35.122955%  |\n",
                  "+---------------+--------------+------------+--------------+\n",
                  "+---------------+--------------+------------+--------------+\n",
                  "| Profit Margin | Sharp Ratio  | Volatility | Max Drawdown |\n",
                  "+---------------+--------------+------------+--------------+\n",
                  "|  -24.667449%  | -1787.568819 | 0.006179%  |  24.684778%  |\n",
                  "+---------------+--------------+------------+--------------+\n",
                  "+---------------+--------------+------------+--------------+\n",
                  "| Profit Margin | Sharp Ratio  | Volatility | Max Drawdown |\n",
                  "+---------------+--------------+------------+--------------+\n",
                  "|  -13.269536%  | -1018.163059 | 0.005452%  |  13.269536%  |\n",
                  "+---------------+--------------+------------+--------------+\n",
                  "Valid Episode: [2/10]\n",
                  "+--------------+------------+--------------+\n",
                  "| Total Return | Volatility | Max Drawdown |\n",
                  "+--------------+------------+--------------+\n",
                  "| -13.725643%  | 0.004899%  |  13.823969%  |\n",
                  "+--------------+------------+--------------+\n",
                  "Train Episode: [3/10]\n",
                  "+---------------+-------------+------------+--------------+\n",
                  "| Profit Margin | Sharp Ratio | Volatility | Max Drawdown |\n",
                  "+---------------+-------------+------------+--------------+\n",
                  "|   -7.471570%  | -690.985539 | 0.004382%  |  7.586687%   |\n",
                  "+---------------+-------------+------------+--------------+\n",
                  "+---------------+-------------+------------+--------------+\n",
                  "| Profit Margin | Sharp Ratio | Volatility | Max Drawdown |\n",
                  "+---------------+-------------+------------+--------------+\n",
                  "|   -6.185789%  | -377.008630 | 0.006602%  |  6.185789%   |\n",
                  "+---------------+-------------+------------+--------------+\n",
                  "+---------------+-------------+------------+--------------+\n",
                  "| Profit Margin | Sharp Ratio | Volatility | Max Drawdown |\n",
                  "+---------------+-------------+------------+--------------+\n",
                  "|  -10.658422%  | -993.021710 | 0.004426%  |  10.658422%  |\n",
                  "+---------------+-------------+------------+--------------+\n",
                  "Valid Episode: [3/10]\n",
                  "+--------------+------------+--------------+\n",
                  "| Total Return | Volatility | Max Drawdown |\n",
                  "+--------------+------------+--------------+\n",
                  "| -13.170535%  | 0.005780%  |  13.239935%  |\n",
                  "+--------------+------------+--------------+\n",
                  "Train Episode: [4/10]\n",
                  "+---------------+-------------+------------+--------------+\n",
                  "| Profit Margin | Sharp Ratio | Volatility | Max Drawdown |\n",
                  "+---------------+-------------+------------+--------------+\n",
                  "|   -4.590676%  | -275.820456 | 0.006640%  |  5.162691%   |\n",
                  "+---------------+-------------+------------+--------------+\n",
                  "+---------------+-------------+------------+--------------+\n",
                  "| Profit Margin | Sharp Ratio | Volatility | Max Drawdown |\n",
                  "+---------------+-------------+------------+--------------+\n",
                  "|   -7.181079%  | -597.487064 | 0.004863%  |  7.347568%   |\n",
                  "+---------------+-------------+------------+--------------+\n",
                  "+---------------+-------------+------------+--------------+\n",
                  "| Profit Margin | Sharp Ratio | Volatility | Max Drawdown |\n",
                  "+---------------+-------------+------------+--------------+\n",
                  "|   -4.068303%  | -187.355544 | 0.008635%  |  4.444695%   |\n",
                  "+---------------+-------------+------------+--------------+\n",
                  "Valid Episode: [4/10]\n",
                  "+--------------+------------+--------------+\n",
                  "| Total Return | Volatility | Max Drawdown |\n",
                  "+--------------+------------+--------------+\n",
                  "|  -1.372845%  | 0.004237%  |  1.707928%   |\n",
                  "+--------------+------------+--------------+\n",
                  "Train Episode: [5/10]\n",
                  "+---------------+-------------+------------+--------------+\n",
                  "| Profit Margin | Sharp Ratio | Volatility | Max Drawdown |\n",
                  "+---------------+-------------+------------+--------------+\n",
                  "|   -6.072037%  | -600.602516 | 0.004067%  |  6.072037%   |\n",
                  "+---------------+-------------+------------+--------------+\n",
                  "+---------------+-------------+------------+--------------+\n",
                  "| Profit Margin | Sharp Ratio | Volatility | Max Drawdown |\n",
                  "+---------------+-------------+------------+--------------+\n",
                  "|   -4.561090%  | -308.007296 | 0.005908%  |  4.581023%   |\n",
                  "+---------------+-------------+------------+--------------+\n",
                  "+---------------+-------------+------------+--------------+\n",
