main.ipynb

{
 "cells": [
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "# Set up"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 1,
   "metadata": {},
   "outputs": [],
   "source": [
    "import gym_super_mario_bros # Importing the Super Mario Bros environment\n",
    "from nes_py.wrappers import JoypadSpace # Importing joypad wrapper\n",
    "from gym_super_mario_bros.actions import SIMPLE_MOVEMENT # Importing simplified controls\n",
    "import os\n",
    "from PIL import Image\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 2,
   "metadata": {},
   "outputs": [],
   "source": [
    "curr = os.getcwd()\n",
    "os.chdir('/teamspace/studios/this_studio/RL-Mario')\n",
    "from save_frame import save_images, save_video"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 3,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "[['NOOP'],\n",
       " ['right'],\n",
       " ['right', 'A'],\n",
       " ['right', 'B'],\n",
       " ['right', 'A', 'B'],\n",
       " ['A'],\n",
       " ['left']]"
      ]
     },
     "execution_count": 3,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "SIMPLE_MOVEMENT"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 4,
   "metadata": {},
   "outputs": [
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "/home/zeus/miniconda3/envs/cloudspace/lib/python3.10/site-packages/gym/envs/registration.py:555: UserWarning: \u001b[33mWARN: The environment SuperMarioBros-v0 is out of date. You should consider upgrading to version `v3`.\u001b[0m\n",
      "  logger.warn(\n",
      "/home/zeus/miniconda3/envs/cloudspace/lib/python3.10/site-packages/gym/envs/registration.py:627: UserWarning: \u001b[33mWARN: The environment creator metadata doesn't include `render_modes`, contains: ['render.modes', 'video.frames_per_second']\u001b[0m\n",
      "  logger.warn(\n"
     ]
    },
    {
     "data": {
      "text/plain": [
       "256"
      ]
     },
     "execution_count": 4,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "\"\"\"\n",
    "    1. Creating the environment\n",
    "    2. Ensuring that the environment is compatible with the OpenAI Gym API\n",
    "    3. Setting the rendering mode to numpy array form\n",
    "\"\"\"\n",
    "env = gym_super_mario_bros.make('SuperMarioBros-v0', apply_api_compatibility=True, render_mode='rgb_array') \n",
    "env.action_space.n # Checking the number of actions"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 5,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "7"
      ]
     },
     "execution_count": 5,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "env = JoypadSpace(env, SIMPLE_MOVEMENT) # Applying the wrapper\n",
    "env.action_space.n # Simplified action space"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 6,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "(240, 256, 3)"
      ]
     },
     "execution_count": 6,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "env.observation_space.shape # Checking the shape of the observation space (height, width, channels)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 7,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "0"
      ]
     },
     "execution_count": 7,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "env.action_space.sample() # Random action"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 8,
   "metadata": {},
   "outputs": [
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "/home/zeus/miniconda3/envs/cloudspace/lib/python3.10/site-packages/gym/utils/passive_env_checker.py:233: DeprecationWarning: `np.bool8` is a deprecated alias for `np.bool_`.  (Deprecated NumPy 1.24)\n",
      "  if not isinstance(terminated, (bool, np.bool8)):\n"
     ]
    },
    {
     "data": {
      "text/plain": [
       "(array([[[104, 136, 252],\n",
       "         [104, 136, 252],\n",
       "         [104, 136, 252],\n",
       "         ...,\n",
       "         [104, 136, 252],\n",
       "         [104, 136, 252],\n",
       "         [104, 136, 252]],\n",
       " \n",
       "        [[104, 136, 252],\n",
       "         [104, 136, 252],\n",
       "         [104, 136, 252],\n",
       "         ...,\n",
       "         [104, 136, 252],\n",
       "         [104, 136, 252],\n",
       "         [104, 136, 252]],\n",
       " \n",
       "        [[104, 136, 252],\n",
       "         [104, 136, 252],\n",
