Ex_04_DQN_approximation_lab.ipynb

{"nbformat":4,"nbformat_minor":0,"metadata":{"colab":{"provenance":[{"file_id":"1OWqBwq0BN0rQCfPDPB6bruBPSErA2BPg","timestamp":1679843587254},{"file_id":"1DiygZpnDaAQE4Tlb0zMxOg_rzenpH36T","timestamp":1648464626623},{"file_id":"1ZqlNaTexUCQ-_Zw89wgpYnovEq2PtvXx","timestamp":1648415770514},{"file_id":"1Jl1NllT9gqQWicZRgOCXE5-20SNX3fdq","timestamp":1643805237849},{"file_id":"1tdHcENw4oLA2G7gWqfgzODRAUT_pNhVL","timestamp":1641227537710}]},"kernelspec":{"name":"python3","display_name":"Python 3"},"language_info":{"name":"python"},"accelerator":"GPU","gpuClass":"standard"},"cells":[{"cell_type":"markdown","metadata":{"id":"iEayTRvYuww1"},"source":["#DQN, Function Approximation, Pefrormance tricks\n","\n","In this lab we study the basics of Q learning with function approximation by neural networks."]},{"cell_type":"code","source":["# Installing dependencies for visualization\n","!apt-get -qq -y install libcusparse8.0 libnvrtc8.0 libnvtoolsext1 > /dev/null\n","!ln -snf /usr/lib/x86_64-linux-gnu/libnvrtc-builtins.so.8.0 /usr/lib/x86_64-linux-gnu/libnvrtc-builtins.so\n","!apt-get -qq -y install xvfb freeglut3-dev ffmpeg> /dev/null\n","!pip -q install gym\n","!pip -q install pyglet\n","!pip -q install pyopengl\n","!pip -q install pyvirtualdisplay"],"metadata":{"id":"I32BwFgeVhHg","colab":{"base_uri":"https://localhost:8080/"},"executionInfo":{"status":"ok","timestamp":1680282212430,"user_tz":-120,"elapsed":37921,"user":{"displayName":"Michał Janik","userId":"10483929333441523386"}},"outputId":"d7cfd829-3b5f-425a-ba7f-39c92ee56a01"},"execution_count":1,"outputs":[{"output_type":"stream","name":"stdout","text":["E: Unable to locate package libcusparse8.0\n","E: Couldn't find any package by glob 'libcusparse8.0'\n","E: Couldn't find any package by regex 'libcusparse8.0'\n","E: Unable to locate package libnvrtc8.0\n","E: Couldn't find any package by glob 'libnvrtc8.0'\n","E: Couldn't find any package by regex 'libnvrtc8.0'\n","\u001b[2K     \u001b[90m━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━\u001b[0m \u001b[32m831.3/831.3 KB\u001b[0m \u001b[31m13.2 MB/s\u001b[0m eta \u001b[36m0:00:00\u001b[0m\n","\u001b[?25h"]}]},{"cell_type":"code","source":["import glob\n","import random\n","import time\n","\n","\n","import gym\n","import matplotlib.pyplot as plt\n","import numpy as np\n","from tensorflow.keras.optimizers import Adam\n","from tensorflow.python.keras.regularizers import l2\n","from tensorflow.keras.layers import Input, Dense\n","from tensorflow.keras.models import Model, clone_model\n","\n","import numpy as np\n","import matplotlib.pyplot as plt\n","import collections\n","from collections import deque\n","\n","from base64 import b64encode\n","from IPython.display import HTML\n","from pyvirtualdisplay import Display\n","\n","# Start virtual display\n","display = Display(visible=0, size=(1024, 768))\n","display.start()\n","\n","def show_video(file_name):\n","    mp4 = open(file_name,'rb').read()\n","    data_url = \"data:video/mp4;base64,\" + b64encode(mp4).decode()\n","    return HTML(\"\"\"\n","    <video width=480 controls>\n","        <source src=\"%s\" type=\"video/mp4\">\n","    </video>\n","    \"\"\" % data_url)"],"metadata":{"id":"ckBLjgWAeQy5","executionInfo":{"status":"ok","timestamp":1680282215875,"user_tz":-120,"elapsed":3450,"user":{"displayName":"Michał Janik","userId":"10483929333441523386"}}},"execution_count":2,"outputs":[]},{"cell_type":"markdown","source":["We will start by defining a useful data structure:"],"metadata":{"id":"hvcjOL9MborZ"}},{"cell_type":"code","source":["Transition = collections.namedtuple('transition', ['state', 'action', 'reward', 'done', 'next_state'])"],"metadata":{"id":"t5U1thRNbyNO","executionInfo":{"status":"ok","timestamp":1680282216162,"user_tz":-120,"elapsed":292,"user":{"displayName":"Michał Janik","userId":"10483929333441523386"}},"colab":{"base_uri":"https://localhost:8080/"},"outputId":"eb9ac38d-a347-4905-9cd9-957261aa9174"},"execution_count":3,"outputs":[{"output_type":"stream","name":"stderr","text":["/usr/local/lib/python3.9/dist-packages/ipykernel/ipkernel.py:283: DeprecationWarning: `should_run_async` will not call `transform_cell` automatically in the future. Please pass the result to `transformed_cell` argument and any exception that happen during thetransform in `preprocessing_exc_tuple` in IPython 7.17 and above.\n","  and should_run_async(code)\n"]}]},{"cell_type":"markdown","source":["## CartPole\n","Debugging DQN is typically a complicated process, thus we have to start with a simple environment, that can be quickly iterated. Let's first construct working DQN for CartPole problem. We will use a small modification of the orginal CartPole env, we do reward reshape (to make problem easier for DQN). Precisely, we add a punishment for ending the episode:"],"metadata":{"id":"bPlVCdaAchgH"}},{"cell_type":"code","source":["class ModifiedCartPole:\n","    def __init__(self):\n","        self.env = gym.make('CartPole-v0')\n","\n","    def reset(self):\n","        return self.env.reset()\n","\n","    def step(self, action):\n","        obs, reward, done, _ = self.env.step(action)\n","        if done:\n","            reward = -10\n","        return obs, reward/10, done, {}"],"metadata":{"id":"YP_8AWXAeFX7","executionInfo":{"status":"ok","timestamp":1680282216162,"user_tz":-120,"elapsed":14,"user":{"displayName":"Michał Janik","userId":"10483929333441523386"}}},"execution_count":4,"outputs":[]},{"cell_type":"markdown","source":["##Q-network. \n","First we must create a network to approximate Q(s, a). We have two natural design choices:\n","- Q-network takes two inputs: state s and action a and predicts one value Q(s,a)\n","- Q-network takes one input: state s, and predicts a vector of Q(s, a) for all possible actions.\n","\n","We will follow the second design choice (one of the reasons is that such network can faster predict the best action).\n","\n","**Exercise: fill the code below to create Q-network**. Create a simple fully connected network with num_layers layers each with 64 neurons. The input is a vector of size 4, and the output is a vector of size 2 (we have 2 actions in cartpole)."],"metadata":{"id":"UrXppGmcc6aB"}},{"cell_type":"code","source":["def make_cartpole_network(input_size=4, num_action=2, num_layers=3, learning_rate=1e-4, weight_decay=0.):\n","  input_state = Input(batch_shape=(None, input_size))\n","  x = input_state\n","  for _ in range(num_layers):\n","    x = Dense(64, activation='relu', kernel_regularizer=l2(weight_decay))(x)\n","  output = Dense(num_action, kernel_regularizer=l2(weight_decay))(x)\n","  model = Model(inputs=input_state, outputs=output)\n","  model.compile(\n","      loss='mse',\n","      optimizer=Adam(learning_rate=learning_rate)\n","  )\n","  return model"],"metadata":{"id":"4hivxFi6BWRi","executionInfo":{"status":"ok","timestamp":1680282216163,"user_tz":-120,"elapsed":15,"user":{"displayName":"Michał Janik","userId":"10483929333441523386"}}},"execution_count":5,"outputs":[]},{"cell_type":"markdown","source":["## Building DQN\n","\n","We will start with some utils functions:\n","\n","**Exercise: read the following functions, to understand them** (will be used later)."],"metadata":{"id":"kIctRLzEgxNN"}},{"cell_type":"code","source":["def predict_q_values(q_network, state):\n","  # Makes a prediction for a single state and returns array of Q-values\n","  return q_network.predict(np.array([state]), verbose=0)[0]\n","\n","def choose_best_action(q_network, state):\n","  # Chooses best action according to Q-network\n","  action_values = predict_q_values(q_network, state)\n","  best_action = np.argmax(action_values)\n","  return best_action\n","\n","def evaluate_state_batch(target_network, state_batch):\n","  '''This function can evaluate the whole batch of states at once, it\n","  is very useful to speedup the training when we calculate targets\n","  Arguments:\n","    - state_batch: list of states to evaluate\n","  Returns:\n","    - best actions: list of best action for every state\n","    - best vals: list of best state-action values for very state\n","    - action_values: list of all action-values for each state\n","\n","  Here we named the argument target network instead of q_network, because this\n","  function will be used with target network.