                  "| Profit Margin | Sharp Ratio | Volatility | Max Drawdown |\n",
                  "+---------------+-------------+------------+--------------+\n",
                  "|   -2.709532%  | -187.849450 | 0.005698%  |  2.876131%   |\n",
                  "+---------------+-------------+------------+--------------+\n",
                  "Valid Episode: [5/10]\n",
                  "+--------------+------------+--------------+\n",
                  "| Total Return | Volatility | Max Drawdown |\n",
                  "+--------------+------------+--------------+\n",
                  "|  -3.582360%  | 0.005009%  |  3.753655%   |\n",
                  "+--------------+------------+--------------+\n",
                  "Train Episode: [6/10]\n",
                  "+---------------+-------------+------------+--------------+\n",
                  "| Profit Margin | Sharp Ratio | Volatility | Max Drawdown |\n",
                  "+---------------+-------------+------------+--------------+\n",
                  "|   -2.591166%  | -172.978546 | 0.005913%  |  2.626120%   |\n",
                  "+---------------+-------------+------------+--------------+\n",
                  "+---------------+-------------+------------+--------------+\n",
                  "| Profit Margin | Sharp Ratio | Volatility | Max Drawdown |\n",
                  "+---------------+-------------+------------+--------------+\n",
                  "|   -7.150399%  | -531.340272 | 0.005444%  |  7.183021%   |\n",
                  "+---------------+-------------+------------+--------------+\n",
                  "+---------------+-------------+------------+--------------+\n",
                  "| Profit Margin | Sharp Ratio | Volatility | Max Drawdown |\n",
                  "+---------------+-------------+------------+--------------+\n",
                  "|   -2.927376%  | -129.170595 | 0.008953%  |  4.023119%   |\n",
                  "+---------------+-------------+------------+--------------+\n",
                  "Valid Episode: [6/10]\n",
                  "+--------------+------------+--------------+\n",
                  "| Total Return | Volatility | Max Drawdown |\n",
                  "+--------------+------------+--------------+\n",
                  "|  -3.579781%  | 0.005222%  |  3.754136%   |\n",
                  "+--------------+------------+--------------+\n",
                  "Train Episode: [7/10]\n",
                  "+---------------+-------------+------------+--------------+\n",
                  "| Profit Margin | Sharp Ratio | Volatility | Max Drawdown |\n",
                  "+---------------+-------------+------------+--------------+\n",
                  "|   -4.947693%  | -552.255063 | 0.003583%  |  4.947693%   |\n",
                  "+---------------+-------------+------------+--------------+\n",
                  "+---------------+-------------+------------+--------------+\n",
                  "| Profit Margin | Sharp Ratio | Volatility | Max Drawdown |\n",
                  "+---------------+-------------+------------+--------------+\n",
                  "|   -7.518199%  | -659.789915 | 0.004619%  |  7.524365%   |\n",
                  "+---------------+-------------+------------+--------------+\n",
                  "+---------------+-------------+------------+--------------+\n",
                  "| Profit Margin | Sharp Ratio | Volatility | Max Drawdown |\n",
                  "+---------------+-------------+------------+--------------+\n",
                  "|   -6.773942%  | -735.281479 | 0.003720%  |  6.777264%   |\n",
                  "+---------------+-------------+------------+--------------+\n",
                  "Valid Episode: [7/10]\n",
                  "+--------------+------------+--------------+\n",
                  "| Total Return | Volatility | Max Drawdown |\n",
                  "+--------------+------------+--------------+\n",
                  "|  -4.735105%  | 0.005104%  |  4.910317%   |\n",
                  "+--------------+------------+--------------+\n",
                  "Train Episode: [8/10]\n",
                  "+---------------+-------------+------------+--------------+\n",
                  "| Profit Margin | Sharp Ratio | Volatility | Max Drawdown |\n",
                  "+---------------+-------------+------------+--------------+\n",
                  "|   -3.684426%  | -226.730691 | 0.006452%  |  3.740554%   |\n",
                  "+---------------+-------------+------------+--------------+\n",
                  "+---------------+-------------+------------+--------------+\n",
                  "| Profit Margin | Sharp Ratio | Volatility | Max Drawdown |\n",
                  "+---------------+-------------+------------+--------------+\n",
                  "|   -3.177255%  | -192.320172 | 0.006541%  |  3.241641%   |\n",
                  "+---------------+-------------+------------+--------------+\n",
                  "+---------------+-------------+------------+--------------+\n",
                  "| Profit Margin | Sharp Ratio | Volatility | Max Drawdown |\n",
                  "+---------------+-------------+------------+--------------+\n",
                  "|   -3.446168%  | -217.534943 | 0.006282%  |  3.461580%   |\n",
                  "+---------------+-------------+------------+--------------+\n",
                  "Valid Episode: [8/10]\n",
                  "+--------------+------------+--------------+\n",
                  "| Total Return | Volatility | Max Drawdown |\n",
                  "+--------------+------------+--------------+\n",