       "         [104, 136, 252],\n",
       "         ...,\n",
       "         [104, 136, 252],\n",
       "         [104, 136, 252],\n",
       "         [104, 136, 252]],\n",
       " \n",
       "        ...,\n",
       " \n",
       "        [[240, 208, 176],\n",
       "         [228,  92,  16],\n",
       "         [228,  92,  16],\n",
       "         ...,\n",
       "         [228,  92,  16],\n",
       "         [228,  92,  16],\n",
       "         [  0,   0,   0]],\n",
       " \n",
       "        [[240, 208, 176],\n",
       "         [228,  92,  16],\n",
       "         [228,  92,  16],\n",
       "         ...,\n",
       "         [228,  92,  16],\n",
       "         [  0,   0,   0],\n",
       "         [  0,   0,   0]],\n",
       " \n",
       "        [[228,  92,  16],\n",
       "         [  0,   0,   0],\n",
       "         [  0,   0,   0],\n",
       "         ...,\n",
       "         [  0,   0,   0],\n",
       "         [  0,   0,   0],\n",
       "         [228,  92,  16]]], dtype=uint8),\n",
       " 0.0,\n",
       " False,\n",
       " False,\n",
       " {'coins': 0,\n",
       "  'flag_get': False,\n",
       "  'life': 2,\n",
       "  'score': 0,\n",
       "  'stage': 1,\n",
       "  'status': 'small',\n",
       "  'time': 400,\n",
       "  'world': 1,\n",
       "  'x_pos': 40,\n",
       "  'y_pos': 79})"
      ]
     },
     "execution_count": 8,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "env.reset()\n",
    "env.step(env.action_space.sample()) # Taking a random action\n",
    "\n",
    "# Ouput format : (observation, reward, done, info, extra_info)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 9,
   "metadata": {},
   "outputs": [],
   "source": [
    "# done = True # Checking if the game is done\n",
    "# for step in range(10000): # Running through the frames\n",
    "#     if done:\n",
    "#         env.reset() # Resetting the environment\n",
    "#     state, reward, done, info, _ = env.step(env.action_space.sample())  # Taking a random action\n",
    "#     frame = env.render() # Rendering the environment\n",
    "#     save_images(frame, step) # Saving the images\n",
    "    \n",
    "# env.close() # Closing the environment"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 10,
   "metadata": {},
   "outputs": [],
   "source": [
    "# save_video('videos', 'random_action.mp4') # Creating the video"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 11,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "(240, 256, 3)"
      ]
     },
     "execution_count": 11,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "state = env.reset()[0] # Resetting the environment\n",
    "state.shape # Checking the shape of"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 12,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "5"
      ]
     },
     "execution_count": 12,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "len(env.step(1)) # observation, reward, done, info , extra_info"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "# Pre-processing"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 13,
   "metadata": {},
   "outputs": [],
   "source": [
    "\"\"\"\n",
    "    GrayScaleObservation : Converting the frames to grayscale for faster processing\n",
    "    VecFrameStack : Stacking the frames for multiple environments\n",
    "    DummyVecEnv : Creating a dummy vectorized environment\n",
    "    plt : Matplotlib for plotting the impact of frame stacking\n",
    "\"\"\"\n",
    "from gym.wrappers import GrayScaleObservation \n",
    "from stable_baselines3.common.vec_env import VecFrameStack, DummyVecEnv\n",
    "from matplotlib import pyplot as plt"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 14,
   "metadata": {},
   "outputs": [
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "/home/zeus/miniconda3/envs/cloudspace/lib/python3.10/site-packages/gym/envs/registration.py:555: UserWarning: \u001b[33mWARN: The environment SuperMarioBros-v0 is out of date. You should consider upgrading to version `v3`.\u001b[0m\n",
      "  logger.warn(\n",
      "/home/zeus/miniconda3/envs/cloudspace/lib/python3.10/site-packages/gym/envs/registration.py:627: UserWarning: \u001b[33mWARN: The environment creator metadata doesn't include `render_modes`, contains: ['render.modes', 'video.frames_per_second']\u001b[0m\n",