\n","  '''\n","  action_values = target_network.predict(np.array(state_batch), verbose=0)\n","  best_actions = np.argmax(action_values, axis=-1)\n","  best_vals = np.max(action_values, axis=-1)\n","  return best_actions, best_vals, action_values\n","\n","\n","def choose_action(q_network, state, epsilon):\n","  if random.random() < epsilon:\n","      return random.randint(0, 1)\n","  else:\n","      return choose_best_action(q_network, state)"],"metadata":{"id":"G-7XROP_nkBm","executionInfo":{"status":"ok","timestamp":1680282216163,"user_tz":-120,"elapsed":15,"user":{"displayName":"Michał Janik","userId":"10483929333441523386"}}},"execution_count":6,"outputs":[]},{"cell_type":"markdown","source":["While running the epizodes we will collect transitions and store them in a replay_buffer, which is just a list of transitions. Before we write a code for running epizodes we must first prepare a function that prepares training (since it is used while running the game) and a one for doing the training.\n","\n","**Exercise: the training protocole is the heart of DQN. Fill the gaps in the following function**"],"metadata":{"id":"DRiSgvoRueyJ"}},{"cell_type":"code","source":["def prepare_train_targets(target_network, replay_buffer, mini_batch_size, q_learning_rate=0.5, gamma=0.99):\n","  ''' Write a code to choose random samples from replay_buffer. \n","  Choose mini_batch_size of samples and collect them in replay_batch. \n","  Replay batch must be a list of transitions. '''\n","\n","  #Hint: you can use random.sample method\n","  replay_batch = random.sample(replay_buffer, mini_batch_size)\n","\n","  # We will collect all states and next_states from the batch of transitions to evaluate them at once\n","  # ( optimization!)\n","  next_state_batch = []\n","  state_batch = []\n","  train_x, train_y = [], []\n","\n","  for transition in replay_batch:\n","    next_state_batch.append(transition.next_state)\n","    state_batch.append(transition.state)\n","\n","  _, next_state_values, _ = evaluate_state_batch(target_network, next_state_batch)\n","  _, _, state_action_vals = evaluate_state_batch(target_network, state_batch)\n","\n","  for transition, state_vals, next_state_value in zip(replay_batch, state_action_vals, next_state_values):\n","    action = transition.action\n","    \n","    ''' Create train datapoints'''\n","    # Copy transition.state to x, use copy() method\n","    x = transition.state.copy()\n","    ''' In out setting y must be a list of two values (Q-values for 2 actions).\n","    Note, that we want to update only of the y values (corresponding to chosen action)'''\n","    # Copy state_vals vector to y\n","    y = state_vals.copy()\n","    \n","    '''Calculate the target value for y[action]. Implement the Q-iteration formula:\n","    Q(s,a) <- q_learning_rate*(reward + gamma * max_a' Q_target(s', a')) + (1-q_learning_rate Q(s,a))\n","    \n","    Remember to separately treat the case where the transition is last in the epizode. \n","    You can check if the transition is last, by looking at transition.done'''\n","    new_q_val = transition.reward + (1 - transition.done) * gamma * next_state_value\n","    y[action] = q_learning_rate * new_q_val + (1 - q_learning_rate) * y[action]\n","    train_x.append(x)\n","    train_y.append(y)\n","\n","  return np.array(train_x), np.array(train_y)"],"metadata":{"id":"ZuMdEiapC5Xp","executionInfo":{"status":"ok","timestamp":1680282216163,"user_tz":-120,"elapsed":14,"user":{"displayName":"Michał Janik","userId":"10483929333441523386"}}},"execution_count":7,"outputs":[]},{"cell_type":"markdown","source":["**Exercise: fill the gaps in the training function**"],"metadata":{"id":"C9XEP4-vGxK9"}},{"cell_type":"code","source":["def train(q_network, target_network, replay_buffer, train_epochs, mini_batch_size, q_learning_rate, gamma):\n","  train_x, train_y =  prepare_train_targets(target_network, replay_buffer, mini_batch_size, q_learning_rate, gamma)\n","  '''For models in keras you can use fit() method that takes x and y as inputs and return history. \n","      Hint: fit() method has verbose atribute that determines what this function prints on the screen \n","      You can set verbose=0 to have no prints.'''\n","  history = q_network.fit(x=train_x, y=train_y,\n","                          batch_size=mini_batch_size,\n","                          epochs=train_epochs,\n","                          verbose=0)\n","  return history.history['loss']"],"metadata":{"id":"5CnS5l9kv3N_","executionInfo":{"status":"ok","timestamp":1680282216164,"user_tz":-120,"elapsed":15,"user":{"displayName":"Michał Janik","userId":"10483929333441523386"}}},"execution_count":8,"outputs":[]},{"cell_type":"markdown","source":["Now, let us code the heart of DQN algorithm: the function that runs an epizode and trains Q-network.\n","\n","**Exercise: fill the code in run_one_episode function**"],"metadata":{"id":"kv7eYsOj37y7"}},{"cell_type":"code","source":["def run_one_episode(q_network, target_network, env, epsilon, steps_so_far, replay_buffer,\n","                    train_epochs, mini_batch_size, train_every_n_steps, update_target_every_n_steps,\n","                    q_learning_rate, gamma):\n","    done = False\n","    episode_steps = 0\n","    state = env.reset()\n","    ep_actions = []\n","    loss_history = []\n","    while not done:\n","        action = choose_action(q_network, state, epsilon)\n","        ep_actions.append(action)\n","        next_state, reward, done, _ = env.step(action)\n","        episode_steps += 1\n","        steps_so_far += 1\n","\n","        new_transition = Transition(\n","            state=state,\n","            action=action,\n","            reward=reward,\n","            done=done,\n","            next_state=next_state\n","        )\n","        replay_buffer.append(new_transition)\n","\n","        state = next_state\n","        \n","        if steps_so_far % train_every_n_steps == 0 and len(replay_buffer) > mini_batch_size:\n","            history = train(q_network, target_network, replay_buffer, train_epochs, mini_batch_size, q_learning_rate, gamma)\n","            loss_history.extend(history)\n","            if steps_so_far % update_target_every_n_steps == 0:\n","                # update weights of the target network i.e. make them equal to the weigths of q_network \n","                target_network.set_weights(q_network.get_weights())\n","\n","    return episode_steps, loss_history"],"metadata":{"id":"1hBpzbmmHhY3","executionInfo":{"status":"ok","timestamp":1680282216164,"user_tz":-120,"elapsed":15,"user":{"displayName":"Michał Janik","userId":"10483929333441523386"}}},"execution_count":9,"outputs":[]},{"cell_type":"markdown","source":["Finally, we can complete the full DQN algorithm.