Repo updated

1c_Breakout-v0-Test.ipynb

@@ -0,0 +1,235 @@
+{
+ "cells": [
+  {
+   "cell_type": "code",
+   "execution_count": 1,
+   "id": "7d1fa20c",
+   "metadata": {},
+   "outputs": [],
+   "source": [
+    "import os, sys\n",
+    "import gc\n",
+    "import matplotlib.pyplot as plt\n",
+    "import numpy as np"
+   ]
+  },
+  {
+   "cell_type": "code",
+   "execution_count": null,
+   "id": "32bb66c4",
+   "metadata": {},
+   "outputs": [],
+   "source": [
+    "sys.path.append('../')\n",
+    "import torch"
+   ]
+  },
+  {
+   "cell_type": "code",
+   "execution_count": 3,
+   "id": "c8c767df",
+   "metadata": {},
+   "outputs": [],
+   "source": [
+    "from get_algos import get_all_algos , run_experiment ,call_paths, plot_mean"
+   ]
+  },
+  {
+   "cell_type": "code",
+   "execution_count": 4,
+   "id": "c9fc3531",
+   "metadata": {},
+   "outputs": [],
+   "source": [
+    "import gym\n",
+    "# from gym.wrappers import Monitor\n",
+    "from stable_baselines3 import PPO, A2C, DDPG, DQN, SAC, TD3\n",
+    "from stable_baselines3.common.vec_env import VecFrameStack,VecVideoRecorder\n",
+    "from stable_baselines3.common.evaluation import evaluate_policy\n",
+    "from stable_baselines3.common.env_util import make_atari_env\n",
+    "from stable_baselines3.common.monitor import Monitor"
+   ]
+  },
+  {
+   "cell_type": "code",
+   "execution_count": 5,
+   "id": "b1e1b76a",
+   "metadata": {},
+   "outputs": [],
+   "source": [
+    "env_name='Breakout-v0'"
+   ]
+  },
+  {
+   "cell_type": "markdown",
+   "id": "54c5cdd5",
+   "metadata": {},
+   "source": [
+    "### 2. Testing\n",
+    "<ol>\n",
+    "    <li>Create the environment</li>\n",
+    "    <li>For each algorithm:</li>\n",
+    "    <ol><li>Load the model</li>\n",
+    "        <li>Evaluate the model for 5 sample iterations </li>\n",
+    "        <li>Output the score  for each algo and each iteration</li>\n",
+    "        <li>Save renders of the output </li>\n",
+    "    </ol>\n",
+    "</ol> "
+   ]
+  },
+  {
+   "cell_type": "code",
+   "execution_count": 6,
+   "id": "65fae03b",
+   "metadata": {},
+   "outputs": [
+    {
+     "name": "stdout",
+     "output_type": "stream",
+     "text": [
+      "['PPO', 'A2C', 'DQN']\n"
+     ]
+    }
+   ],
+   "source": [
+    "algo_list=get_all_algos(gym.make(env_name))\n",
+    "print(algo_list)"
+   ]
+  },
+  {
+   "cell_type": "markdown",
+   "id": "7e8fab60",
+   "metadata": {},
+   "source": [
+    "### 3. Testing"
+   ]
+  },
+  {
+   "cell_type": "code",
+   "execution_count": 13,
+   "id": "972110da",
+   "metadata": {},
+   "outputs": [],
+   "source": [
+    "def testing_model(algo_list, env_name,n_steps,vid_length):\n",
+    "    total_rewards={}\n",
+    "    for algo_name in reversed(algo_list):\n",
+    "        algo_rewards=[]\n",
+    "        device='cpu' if algo_name=='DQN' else 'cuda'\n",
+    "        log_path, render_path, model_path=call_paths(algo_name,env_name,n_steps)\n",
+    "        env=VecVideoRecorder(VecFrameStack(make_atari_env(env_name,monitor_dir=render_path),n_stack=6),render_path,record_video_trigger=lambda step: step == 0,video_length=vid_length)\n",
+    "        model_name=env_name+\"_\"+algo_name+\"_model\"\n",
+    "        model=eval(algo_name).load(os.path.join(model_path, model_name),env,device=device)\n",
+    "        for i in range(5):\n",
+    "            state=env.reset()\n",
+    "            epi_rewards=0\n",
+    "            while True:\n",
+    "                action,_=model.predict(state)\n",
+    "                state,reward,done,_=env.step(action)\n",
+    "                epi_rewards+=reward[0]\n",
+    "                env.render()\n",
+    "                if done:\n",
+    "                    algo_rewards.append(epi_rewards)\n",
+    "                    break\n",
+    "        total_rewards[algo_name]=algo_rewards\n",
+    "        del model\n",
+    "        env.close()\n",
+    "    return total_rewards"
+   ]
+  },
+  {
+   "cell_type": "code",
+   "execution_count": 14,
+   "id": "2f2d2d60",
+   "metadata": {},
+   "outputs": [
+    {
+     "name": "stdout",
+     "output_type": "stream",
+     "text": [
+      "Wrapping the env in a VecTransposeImage.\n"
+     ]
+    },
+    {
+     "name": "stderr",
+     "output_type": "stream",
+     "text": [
+      "E:\\Anaconda\\envs\\ai_gym\\lib\\site-packages\\stable_baselines3\\common\\buffers.py:229: UserWarning: This system does not have apparently enough memory to store the complete replay buffer 84.69GB > 11.00GB\n",
+      "  \"This system does not have apparently enough memory to store the complete \"\n"
+     ]
+    },
+    {
+     "name": "stdout",
+     "output_type": "stream",
+     "text": [
+      "Wrapping the env in a VecTransposeImage.\n",
+      "Wrapping the env in a VecTransposeImage.\n"
+     ]
+    }
+   ],
+   "source": [
+    "test=total_rewards=testing_model(algo_list, env_name,n_steps=200000,vid_length=1000)"
+   ]
+  },
+  {
+   "cell_type": "code",
+   "execution_count": 15,
+   "id": "63f6b45f",
+   "metadata": {},
+   "outputs": [
+    {
+     "data": {
+      "text/plain": [
+       "{'DQN': [0.0, 0.0, 0.0, 5.0, 0.0],\n",
+       " 'A2C': [4.0, 0.0, 0.0, 0.0, 0.0],\n",
+       " 'PPO': [3.0, 0.0, 0.0, 3.0, 0.0]}"
+      ]
+     },
+     "execution_count": 15,
+     "metadata": {},
+     "output_type": "execute_result"
+    }
+   ],
+   "source": [
+    "test"
+   ]
+  },
+  {
+   "cell_type": "markdown",
+   "id": "870561de",
+   "metadata": {},
+   "source": [
+    "We see that PPO seems to consistently provide rewards with the highest average reward. Hence, we will use PPO as the algorithm for subsequent training of <b> 5,000,000 </b>steps"
+   ]
+  },
+  {
+   "cell_type": "code",
+   "execution_count": null,
+   "id": "13d01daf",
+   "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": "3.7.0"
+  }
+ },
+ "nbformat": 4,
+ "nbformat_minor": 5
+}