      "  logger.warn(\n",
      "/home/zeus/miniconda3/envs/cloudspace/lib/python3.10/site-packages/stable_baselines3/common/vec_env/patch_gym.py:49: UserWarning: You provided an OpenAI Gym environment. We strongly recommend transitioning to Gymnasium environments. Stable-Baselines3 is automatically wrapping your environments in a compatibility layer, which could potentially cause issues.\n",
      "  warnings.warn(\n"
     ]
    },
    {
     "data": {
      "text/plain": [
       "(240, 256, 4)"
      ]
     },
     "execution_count": 14,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "# Creating a base environment\n",
    "env = gym_super_mario_bros.make('SuperMarioBros-v0', apply_api_compatibility=True, render_mode='rgb_array')\n",
    "# Simplifying the controls\n",
    "env = JoypadSpace(env, SIMPLE_MOVEMENT)\n",
    "# Grayscale\n",
    "env = GrayScaleObservation(env, keep_dim=True) # Keep the dimension : True (height, width, 1) False (height, width)\n",
    "# Wrap inside DummyVecEnv\n",
    "env = DummyVecEnv([lambda: env]) # Adds 1 dimension to the observation space for the batch size\n",
    "# # Stack 4 frames\n",
    "env = VecFrameStack(env, n_stack=4, channels_order='last') \n",
    "# before (batch, height, width, channels) -> after (batch, height, width, channels * n_stack)\n",
    "env.observation_space.shape # Checking the shape of the observation space"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 15,
   "metadata": {},
   "outputs": [],
   "source": [
    "JoypadSpace.reset = lambda self, **kwargs: self.env.reset(**kwargs) # JoyPadSpace reset function \n",
    "\n",
    "state = env.reset()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 16,
   "metadata": {},
   "outputs": [
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "/home/zeus/miniconda3/envs/cloudspace/lib/python3.10/site-packages/gym/utils/passive_env_checker.py:233: DeprecationWarning: `np.bool8` is a deprecated alias for `np.bool_`.  (Deprecated NumPy 1.24)\n",
      "  if not isinstance(terminated, (bool, np.bool8)):\n"
     ]
    },
    {
     "data": {
      "image/png": "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",
      "text/plain": [
       "<Figure size 2000x1600 with 4 Axes>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "state, reward, done, info = env.step([env.action_space.sample()])  # Taking a random action\n",
    "\n",
    "plt.figure(figsize=(20, 16))\n",
    "for idx in range(state.shape[3]):\n",
    "    plt.subplot(1, 4, idx + 1)\n",
    "    plt.imshow(state[0][:, :, idx])\n",
    "    plt.axis('off')"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "# Train"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 17,
   "metadata": {},
   "outputs": [],
   "source": [
    "import os \n",
    "from stable_baselines3 import PPO # algo\n",
    "from stable_baselines3.common.callbacks import BaseCallback # saving the model"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 18,
   "metadata": {},
   "outputs": [],
   "source": [
    "class TrainAndLoggingCallback(BaseCallback):\n",
    "    def __init__(self, check_freq, save_path, verbose=1): # verbose : 1 (basic info)\n",
    "        super().__init__()\n",
    "        self.check_freq = check_freq # Setting the check frequency for saving the model\n",
    "        self.save_path = save_path \n",
    "\n",
    "    def _init_callback(self):\n",
    "        \"\"\" Create the folder for saving the model \"\"\"\n",
    "        if self.save_path is not None:\n",
    "            os.makedirs(self.save_path, exist_ok=True)\n",
    "\n",
    "\n",
    "    def _on_step(self) -> bool:\n",
    "        if self.n_calls % self.check_freq == 0:\n",
    "            # Saving the model every 1000 steps\n",
    "            model_path = os.path.join(self.save_path, f'best_model_{self.n_calls}')\n",
    "            self.model.save(model_path)\n",
    "        return True"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 19,
   "metadata": {},
   "outputs": [],
   "source": [
    "CHECKPOINT_DIR = './train/'\n",
    "LOG_DIR = './logs/'"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 20,
   "metadata": {},
   "outputs": [],
   "source": [
    "callback = TrainAndLoggingCallback(check_freq=1000000, save_path=CHECKPOINT_DIR)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 21,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Using cuda device\n",