\n","\n","**Exercise: complete the full training loop code**"],"metadata":{"id":"MdYqQU9l1a8X"}},{"cell_type":"code","source":["def run_dqn(train_steps=10000, use_target_network=True, max_replay_buffer_size=None):\n","    env = ModifiedCartPole()\n","    \n","    # Here is a set of default parameters (tested), you can try to find better values\n","    epsilon = 0.6\n","    min_epsilon = 0.1\n","    epsilon_decay=0.995\n","    gamma=0.99\n","    q_learning_rate = 0.6\n","    train_every_n_steps=32\n","    mini_batch_size=128\n","    update_target_every_n_steps=128\n","    train_epochs=2\n","\n","    q_checkpoints = []\n","\n","    if max_replay_buffer_size is None:\n","      replay_buffer = []\n","    else:\n","      assert max_replay_buffer_size >= mini_batch_size\n","      replay_buffer = deque(maxlen=max_replay_buffer_size)\n","\n","    q_network = make_cartpole_network()\n","    if use_target_network:\n","      target_network = make_cartpole_network()\n","      target_network.set_weights(q_network.get_weights())\n","    else:\n","      target_network = q_network\n","    \n","    steps_so_far = 0\n","\n","    episode_lengths, loss_history = [], []\n","    episode_num = 0\n","\n","    while steps_so_far < train_steps:\n","        episode_length, loss = run_one_episode(\n","            q_network, target_network, env,epsilon, steps_so_far, replay_buffer,\n","            train_epochs, mini_batch_size, train_every_n_steps,\n","            update_target_every_n_steps, q_learning_rate, gamma\n","        )\n","        if epsilon > min_epsilon:\n","            epsilon *= epsilon_decay\n","        episode_num += 1\n","        episode_lengths.append(episode_length)\n","        if loss is not None:\n","          loss_history.extend(loss)\n","        steps_so_far += episode_length\n","        q_checkpoints.append(q_network.get_weights())\n","        print(f'Episode = {episode_num} | steps =  {steps_so_far} | episode_length = {episode_length} | epsilon = {epsilon}')\n","\n","    return episode_lengths, loss_history, q_checkpoints"],"metadata":{"id":"T9L7S4yMJhy0","executionInfo":{"status":"ok","timestamp":1680282216164,"user_tz":-120,"elapsed":14,"user":{"displayName":"Michał Janik","userId":"10483929333441523386"}}},"execution_count":10,"outputs":[]},{"cell_type":"markdown","source":["Let us now run the training (it may take several minutes to take the training of 5000-8000 steps). Do not expect the reward to grow monotonically. The training typically looks like a noisy process with some drift towards higher returns."],"metadata":{"id":"I3_LzNLP464R"}},{"cell_type":"code","source":["progress_standard, loss_history_standard, q_checkpoints_standard = run_dqn(8000)"],"metadata":{"id":"BOKTRkSS490q","colab":{"base_uri":"https://localhost:8080/"},"executionInfo":{"status":"ok","timestamp":1680282575833,"user_tz":-120,"elapsed":359683,"user":{"displayName":"Michał Janik","userId":"10483929333441523386"}},"outputId":"8dc91263-766a-49f5-d25a-c6879046da6c"},"execution_count":11,"outputs":[{"output_type":"stream","name":"stderr","text":["/usr/local/lib/python3.9/dist-packages/gym/envs/registration.py:593: UserWarning: \u001b[33mWARN: The environment CartPole-v0 is out of date. You should consider upgrading to version `v1`.\u001b[0m\n","  logger.warn(\n","/usr/local/lib/python3.9/dist-packages/gym/core.py:317: DeprecationWarning: \u001b[33mWARN: Initializing wrapper in old step API which returns one bool instead of two. It is recommended to set `new_step_api=True` to use new step API. This will be the default behaviour in future.\u001b[0m\n","  deprecation(\n","/usr/local/lib/python3.9/dist-packages/gym/wrappers/step_api_compatibility.py:39: DeprecationWarning: \u001b[33mWARN: Initializing environment in old step API which returns one bool instead of two. It is recommended to set `new_step_api=True` to use new step API. This will be the default behaviour in future.\u001b[0m\n","  deprecation(\n"]},{"output_type":"stream","name":"stdout","text":["Episode = 1 | steps =  15 | episode_length = 15 | epsilon = 0.597\n","Episode = 2 | steps =  28 | episode_length = 13 | epsilon = 0.594015\n","Episode = 3 | steps =  40 | episode_length = 12 | epsilon = 0.5910449249999999\n","Episode = 4 | steps =  71 | episode_length = 31 | epsilon = 0.588089700375\n","Episode = 5 | steps =  80 | episode_length = 9 | epsilon = 0.5851492518731249\n","Episode = 6 | steps =  91 | episode_length = 11 | epsilon = 0.5822235056137594\n","Episode = 7 | steps =  101 | episode_length = 10 | epsilon = 0.5793123880856905\n","Episode = 8 | steps =  112 | episode_length = 11 | epsilon = 0.5764158261452621\n","Episode = 9 | steps =  132 | episode_length = 20 | epsilon = 0.5735337470145357\n","Episode = 10 | steps =  144 | episode_length = 12 | epsilon = 0.570666078279463\n","Episode = 11 | steps =  154 | episode_length = 10 | epsilon = 0.5678127478880657\n","Episode = 12 | steps =  167 | episode_length = 13 | epsilon = 0.5649736841486254\n","Episode = 13 | steps =  179 | episode_length = 12 | epsilon = 0.5621488157278822\n","Episode = 14 | steps =  190 | episode_length = 11 | epsilon = 0.5593380716492428\n","Episode = 15 | steps =  211 | episode_length = 21 | epsilon = 0.5565413812909965\n","Episode = 16 | steps =  222 | episode_length = 11 | epsilon = 0.5537586743845415\n","Episode = 17 | steps =  235 | episode_length = 13 | epsilon = 0.5509898810126188\n","Episode = 18 | steps =  255 | episode_length = 20 | epsilon = 0.5482349316075558\n","Episode = 19 | steps =  275 | episode_length = 20 | epsilon = 0.545493756949518\n","Episode = 20 | steps =  296 | episode_length = 21 | epsilon = 0.5427662881647703\n","Episode = 21 | steps =  306 | episode_length = 10 | epsilon = 0.5400524567239465\n","Episode = 22 | steps =  343 | episode_length = 37 | epsilon = 0.5373521944403268\n","Episode = 23 | steps =  355 | episode_length = 12 | epsilon = 0.5346654334681251\n","Episode = 24 | steps =  374 | episode_length = 19 | epsilon = 0.5319921063007845\n","Episode = 25 | steps =  391 | episode_length = 17 | epsilon = 0.5293321457692806\n","Episode = 26 | steps =  399 | episode_length = 8 | epsilon = 0.5266854850404341\n","Episode = 27 | steps =  420 | episode_length = 21 | epsilon = 0.5240520576152319\n","Episode = 28 | steps =  457 | episode_length = 37 | epsilon = 0.5214317973271557\n","Episode = 29 | steps =  480 | episode_length = 23 | epsilon = 0.51882463834052\n","Episode = 30 | steps =  518 | episode_length = 38 | epsilon = 0.5162305151488173\n","Episode = 31 | steps =  547 | episode_length = 29 | epsilon = 0.5136493625730733\n","Episode = 32 | steps =  565 | episode_length = 18 | epsilon = 0.5110811157602079\n","Episode = 33 | steps =  608 | episode_length = 43 | epsilon = 0.5085257101814068\n","Episode = 34 | steps =  629 | episode_length = 21 | epsilon = 0.5059830816304998\n","Episode = 35 | steps =  650 | episode_length = 21 | epsilon = 0.5034531662223473\n","Episode = 36 | steps =  684 | episode_length = 34 | epsilon = 0.5009359003912355\n","Episode = 37 | steps =  708 | episode_length = 24 | epsilon = 0.4984312208892793\n","Episode = 38 | steps =  735 | episode_length = 27 | epsilon = 0.4959390647848329\n","Episode = 39 | steps =  752 | episode_length = 17 | epsilon = 0.49345936946090874\n","Episode = 40 | steps =  769 | episode_length = 17 | epsilon = 0.49099207261360417\n","Episode = 41 | steps =  779 | episode_length = 10 | epsilon = 0.48853711225053614\n","Episode = 42 | steps =  792 | episode_length = 13 | epsilon = 0.4860944266892835\n","Episode = 43 | steps =  816 | episode_length = 24 | epsilon = 0.48366395455583705\n","Episode = 44 | steps =  845 | episode_length = 29 | epsilon = 0.4812456347830579\n","Episode = 45 | steps =  861 | episode_length = 16 | epsilon = 