                  "|  -5.751130%  | 0.005604%  |  5.920226%   |\n",
                  "+--------------+------------+--------------+\n",
                  "Train Episode: [9/10]\n",
                  "+---------------+-------------+------------+--------------+\n",
                  "| Profit Margin | Sharp Ratio | Volatility | Max Drawdown |\n",
                  "+---------------+-------------+------------+--------------+\n",
                  "|   -3.937763%  | -173.505094 | 0.009017%  |  4.554533%   |\n",
                  "+---------------+-------------+------------+--------------+\n",
                  "+---------------+-------------+------------+--------------+\n",
                  "| Profit Margin | Sharp Ratio | Volatility | Max Drawdown |\n",
                  "+---------------+-------------+------------+--------------+\n",
                  "|   -3.535627%  | -226.197034 | 0.006202%  |  3.581861%   |\n",
                  "+---------------+-------------+------------+--------------+\n",
                  "+---------------+-------------+------------+--------------+\n",
                  "| Profit Margin | Sharp Ratio | Volatility | Max Drawdown |\n",
                  "+---------------+-------------+------------+--------------+\n",
                  "|   -7.156700%  | -483.521282 | 0.005987%  |  7.177333%   |\n",
                  "+---------------+-------------+------------+--------------+\n",
                  "Valid Episode: [9/10]\n",
                  "+--------------+------------+--------------+\n",
                  "| Total Return | Volatility | Max Drawdown |\n",
                  "+--------------+------------+--------------+\n",
                  "|  -3.991204%  | 0.005238%  |  4.144199%   |\n",
                  "+--------------+------------+--------------+\n",
                  "Train Episode: [10/10]\n",
                  "+---------------+-------------+------------+--------------+\n",
                  "| Profit Margin | Sharp Ratio | Volatility | Max Drawdown |\n",
                  "+---------------+-------------+------------+--------------+\n",
                  "|   -7.438490%  | -454.212161 | 0.006634%  |  7.552669%   |\n",
                  "+---------------+-------------+------------+--------------+\n",
                  "+---------------+-------------+------------+--------------+\n",
                  "| Profit Margin | Sharp Ratio | Volatility | Max Drawdown |\n",
                  "+---------------+-------------+------------+--------------+\n",
                  "|   -4.530511%  | -186.735150 | 0.009669%  |  4.530511%   |\n",
                  "+---------------+-------------+------------+--------------+\n",
                  "+---------------+-------------+------------+--------------+\n",
                  "| Profit Margin | Sharp Ratio | Volatility | Max Drawdown |\n",
                  "+---------------+-------------+------------+--------------+\n",
                  "|   -7.979081%  | -786.848392 | 0.004121%  |  7.979081%   |\n",
                  "+---------------+-------------+------------+--------------+\n",
                  "Valid Episode: [10/10]\n",
                  "+--------------+------------+--------------+\n",
                  "| Total Return | Volatility | Max Drawdown |\n",
                  "+--------------+------------+--------------+\n",
                  "|  -6.387074%  | 0.005365%  |  6.527131%   |\n",
                  "+--------------+------------+--------------+\n"
               ]
            }
         ],
         "source": [
            "trainer.train_and_valid()"
         ]
      },
      {
         "cell_type": "markdown",
         "metadata": {},
         "source": [
            "## Step 10: RL Agent Testing"
         ]
      },
      {
         "cell_type": "code",
         "execution_count": 12,
         "metadata": {
            "pycharm": {
               "name": "#%%\n"
            }
         },
         "outputs": [
            {
               "name": "stdout",
               "output_type": "stream",
               "text": [
                  "Resume checkpoint /data/sunshuo/qml/TradeMaster/work_dir/high_frequency_trading_BTC_high_frequency_trading_dqn_ddqn_adam_mse/checkpoints/best.pth\n",
                  "Test Best Episode\n",
                  "+--------------+------------+--------------+\n",
                  "| Total Return | Volatility | Max Drawdown |\n",
                  "+--------------+------------+--------------+\n",
                  "|  0.479027%   | 0.009575%  |  0.570660%   |\n",
                  "+--------------+------------+--------------+\n"
               ]
            }
         ],
         "source": [
            "trainer.test();"
         ]
      },
      {
         "cell_type": "code",
         "execution_count": 13,
         "metadata": {},
         "outputs": [
            {
               "data": {
                  "image/png": "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",
                  "text/plain": [
                     "<Figure size 432x288 with 1 Axes>"
                  ]
               },
               "metadata": {
                  "needs_background": "light"
               },
               "output_type": "display_data"
            }
         ],
         "source": [
            "plot(trainer.test_environment.save_asset_memoey(),alg=\"DDQN\")"
         ]
      }
   ],
   "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": "3.7.13"
      },
      "vscode": {
         "interpreter": {
            "hash": "a1f97403911abd3f02553c8f2b0c54537fddc7efadd9f5d3e31784db6e40c347"
         }
      }
   },
   "nbformat": 4,
   "nbformat_minor": 2
}