      "Wrapping the env in a VecTransposeImage.\n"
     ]
    }
   ],
   "source": [
    "model = PPO('CnnPolicy', env, verbose=1, tensorboard_log=LOG_DIR, learning_rate=1e-6,\n",
    "            n_steps=2048, batch_size=2048 ) #n_steps -> update frequency per frame n_steps * n_envs"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 22,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Logging to ./logs/PPO_1\n",
      "-----------------------------\n",
      "| time/              |      |\n",
      "|    fps             | 204  |\n",
      "|    iterations      | 1    |\n",
      "|    time_elapsed    | 9    |\n",
      "|    total_timesteps | 2048 |\n",
      "-----------------------------\n",
      "------------------------------------------\n",
      "| time/                   |              |\n",
      "|    fps                  | 164          |\n",
      "|    iterations           | 2            |\n",
      "|    time_elapsed         | 24           |\n",
      "|    total_timesteps      | 4096         |\n",
      "| train/                  |              |\n",
      "|    approx_kl            | 4.331232e-07 |\n",
      "|    clip_fraction        | 0            |\n",
      "|    clip_range           | 0.2          |\n",
      "|    entropy_loss         | -1.95        |\n",
      "|    explained_variance   | 0.00179      |\n",
      "|    learning_rate        | 1e-06        |\n",
      "|    loss                 | 89.7         |\n",
      "|    n_updates            | 10           |\n",
      "|    policy_gradient_loss | -1.6e-05     |\n",
      "|    value_loss           | 183          |\n",
      "------------------------------------------\n",
      "-------------------------------------------\n",
      "| time/                   |               |\n",
      "|    fps                  | 157           |\n",
      "|    iterations           | 3             |\n",
      "|    time_elapsed         | 39            |\n",
      "|    total_timesteps      | 6144          |\n",
      "| train/                  |               |\n",
      "|    approx_kl            | 3.7273276e-07 |\n",
      "|    clip_fraction        | 0             |\n",
      "|    clip_range           | 0.2           |\n",
      "|    entropy_loss         | -1.95         |\n",
      "|    explained_variance   | 0.00162       |\n",
      "|    learning_rate        | 1e-06         |\n",
      "|    loss                 | 0.584         |\n",
      "|    n_updates            | 20            |\n",
      "|    policy_gradient_loss | -2.6e-05      |\n",
      "|    value_loss           | 1.41          |\n",
      "-------------------------------------------\n"
     ]
    }
   ],
   "source": [
    "model.learn(total_timesteps=10000000, callback=callback)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "# Test"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": [
    "model = PPO.load('RL-Mario/train/best_model_9000000')"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "['left']"
      ]
     },
     "execution_count": 22,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "state = env.reset()\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "(array([4]), None)"
      ]
     },
     "execution_count": 23,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "model.predict(state)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "['A']"
      ]
     },
     "execution_count": 24,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "SIMPLE_MOVEMENT[model.predict(state)[0][0]]"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": [
    "step = 0\n",
    "while True:\n",
    "    action, _ = model.predict(state)\n",
    "    state, reward, done, info = env.step(action)\n",
    "    frame = env.render()\n",
    "    save_images(frame, step)\n",
    "    step += 1 \n",
    "    if done:\n",
    "        break"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Video saved to ppo_model.mp4\n"
     ]
    }
   ],
   "source": [
    "save_video('videos', 'ppo_model.mp4')"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": []
  }
 ],
 "metadata": {
  "kernelspec": {
   "display_name": "cloudspace",
   "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": "3.10.10"
  }
 },
 "nbformat": 4,
 "nbformat_minor": 2
}