0.4788394066091426\n","Episode = 46 | steps =  880 | episode_length = 19 | epsilon = 0.4764452095760969\n","Episode = 47 | steps =  924 | episode_length = 44 | epsilon = 0.4740629835282164\n","Episode = 48 | steps =  989 | episode_length = 65 | epsilon = 0.47169266861057535\n","Episode = 49 | steps =  1020 | episode_length = 31 | epsilon = 0.46933420526752245\n","Episode = 50 | steps =  1039 | episode_length = 19 | epsilon = 0.46698753424118483\n","Episode = 51 | steps =  1075 | episode_length = 36 | epsilon = 0.4646525965699789\n","Episode = 52 | steps =  1112 | episode_length = 37 | epsilon = 0.462329333587129\n","Episode = 53 | steps =  1144 | episode_length = 32 | epsilon = 0.46001768691919337\n","Episode = 54 | steps =  1186 | episode_length = 42 | epsilon = 0.4577175984845974\n","Episode = 55 | steps =  1207 | episode_length = 21 | epsilon = 0.45542901049217444\n","Episode = 56 | steps =  1285 | episode_length = 78 | epsilon = 0.4531518654397136\n","Episode = 57 | steps =  1299 | episode_length = 14 | epsilon = 0.450886106112515\n","Episode = 58 | steps =  1353 | episode_length = 54 | epsilon = 0.44863167558195244\n","Episode = 59 | steps =  1404 | episode_length = 51 | epsilon = 0.44638851720404266\n","Episode = 60 | steps =  1444 | episode_length = 40 | epsilon = 0.44415657461802244\n","Episode = 61 | steps =  1468 | episode_length = 24 | epsilon = 0.4419357917449323\n","Episode = 62 | steps =  1485 | episode_length = 17 | epsilon = 0.43972611278620766\n","Episode = 63 | steps =  1522 | episode_length = 37 | epsilon = 0.4375274822222766\n","Episode = 64 | steps =  1559 | episode_length = 37 | epsilon = 0.4353398448111652\n","Episode = 65 | steps =  1604 | episode_length = 45 | epsilon = 0.4331631455871094\n","Episode = 66 | steps =  1696 | episode_length = 92 | epsilon = 0.43099732985917383\n","Episode = 67 | steps =  1742 | episode_length = 46 | epsilon = 0.42884234320987796\n","Episode = 68 | steps =  1787 | episode_length = 45 | epsilon = 0.42669813149382857\n","Episode = 69 | steps =  1818 | episode_length = 31 | epsilon = 0.4245646408363594\n","Episode = 70 | steps =  1959 | episode_length = 141 | epsilon = 0.42244181763217764\n","Episode = 71 | steps =  1993 | episode_length = 34 | epsilon = 0.42032960854401674\n","Episode = 72 | steps =  2017 | episode_length = 24 | epsilon = 0.4182279605012967\n","Episode = 73 | steps =  2049 | episode_length = 32 | epsilon = 0.4161368206987902\n","Episode = 74 | steps =  2092 | episode_length = 43 | epsilon = 0.4140561365952963\n","Episode = 75 | steps =  2110 | episode_length = 18 | epsilon = 0.4119858559123198\n","Episode = 76 | steps =  2148 | episode_length = 38 | epsilon = 0.4099259266327582\n","Episode = 77 | steps =  2181 | episode_length = 33 | epsilon = 0.4078762969995944\n","Episode = 78 | steps =  2211 | episode_length = 30 | epsilon = 0.4058369155145964\n","Episode = 79 | steps =  2260 | episode_length = 49 | epsilon = 0.4038077309370234\n","Episode = 80 | steps =  2281 | episode_length = 21 | epsilon = 0.40178869228233827\n","Episode = 81 | steps =  2335 | episode_length = 54 | epsilon = 0.3997797488209266\n","Episode = 82 | steps =  2357 | episode_length = 22 | epsilon = 0.39778085007682196\n","Episode = 83 | steps =  2382 | episode_length = 25 | epsilon = 0.39579194582643784\n","Episode = 84 | steps =  2409 | episode_length = 27 | epsilon = 0.39381298609730564\n","Episode = 85 | steps =  2442 | episode_length = 33 | epsilon = 0.3918439211668191\n","Episode = 86 | steps =  2478 | episode_length = 36 | epsilon = 0.389884701560985\n","Episode = 87 | steps =  2497 | episode_length = 19 | epsilon = 0.38793527805318007\n","Episode = 88 | steps =  2518 | episode_length = 21 | epsilon = 0.3859956016629142\n","Episode = 89 | steps =  2546 | episode_length = 28 | epsilon = 0.3840656236545996\n","Episode = 90 | steps =  2574 | episode_length = 28 | epsilon = 0.38214529553632665\n","Episode = 91 | steps =  2601 | episode_length = 27 | epsilon = 0.380234569058645\n","Episode = 92 | steps =  2620 | episode_length = 19 | epsilon = 0.37833339621335177\n","Episode = 93 | steps =  2638 | episode_length = 18 | epsilon = 0.376441729232285\n","Episode = 94 | steps =  2650 | episode_length = 12 | epsilon = 0.3745595205861236\n","Episode = 95 | steps =  2676 | episode_length = 26 | epsilon = 0.37268672298319294\n","Episode = 96 | steps =  2705 | episode_length = 29 | epsilon = 0.37082328936827696\n","Episode = 97 | steps =  2727 | episode_length = 22 | epsilon = 0.36896917292143555\n","Episode = 98 | steps =  2788 | episode_length = 61 | epsilon = 0.36712432705682835\n","Episode = 99 | steps =  2829 | episode_length = 41 | epsilon = 0.3652887054215442\n","Episode = 100 | steps =  2849 | episode_length = 20 | epsilon = 0.3634622618944365\n","Episode = 101 | steps =  2868 | episode_length = 19 | epsilon = 0.3616449505849643\n","Episode = 102 | steps =  2897 | episode_length = 29 | epsilon = 0.3598367258320395\n","Episode = 103 | steps =  2930 | episode_length = 33 | epsilon = 0.3580375422028793\n","Episode = 104 | steps =  2949 | episode_length = 19 | epsilon = 0.3562473544918649\n","Episode = 105 | steps =  2975 | episode_length = 26 | epsilon = 0.35446611771940556\n","Episode = 106 | steps =  2991 | episode_length = 16 | epsilon = 0.3526937871308085\n","Episode = 107 | steps =  3019 | episode_length = 28 | epsilon = 0.35093031819515447\n","Episode = 108 | steps =  3033 | episode_length = 14 | epsilon = 0.3491756666041787\n","Episode = 109 | steps =  3046 | episode_length = 13 | epsilon = 0.3474297882711578\n","Episode = 110 | steps =  3075 | episode_length = 29 | epsilon = 0.345692639329802\n","Episode = 111 | steps =  3093 | episode_length = 18 | epsilon = 0.343964176133153\n","Episode = 112 | steps =  3127 | episode_length = 34 | epsilon = 0.34224435525248725\n","Episode = 113 | steps =  3143 | episode_length = 16 | epsilon = 0.3405331334762248\n","Episode = 114 | steps =  3158 | episode_length = 15 | epsilon = 0.3388304678088437\n","Episode = 115 | steps =  3180 | episode_length = 22 | epsilon = 0.33713631546979944\n","Episode = 116 | steps =  3202 | episode_length = 22 | epsilon = 0.33545063389245044\n","Episode = 117 | steps =  3222 | episode_length = 20 | epsilon = 0.3337733807229882\n","Episode = 118 | steps =  3246 | episode_length = 24 | epsilon = 0.3321045138193732\n","Episode = 119 | steps =  3272 | episode_length = 26 | epsilon = 0.33044399125027635\n","Episode = 120 | steps =  3293 | episode_length = 21 | epsilon = 0.328791771294025\n","Episode = 121 | steps =  3314 | episode_length = 21 | epsilon = 0.32714781243755486\n","Episode = 122 | steps =  3337 | episode_length = 23 | epsilon = 0.3255120733753671\n","Episode = 123 | steps =  3353 | episode_length = 16 | epsilon = 0.32388451300849025\n","Episode = 124 | steps =  3367 | episode_length = 14 | epsilon = 0.3222650904434478\n","Episode = 125 | steps =  3395 | episode_length = 28 | epsilon = 0.32065376499123055\n","Episode = 126 | steps =  3417 | episode_length = 22 | epsilon = 0.3190504961662744\n","Episode = 127 | steps =  3438 | episode_length = 21 | epsilon = 0.31745524368544303\n","Episode = 128 | steps =  3457 | episode_length = 19 | epsilon = 0.31586796746701584\n","Episode = 129 | steps =  3493 | episode_length = 36 | epsilon = 0.31428862762968074\n","Episode = 130 | steps =  3509 | episode_length = 16 | epsilon = 0.31271718449153235\n","Episode = 131 | steps =  3530 | episode_length = 21 | epsilon = 0.3111535985690747\n","Episode = 132 | steps =  3544 | episode_length = 14 | epsilon = 0.30959783057622936\n","Episode = 133 | steps =  3560 | episode_length = 16 | epsilon = 0.30804984142334824\n","Episode = 134 | steps =  3594 | episode_length = 34 | epsilon = 0.3065095922162315\n","Episode = 135 | steps =  3607 | episode_length = 13 | epsilon = 0.3049770442551503\n","Episode = 136 | steps =  3626 | episode_length = 19 | epsilon = 0.30345215903387457\n","Episode = 137 | steps =  3644 | episode_length = 18 | epsilon = 0.3019348982387052\n","Episode = 138 | steps =  3670 | episode_length = 26 | epsilon = 0.3004252237475116\n","Episode = 139 | steps =  3686 | episode_length = 16 | epsilon = 0.29892309762877406\n","Episode = 140 | steps =  3705 | episode_length = 19 | epsilon = 0.29742848214063017\n","Episode = 141 | steps =  3730 | episode_length = 25 | epsilon = 0.29594133972992703\n","Episode = 142 | steps =  3744 | episode_length = 14 | epsilon = 0.29446163303127737\n","Episode = 143 | steps =  3765 | episode_length = 21 | epsilon = 0.292989324866121\n","Episode = 144 | steps =  3783 | episode_length = 18 | epsilon = 0.2915243782417904\n","Episode = 145 | steps =  3802 | episode_length = 19 | epsilon = 0.29006675635058143\n","Episode = 146 | steps =  3828 | episode_length = 26 | epsilon = 0.2886164225688285\n","Episode = 147 | steps =  3845 | episode_length = 17 | epsilon = 0.28717334045598436\n","Episode = 148 | steps =  3871 | episode_length = 26 | epsilon = 0.28573747375370445\n","Episode = 149 | steps =  3889 | episode_length = 18 | epsilon = 0.28430878638493595\n","Episode = 150 | steps =  3915 | episode_length = 26 | epsilon = 0.2828872424530113\n","Episode = 151 | steps =  3944 | episode_length = 29 | epsilon = 0.2814728062407462\n","Episode = 152 | steps =  3962 | episode_length = 18 | epsilon = 0.2800654422095425\n","Episode = 153 | steps =  3985 | episode_length = 23 | epsilon = 0.2786651149984948\n","Episode = 154 | steps =  4006 | episode_length = 21 | epsilon = 0.2772717894235023\n","Episode = 155 | steps =  4040 | episode_length = 34 | epsilon = 0.27588543047638475\n","Episode = 156 | steps =  4070 | episode_length = 30 | epsilon = 0.2745060033240028\n","Episode = 157 | steps =  4093 | episode_length = 23 | epsilon = 0.2731334733073828\n","Episode = 158 | steps =  4122 | episode_length = 29 | epsilon = 0.2717678059408459\n","Episode = 159 | steps =  4138 | episode_length = 16 | epsilon = 0.27040896691114163\n","Episode = 160 | steps =  4156 | episode_length = 18 | epsilon = 0.26905692207658594\n","Episode = 161 | steps =  4179 | episode_length = 23 | epsilon = 0.267711637466203\n","Episode = 162 | steps =  4220 | episode_length = 41 | epsilon = 0.266373079278872\n","Episode = 163 | steps =  4237 | episode_length = 17 | epsilon = 0.26504121388247764\n","Episode = 164 | steps =  4266 | episode_length = 29 | epsilon = 0.26371600781306526\n","Episode = 165 | steps =  4291 | episode_length = 25 | epsilon = 0.26239742777399994\n","Episode = 166 | steps =  4311 | episode_length = 20 | epsilon = 0.26108544063512995\n","Episode = 167 | steps =  4361 | episode_length = 50 | epsilon = 0.2597800134319543\n","Episode = 168 | steps =  4380 | episode_length = 19 | epsilon = 0.2584811133647945\n","Episode = 169 | steps =  4396 | episode_length = 16 | epsilon = 0.25718870779797054\n","Episode = 170 | steps =  4421 | episode_length = 25 | epsilon = 0.2559027642589807\n","Episode = 171 | steps =  4446 | episode_length = 25 | epsilon = 0.25462325043768574\n","Episode = 172 | steps =  4471 | episode_length = 25 | epsilon = 0.2533501341854973\n","Episode = 173 | steps =  4489 | episode_length = 18 | epsilon = 0.2520833835145698\n","Episode = 174 | steps =  4513 | episode_length = 24 | epsilon = 0.250822966596997\n","Episode = 175 | steps =  4531 | episode_length = 18 | epsilon = 0.249568851764012\n","Episode = 176 | steps =  4564 | episode_length = 33 | epsilon = 0.24832100750519193\n","Episode = 177 | steps =  4591 | episode_length = 27 | epsilon = 0.24707940246766597\n","Episode = 178 | steps =  4615 | episode_length = 24 | epsilon = 0.24584400545532764\n","Episode = 179 | steps =  4642 | episode_length = 27 | epsilon = 0.244614785428051\n","Episode = 180 | steps =  4665 | episode_length = 23 | epsilon = 0.24339171150091077\n","Episode = 181 | steps =  4692 | episode_length = 27 | epsilon = 0.2421747529434062\n","Episode = 182 | steps =  4715 | episode_length = 23 | epsilon = 0.24096387917868917\n","Episode = 183 | steps =  4733 | episode_length = 18 | epsilon = 0.2397590597827957\n","Episode = 184 | steps =  4756 | episode_length = 23 | epsilon = 0.23856026448388173\n","Episode = 185 | steps =  4779 | episode_length = 23 | epsilon = 0.23736746316146232\n","Episode = 186 | steps =  4800 | episode_length = 21 | epsilon = 0.236180625845655\n","Episode = 187 | steps =  4821 | episode_length = 21 | epsilon = 0.23499972271642672\n","Episode = 188 | steps =  4840 | episode_length = 19 | epsilon = 0.23382472410284458\n","Episode = 189 | steps =  4863 | episode_length = 23 | epsilon = 0.23265560048233036\n","Episode = 190 | steps =  4892 | episode_length = 29 | epsilon = 0.23149232247991872\n","Episode = 191 | steps =  4912 | episode_length = 20 | epsilon = 0.23033486086751911\n","Episode = 192 | steps =  4937 | episode_length = 25 | epsilon = 0.2291831865631815\n","Episode = 193 | steps =  4958 | episode_length = 21 | epsilon = 0.2280372706303656\n","Episode = 194 | steps =  4977 | episode_length = 19 | epsilon = 0.22689708427721378\n","Episode = 195 | steps =  5004 | episode_length = 27 | epsilon = 0.2257625988558277\n","Episode = 196 | steps =  5027 | episode_length = 23 | epsilon = 0.22463378586154856\n","Episode = 197 | steps =  5047 | episode_length = 20 | epsilon = 0.2235106169322408\n","Episode = 198 | steps =  5074 | episode_length = 27 | epsilon = 0.2223930638475796\n","Episode = 199 | steps =  5097 | episode_length = 23 | epsilon = 0.22128109852834169\n","Episode = 200 | steps =  5117 | episode_length = 20 | epsilon = 0.22017469303569998\n","Episode = 201 | steps =  5138 | episode_length = 21 | epsilon = 0.21907381957052147\n","Episode = 202 | steps =  5161 | episode_length = 23 | epsilon = 0.21797845047266887\n","Episode = 203 | steps =  5182 | episode_length = 21 | epsilon = 0.21688855822030553\n","Episode = 204 | steps =  5203 | episode_length = 21 | epsilon = 0.215804115429204\n","Episode = 205 | steps =  5226 | episode_length = 23 | epsilon = 0.21472509485205798\n","Episode = 206 | steps =  5250 | episode_length = 24 | epsilon = 0.2136514693777977\n","Episode = 207 | steps =  5274 | episode_length = 24 | epsilon = 0.2125832120309087\n","Episode = 208 | steps =  5298 | episode_length = 24 | epsilon = 0.21152029597075417\n","Episode = 209 | steps =  5332 | episode_length = 34 | epsilon = 0.2104626944909004\n","Episode = 210 | steps =  5358 | episode_length = 26 | epsilon = 0.20941038101844592\n","Episode = 211 | steps =  5390 | episode_length = 32 | epsilon = 0.2083633291133537\n","Episode = 212 | steps =  5411 | episode_length = 21 | epsilon = 0.20732151246778693\n","Episode = 213 | steps =  5432 | episode_length = 21 | epsilon = 0.20628490490544799\n","Episode = 214 | steps =  5459 | episode_length = 27 | epsilon = 0.20525348038092076\n","Episode = 215 | steps =  5483 | episode_length = 24 | epsilon = 0.20422721297901614\n","Episode = 216 | steps =  5509 | episode_length = 26 | epsilon = 0.20320607691412104\n","Episode = 217 | steps =  5550 | episode_length = 41 | epsilon = 0.20219004652955044\n","Episode = 218 | steps =  5583 | episode_length = 33 | epsilon = 0.2011790962969027\n","Episode = 219 | steps =  5621 | episode_length = 38 | epsilon = 0.20017320081541817\n","Episode = 220 | steps =  5650 | episode_length = 29 | epsilon = 0.19917233481134108\n","Episode = 221 | steps =  5682 | episode_length = 32 | epsilon = 0.19817647313728437\n","Episode = 222 | steps =  5715 | episode_length = 33 | epsilon = 0.19718559077159795\n","Episode = 223 | steps =  5747 | episode_length = 32 | epsilon = 0.19619966281773996\n","Episode = 224 | steps =  5779 | episode_length = 32 | epsilon = 0.19521866450365125\n","Episode = 225 | steps =  5805 | episode_length = 26 | epsilon = 0.194242571181133\n","Episode = 226 | steps =  5842 | episode_length = 37 | epsilon = 0.19327135832522732\n","Episode = 227 | steps =  5869 | episode_length = 27 | epsilon = 0.1923050015336012\n","Episode = 228 | steps =  5901 | episode_length = 32 | epsilon = 0.1913434765259332\n","Episode = 229 | steps =  5932 | episode_length = 31 | epsilon = 0.19038675914330352\n","Episode = 230 | steps =  5961 | episode_length = 29 | epsilon = 0.18943482534758702\n","Episode = 231 | steps =  5988 | episode_length = 27 | epsilon = 0.18848765122084907\n","Episode = 232 | steps =  6018 | episode_length = 30 | epsilon = 0.18754521296474483\n","Episode = 233 | steps =  6042 | episode_length = 24 | epsilon = 0.1866074868999211\n","Episode = 234 | steps =  6065 | episode_length = 23 | epsilon = 0.1856744494654215\n","Episode = 235 | steps =  6092 | episode_length = 27 | epsilon = 0.18474607721809438\n","Episode = 236 | steps =  6121 | episode_length = 29 | epsilon = 0.1838223468320039\n","Episode = 237 | steps =  6145 | episode_length = 24 | epsilon = 0.1829032350978439\n","Episode = 238 | steps =  6184 | episode_length = 39 | epsilon = 0.1819887189223547\n","Episode = 239 | steps =  6215 | episode_length = 31 | epsilon = 0.1810787753277429\n","Episode = 240 | steps =  6256 | episode_length = 41 | epsilon = 0.1801733814511042\n","Episode = 241 | steps =  6282 | episode_length = 26 | epsilon = 0.17927251454384868\n","Episode = 242 | steps =  6311 | episode_length = 29 | epsilon = 0.17837615197112944\n","Episode = 243 | steps =  6347 | episode_length = 36 | epsilon = 0.1774842712112738\n","Episode = 244 | steps =  6380 | episode_length = 33 | epsilon = 0.17659684985521742\n","Episode = 245 | steps =  6418 | episode_length = 38 | epsilon = 0.17571386560594135\n","Episode = 246 | steps =  6457 | episode_length = 39 | epsilon = 0.17483529627791164\n","Episode = 247 | steps =  6498 | episode_length = 41 | epsilon = 0.17396111979652207\n","Episode = 248 | steps =  6536 | episode_length = 38 | epsilon = 0.17309131419753945\n","Episode = 249 | steps =  6570 | episode_length = 34 | epsilon = 0.17222585762655176\n","Episode = 250 | steps =  6607 | episode_length = 37 | epsilon = 0.171364728338419\n","Episode = 251 | steps =  6645 | episode_length = 38 | epsilon = 0.17050790469672691\n","Episode = 252 | steps =  6674 | episode_length = 29 | epsilon = 0.16965536517324328\n","Episode = 253 | steps =  6707 | episode_length = 33 | epsilon = 0.16880708834737707\n","Episode = 254 | steps =  6828 | episode_length = 121 | epsilon = 0.16796305290564018\n","Episode = 255 | steps =  6916 | episode_length = 88 | epsilon = 0.16712323764111198\n","Episode = 256 | steps =  7077 | episode_length = 161 | epsilon = 0.1662876214529064\n","Episode = 257 | steps =  7159 | episode_length = 82 | epsilon = 0.16545618334564188\n","Episode = 258 | steps =  7243 | episode_length = 84 | epsilon = 0.16462890242891368\n","Episode = 259 | steps =  7333 | episode_length = 90 | epsilon = 0.1638057579167691\n","Episode = 260 | steps =  7430 | episode_length = 97 | epsilon = 0.16298672912718526\n","Episode = 261 | steps =  7522 | episode_length = 92 | epsilon = 0.16217179548154934\n","Episode = 262 | steps =  7617 | episode_length = 95 | epsilon = 0.1613609365041416\n","Episode = 263 | steps =  7704 | episode_length = 87 | epsilon = 0.1605541318216209\n","Episode = 264 | steps =  7786 | episode_length = 82 | epsilon = 0.1597513611625128\n","Episode = 265 | steps =  7906 | episode_length = 120 | epsilon = 0.15895260435670022\n","Episode = 266 | steps =  8013 | episode_length = 107 | epsilon = 0.15815784133491673\n"]}]},{"cell_type":"markdown","source":["Let us now plot graphs:"],"metadata":{"id":"hLAS8exSTBLN"}},{"cell_type":"code","source":["def visualize_progress(progress, loss_history):\n","  plt.clf()\n","  plt.plot(progress, label=\"DQN progress\")\n","  smoothed_progress = [0]\n","  for x in progress:\n","    smoothed_progress.append(0.8*smoothed_progress[-1] + 0.2*x)\n","  plt.plot(smoothed_progress, label=\"DQN learning (smoothed)\")\n","  plt.legend(loc=\"upper left\")\n","  plt.show()\n","\n","  plt.clf()\n","  plt.plot(loss_history, label=\"Loss\")\n","  plt.legend(loc=\"upper left\")\n","  plt.show()"],"metadata":{"id":"NUkrjbY41kxF","executionInfo":{"status":"ok","timestamp":1680282575834,"user_tz":-120,"elapsed":17,"user":{"displayName":"Michał Janik","userId":"10483929333441523386"}}},"execution_count":12,"outputs":[]},{"cell_type":"code","source":["visualize_progress(progress_standard, loss_history_standard)"],"metadata":{"id":"kovVfKlO6CrJ","colab":{"base_uri":"https://localhost:8080/","height":843},"executionInfo":{"status":"ok","timestamp":1680282576659,"user_tz":-120,"elapsed":828,"user":{"displayName":"Michał Janik","userId":"10483929333441523386"}},"outputId":"665312ea-455f-47a3-cd83-83f55151f719"},"execution_count":13,"outputs":[{"output_type":"display_data","data":{"text/plain":["<Figure size 640x480 with 1 Axes>"],"image/png":"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\n"},"metadata":{}},{"output_type":"display_data","data":{"text/plain":["<Figure size 640x480 with 1 Axes>"],"image/png":"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\n"},"metadata":{}}]},{"cell_type":"markdown","source":["Let us see how the agent performs across the training:"],"metadata":{"id":"VokUnDv0GszE"}},{"cell_type":"code","source":["def record_checkpoint(checkpoint):\n","  # This function records an episode of the agent equipped with a given checkpoint\n","  env = gym.make(\"CartPole-v0\")\n","  model = make_cartpole_network()\n","  model.set_weights(checkpoint)\n","  max_ep_len=200\n","  envw = gym.wrappers.RecordVideo(env, 'video')\n","  o, d, ep_len = envw.reset(), False, 0\n","  envw.start_video_recorder()\n","  while not (d or (ep_len == max_ep_len)):\n","    action = choose_best_action(model, o)\n","    o, r, d, info = envw.step(action)\n","  envw.close()"],"metadata":{"id":"1K0kLsJ7hN6h","executionInfo":{"status":"ok","timestamp":1680282576660,"user_tz":-120,"elapsed":5,"user":{"displayName":"Michał Janik","userId":"10483929333441523386"}}},"execution_count":14,"outputs":[]},{"cell_type":"markdown","source":["Lets take a look at first saved chekpoint:"],"metadata":{"id":"ZmLU4eEgIxno"}},{"cell_type":"code","source":["record_checkpoint(q_checkpoints_standard[0])\n","file_name = glob.glob('video/rl-video-episode-*.mp4')[0]\n","show_video(file_name)"],"metadata":{"id":"zTkCCWwwjqc-","colab":{"base_uri":"https://localhost:8080/","height":593},"executionInfo":{"status":"ok","timestamp":1680282578628,"user_tz":-120,"elapsed":1972,"user":{"displayName":"Michał Janik","userId":"10483929333441523386"}},"outputId":"cb2f28e0-b8eb-4e39-a7f4-e6114d3e5c74"},"execution_count":15,"outputs":[{"output_type":"stream","name":"stderr","text":["/usr/local/lib/python3.9/dist-packages/gym/wrappers/monitoring/video_recorder.py:78: DeprecationWarning: \u001b[33mWARN: Recording ability for environment CartPole-v0 initialized with `render_mode=None` is marked as deprecated and will be removed in the future.\u001b[0m\n","  logger.deprecation(\n","/usr/local/lib/python3.9/dist-packages/gym/core.py:43: DeprecationWarning: \u001b[33mWARN: The argument mode in render method is deprecated; use render_mode during environment initialization instead.\n","See here for more information: https://www.gymlibrary.ml/content/api/\u001b[0m\n","  deprecation(\n","/usr/local/lib/python3.9/dist-packages/gym/wrappers/monitoring/video_recorder.py:78: DeprecationWarning: \u001b[33mWARN: Recording ability for environment CartPole-v0 initialized with `render_mode=None` is marked as deprecated and will be removed in the future.\u001b[0m\n","  logger.deprecation(\n","/usr/local/lib/python3.9/dist-packages/gym/core.py:43: DeprecationWarning: \u001b[33mWARN: The argument mode in render method is deprecated; use render_mode during environment initialization instead.\n","See here for more information: https://www.gymlibrary.ml/content/api/\u001b[0m\n","  deprecation(\n","/usr/local/lib/python3.9/dist-packages/gym/core.py:43: DeprecationWarning: \u001b[33mWARN: The argument mode in render method is deprecated; use render_mode during environment initialization instead.\n","See here for more information: https://www.gymlibrary.ml/content/api/\u001b[0m\n","  deprecation(\n"]},{"output_type":"execute_result","data":{"text/plain":["<IPython.core.display.HTML object>"],"text/html":["\n","    <video width=480 controls>\n","        <source src=\"data:video/mp4;base64,AAAAIGZ0eXBpc29tAAACAGlzb21pc28yYXZjMW1wNDEAAAAIZnJlZQAACg1tZGF0AAACrgYF//+q3EXpvebZSLeWLNgg2SPu73gyNjQgLSBjb3JlIDE1NSByMjkxNyAwYTg0ZDk4IC0gSC4yNjQvTVBFRy00IEFWQyBjb2RlYyAtIENvcHlsZWZ0IDIwMDMtMjAxOCAtIGh0dHA6Ly93d3cudmlkZW9sYW4ub3JnL3gyNjQuaHRtbCAtIG9wdGlvbnM6IGNhYmFjPTEgcmVmPTMgZGVibG9jaz0xOjA6MCBhbmFseXNlPTB4MzoweDExMyBtZT1oZXggc3VibWU9NyBwc3k9MSBwc3lfcmQ9MS4wMDowLjAwIG1peGVkX3JlZj0xIG1lX3JhbmdlPTE2IGNocm9tYV9tZT0xIHRyZWxsaXM9MSA4eDhkY3Q9MSBjcW09MCBkZWFkem9uZT0yMSwxMSBmYXN0X3Bza2lwPTEgY2hyb21hX3FwX29mZnNldD0tMiB0aHJlYWRzPTMgbG9va2FoZWFkX3RocmVhZHM9MSBzbGljZWRfdGhyZWFkcz0wIG5yPTAgZGVjaW1hdGU9MSBpbnRlcmxhY2VkPTAgYmx1cmF5X2NvbXBhdD0wIGNvbnN0cmFpbmVkX2ludHJhPTAgYmZyYW1lcz0zIGJfcHlyYW1pZD0yIGJfYWRhcHQ9MSBiX2JpYXM9MCBkaXJlY3Q9MSB3ZWlnaHRiPTEgb3Blbl9nb3A9MCB3ZWlnaHRwPTIga2V5aW50PTI1MCBrZXlpbnRfbWluPTI1IHNjZW5lY3V0PTQwIGludHJhX3JlZnJlc2g9MCByY19sb29rYWhlYWQ9NDAgcmM9Y3JmIG1idHJlZT0xIGNyZj0yMy4wIHFjb21wPTAuNjAgcXBtaW49MCBxcG1heD02OSBxcHN0ZXA9NCBpcF9yYXRpbz0xLjQwIGFxPTE6MS4wMACAAAACB2WIhAAz//727L4FNf2f0JcRLMXaSnA+KqSAgHc0wAAAAwAAAwAAFgn0I7DkqgN3QAAAHGAFBCwCPCVC2EhH2OkN/wTj6yAOOkz9DSd0e4Ltidrv2b7f14mtP8wPmbboTmK1tRz6cHkULhNH8DX9J5IsC+66TfejyM2hQMtcax7H7VxXd9Fhp8WAQcDF0n60M5kdW0oRIkfXfYd8ulyn8E72mZcgMa33PRIhKp+JBVHkCvz6RKVKt4UAKG2pbJsA4K5xxEOnTfDtGimF3ivOIv34xnqoTJkADj8f79ZVZ30va6FNZUv3kcXMdHe28QPwwC6+5gQ/6kAp40inyXUAzYLqlOFBAFwpk3Cr0VJAJ28p6lT6ix+KxBZm2W9CvA0c+fVbqf7jcR0NFICUvi5SQJ05R6Q8TXAqSgOKTzzSAswiQtvzTTgsAKBUz+EdzW1PlVmuQTkC470XBdrpTxk313fJAUkbHhic0RQzQjCdjmbyhNw/3JQ3SDf+rwgFEL80/x2LB3Suf9w26XhBNKc6AZc+hRhnvouUQXF5JJcjr9vkQvnOZufvbvfZl+nxpC2551IfyHDq0D+YPex8z89fK6ggABSVXoEfvg2FEUjskFqWPBzXWaemfXJnzN8HzjAhEAcR0jKl12OxVmiouHyI0L+iL48u4uNAACtGOkdEoAAAAwAAAwAChwAAAFNBmiRsQv/+jLAAAAoPh8ckUXJMZIRghDNr9XNmoyeyC0DYp/K7G2qDMzWQA2ndsyoz8mgznFEBuyVUZJDWnDd/bWIrBHM+19iCGW66czCOItjoQAAAAClBnkJ4hH8AAAMDOXGHwACMmygSsFUNXTlSXbh7FCme69oDuK6ZMc8tmQAAABMBnmF0R/8AAAMAAjsKkm3q33jTAAAAIgGeY2pH/wAABR8wl0vsItRLe3+icRqHqmr6L10L0ii3E4EAAABRQZpoSahBaJlMCF///oywAAAar0e0AQC7YUz7lZ942dfQeVvgRSdHrvOyuRDGsrA18ogBdYHsPlPRKZ4mBwyUb0SPz/1w9EWgyrldl9brDMIRAAAAI0GehkURLCP/AAADAzhNZu1rS9VGjEUqUcw82Ukvd3fPcRvhAAAAHQGepXRH/wAABR/L7RU0qpWkNegC0FzRtDQ3rl3RAAAADgGep2pH/wAAAwAAAwGpAAAAYEGarEmoQWyZTAhf//6MsAAARgL28noQdVHCA79lYr8r9XjG9qgmfcnYKfSV0hJVub38hTBXtWMwlNsRjD7qPiBNLVtBbPGojjkDh3wThh08O6ojgmPasPJ+2cmp/hGx8AAAACxBnspFFSwj/wAAFrUrNP2bc6hhvasqbWKfKMf1bDdlGu4wYLV2uimijBVCFwAAAB4Bnul0R/8AACOr+F16b/1UnwHOAL4kXdgbQpIDUgIAAAAbAZ7rakf/AAAFHzCWe4ktM1cKz3oaHwCClTJMAAAAuEGa8EmoQWyZTAhX//44QAABDuJeBFoQV/dTfp3bsGfupEoQTtoZH+WrRrLDHdL79mJiLF+yuCHUWyxS0NQU7ebmnQQrRD3TXyG4xpneqaHPuP9jxvY8fbWmIoW889npoWMASuE1NdzfgIZePBeA18CQqaNpkysZJlrOG/dqwotiZsmoaGT7ZcllLoXbrqlGXu2tZc1LHGAO1OqxsKmTgjXWDt4UYdX6zN3ZRcPEBsRCK6kR7za+NLUAAABKQZ8ORRUsI/8AABa7X260W6OCHIlVTdiwngYhT1LFN3GNKh+FZDg1THzSwBDW2AC6AlLxtX74FND6rIMycwjKqdaMT6AY8wS1t8EAAAAuAZ8tdEf/AAAjwLjfwGUuv2WLe5iUs/SrGA+xpEIPtK3UNg91kiJ06zSMeHYSTQAAAEsBny9qR/8AACO9la24/w9j2nP1N8cdmLjbn5uYRkTnvjm2U6QYgzkk1wZ5fnq7GVEJX5UI5RhbjCNP18AAz43L+7+UMrO5swDYB24AAAClQZsySahBbJlMFEwj//3hAAAEG6JrVu6gUI6rZly8tgf+iQYCol2bfinmQAsYMrrGZRoe/jGzV4aLHcPlPCuiqfY7oiWeQHUAaZDF9izvNIlVDm7LkrmHadb8AIhYhpu8xIaURByCevQIh3UsaLF785Yk6lSZzuletbmONnLslqH2I2sUK1Z2fGKdA52C0lSYWiC4iFyPLyhNcljPPuzYsnkYdupAAAAAWQGfUWpH/wAAI7HnbTGY5aOv6oSa7zoQdgBjqfNBohUx6D7TpvbsSVOVhR6Kk0+4C3ynwMdXphp6z9ACUHQ4L5UhQXZ2ppJu0TkNbKYqnVWYuUwxGtwSyzitAAAAbkGbU0nhClJlMCP//IQAAA97+zgsObLLmVbQfAkagAG6AgRjA4KgWF/ju8sqtVgUEgd2fgNB3qroDRaI7qN6qSqQA6Vs1wJL56NWgcznGijG6svKZpMkslqlbCd6jJVbxuqXgoiiZ/KRqqYwm+6AAAAEA21vb3YAAABsbXZoZAAAAAAAAAAAAAAAAAAAA+gAAAGQAAEAAAEAAAAAAAAAAAAAAAABAAAAAAAAAAAAAAAAAAAAAQAAAAAAAAAAAAAAAAAAQAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAIAAAMtdHJhawAAAFx0a2hkAAAAAwAAAAAAAAAAAAAAAQAAAAAAAAGQAAAAAAAAAAAAAAAAAAAAAAABAAAAAAAAAAAAAAAAAAAAAQAAAAAAAAAAAAAAAAAAQAAAAAJYAAABkAAAAAAAJGVkdHMAAAAcZWxzdAAAAAAAAAABAAABkAAAAgAAAQAAAAACpW1kaWEAAAAgbWRoZAAAAAAAAAAAAAAAAAAAMgAAABQAVcQAAAAAAC1oZGxyAAAAAAAAAAB2aWRlAAAAAAAAAAAAAAAAVmlkZW9IYW5kbGVyAAAAAlBtaW5mAAAAFHZtaGQAAAABAAAAAAAAAAAAAAAkZGluZgAAABxkcmVmAAAAAAAAAAEAAAAMdXJsIAAAAAEAAAIQc3RibAAAAJhzdHNkAAAAAAAAAAEAAACIYXZjMQAAAAAAAAABAAAAAAAAAAAAAAAAAAAAAAJYAZAASAAAAEgAAAAAAAAAAQAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAABj//wAAADJhdmNDAWQAH//hABlnZAAfrNlAmDPl4QAAAwABAAADAGQPGDGWAQAGaOvjyyLAAAAAGHN0dHMAAAAAAAAAAQAAABQAAAEAAAAAFHN0c3MAAAAAAAAAAQAAAAEAAACwY3R0cwAAAAAAAAAUAAAAAQAAAgAAAAABAAAFAAAAAAEAAAIAAAAAAQAAAAAAAAABAAABAAAAAAEAAAUAAAAAAQAAAgAAAAABAAAAAAAAAAEAAAEAAAAAAQAABQAAAAABAAACAAAAAAEAAAAAAAAAAQAAAQAAAAABAAAFAAAAAAEAAAIAAAAAAQAAAAAAAAABAAABAAAAAAEAAAMAAAAAAQAAAQAAAAABAAACAAAAABxzdHNjAAAAAAAAAAEAAAABAAAAFAAAAAEAAABkc3RzegAAAAAAAAAAAAAAFAAABL0AAABXAAAALQAAABcAAAAmAAAAVQAAACcAAAAhAAAAEgAAAGQAAAAwAAAAIgAAAB8AAAC8AAAATgAAADIAAABPAAAAqQAAAF0AAAByAAAAFHN0Y28AAAAAAAAAAQAAADAAAABidWR0YQAAAFptZXRhAAAAAAAAACFoZGxyAAAAAAAAAABtZGlyYXBwbAAAAAAAAAAAAAAAAC1pbHN0AAAAJal0b28AAAAdZGF0YQAAAAEAAAAATGF2ZjU4LjI5LjEwMA==\" type=\"video/mp4\">\n","    </video>\n","    "]},"metadata":{},"execution_count":15}]},{"cell_type":"markdown","source":["And the last:"],"metadata":{"id":"HeN9OPMVJBsX"}},{"cell_type":"code","source":["record_checkpoint(q_checkpoints_standard[-1])\n","file_name = glob.glob('video/rl-video-episode-*.mp4')[0]\n","show_video(file_name)"],"metadata":{"id":"TcrqaiHuJA2O","colab":{"base_uri":"https://localhost:8080/","height":718},"executionInfo":{"status":"ok","timestamp":1680282585306,"user_tz":-120,"elapsed":6683,"user":{"displayName":"Michał Janik","userId":"10483929333441523386"}},"outputId":"27f7b478-f011-4fb0-f726-9fe95f4fa5ab"},"execution_count":16,"outputs":[{"output_type":"stream","name":"stderr","text":["/usr/local/lib/python3.9/dist-packages/ipykernel/ipkernel.py:283: DeprecationWarning: `should_run_async` will not call `transform_cell` automatically in the future. Please pass the result to `transformed_cell` argument and any exception that happen during thetransform in `preprocessing_exc_tuple` in IPython 7.17 and above.\n","  and should_run_async(code)\n","/usr/local/lib/python3.9/dist-packages/gym/envs/registration.py:593: UserWarning: \u001b[33mWARN: The environment CartPole-v0 is out of date. You should consider upgrading to version `v1`.\u001b[0m\n","  logger.warn(\n","/usr/local/lib/python3.9/dist-packages/gym/core.py:317: DeprecationWarning: \u001b[33mWARN: Initializing wrapper in old step API which returns one bool instead of two. It is recommended to set `new_step_api=True` to use new step API. This will be the default behaviour in future.\u001b[0m\n","  deprecation(\n","/usr/local/lib/python3.9/dist-packages/gym/wrappers/step_api_compatibility.py:39: DeprecationWarning: \u001b[33mWARN: Initializing environment in old step API which returns one bool instead of two. It is recommended to set `new_step_api=True` to use new step API. This will be the default behaviour in future.\u001b[0m\n","  deprecation(\n","/usr/local/lib/python3.9/dist-packages/gym/wrappers/record_video.py:78: UserWarning: \u001b[33mWARN: Overwriting existing videos at /content/video folder (try specifying a different `video_folder` for the `RecordVideo` wrapper if this is not desired)\u001b[0m\n","  logger.warn(\n","/usr/local/lib/python3.9/dist-packages/gym/wrappers/monitoring/video_recorder.py:78: DeprecationWarning: \u001b[33mWARN: Recording ability for environment CartPole-v0 initialized with `render_mode=None` is marked as deprecated and will be removed in the future.\u001b[0m\n","  logger.deprecation(\n","/usr/local/lib/python3.9/dist-packages/gym/wrappers/monitoring/video_recorder.py:78: DeprecationWarning: \u001b[33mWARN: Recording ability for environment CartPole-v0 initialized with `render_mode=None` is marked as deprecated and will be removed in the future.\u001b[0m\n","  logger.deprecation(\n","/usr/local/lib/python3.9/dist-packages/gym/core.py:43: DeprecationWarning: \u001b[33mWARN: The argument mode in render method is deprecated; use render_mode during environment initialization instead.\n","See here for more information: https://www.gymlibrary.ml/content/api/\u001b[0m\n","  deprecation(\n","/usr/local/lib/python3.9/dist-packages/gym/core.py:43: DeprecationWarning: \u001b[33mWARN: The argument mode in render method is deprecated; use render_mode during environment initialization instead.\n","See here for more information: https://www.gymlibrary.ml/content/api/\u001b[0m\n","  deprecation(\n"]},{"output_type":"execute_result","data":{"text/plain":["<IPython.core.display.HTML object>"],"text/html":["\n","    <video width=480 controls>\n","        <source src=\"data:video/mp4;base64,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\" type=\"video/mp4\">\n","    </video>\n","    "]},"metadata":{},"execution_count":16}]},{"cell_type":"markdown","source":["#Ablation study\n","Let's see the what happens to DQN performance after turning off some of its mechanisms:\n","- target network\n","- sampling from replay_buffer\n","\n","**Exercise: turn off the usage of target networks.** You can for example modify the code of run_dqn() and set target_network = q_network. Compare the results with previous run.\n","\n","**Exercise: add the size limit to replay buffer.** Add a code to run_dqn() that clips its size to a given limit**. What happens if the replay buffer is very small?"],"metadata":{"id":"edjScjA4KLJJ"}},{"cell_type":"code","source":["progress_no_target, loss_history_no_target, _ = run_dqn(8000, use_target_network=False)"],"metadata":{"id":"tx7GxfgMJkVc","colab":{"base_uri":"https://localhost:8080/"},"executionInfo":{"status":"ok","timestamp":1680282950051,"user_tz":-120,"elapsed":364747,"user":{"displayName":"Michał Janik","userId":"10483929333441523386"}},"outputId":"e2b826f7-2186-444b-d852-de3c9b530fc7"},"execution_count":17,"outputs":[{"output_type":"stream","name":"stderr","text":["/usr/local/lib/python3.9/dist-packages/ipykernel/ipkernel.py:283: DeprecationWarning: `should_run_async` will not call `transform_cell` automatically in the future. Please pass the result to `transformed_cell` argument and any exception that happen during thetransform in `preprocessing_exc_tuple` in IPython 7.17 and above.\n","  and should_run_async(code)\n","/usr/local/lib/python3.9/dist-packages/gym/envs/registration.py:593: UserWarning: \u001b[33mWARN: The environment CartPole-v0 is out of date. You should consider upgrading to version `v1`.\u001b[0m\n","  logger.warn(\n","/usr/local/lib/python3.9/dist-packages/gym/core.py:317: DeprecationWarning: \u001b[33mWARN: Initializing wrapper in old step API which returns one bool instead of two. It is recommended to set `new_step_api=True` to use new step API. This will be the default behaviour in future.\u001b[0m\n","  deprecation(\n","/usr/local/lib/python3.9/dist-packages/gym/wrappers/step_api_compatibility.py:39: DeprecationWarning: \u001b[33mWARN: Initializing environment in old step API which returns one bool instead of two. It is recommended to set `new_step_api=True` to use new step API. This will be the default behaviour in future.\u001b[0m\n","  deprecation(\n"]},{"output_type":"stream","name":"stdout","text":["Episode = 1 | steps =  30 | episode_length = 30 | epsilon = 0.597\n","Episode = 2 | steps =  42 | episode_length = 12 | epsilon = 0.594015\n","Episode = 3 | steps =  58 | episode_length = 16 | epsilon = 0.5910449249999999\n","Episode = 4 | steps =  82 | episode_length = 24 | epsilon = 0.588089700375\n","Episode = 5 | steps =  100 | episode_length = 18 | epsilon = 0.5851492518731249\n","Episode = 6 | steps =  119 | episode_length = 19 | epsilon = 0.5822235056137594\n","Episode = 7 | steps =  135 | episode_length = 16 | epsilon = 0.5793123880856905\n","Episode = 8 | steps =  156 | episode_length = 21 | epsilon = 0.5764158261452621\n","Episode = 9 | steps =  168 | episode_length = 12 | epsilon = 0.5735337470145357\n","Episode = 10 | steps =  186 | episode_length = 18 | epsilon = 0.570666078279463\n","Episode = 11 | steps =  196 | episode_length = 10 | epsilon = 0.5678127478880657\n","Episode = 12 | steps =  210 | episode_length = 14 | epsilon = 0.5649736841486254\n","Episode = 13 | steps =  219 | episode_length = 9 | epsilon = 0.5621488157278822\n","Episode = 14 | steps =  229 | episode_length = 10 | epsilon = 0.5593380716492428\n","Episode = 15 | steps =  252 | episode_length = 23 | epsilon = 0.5565413812909965\n","Episode = 16 | steps =  262 | episode_length = 10 | epsilon = 0.5537586743845415\n","Episode = 17 | steps =  272 | episode_length = 10 | epsilon = 0.5509898810126188\n","Episode = 18 | steps =  284 | episode_length = 12 | epsilon = 0.5482349316075558\n","Episode = 19 | steps =  305 | episode_length = 21 | epsilon = 0.545493756949518\n","Episode = 20 | steps =  321 | episode_length = 16 | epsilon = 0.5427662881647703\n","Episode = 21 | steps =  336 | episode_length = 15 | epsilon = 0.5400524567239465\n","Episode = 22 | steps =  346 | episode_length = 10 | epsilon = 0.5373521944403268\n","Episode = 23 | steps =  355 | episode_length = 9 | epsilon = 0.5346654334681251\n","Episode = 24 | steps =  368 | episode_length = 13 | epsilon = 0.5319921063007845\n","Episode = 25 | steps =  380 | episode_length = 12 | epsilon = 0.5293321457692806\n","Episode = 26 | steps =  390 | episode_length = 10 | epsilon = 0.5266854850404341\n","Episode = 27 | steps =  404 | episode_length = 14 | epsilon = 0.5240520576152319\n","Episode = 28 | steps =  445 | episode_length = 41 | epsilon = 0.5214317973271557\n","Episode = 29 | steps =  458 | episode_length = 13 | epsilon = 0.51882463834052\n","Episode = 30 | steps =  466 | episode_length = 8 | epsilon = 0.5162305151488173\n","Episode = 31 | steps =  491 | episode_length = 25 | epsilon = 0.5136493625730733\n","Episode = 32 | steps =  510 | episode_length = 19 | epsilon = 0.5110811157602079\n","Episode = 33 | steps =  527 | episode_length = 17 | epsilon = 0.5085257101814068\n","Episode = 34 | steps =  537 | episode_length = 10 | epsilon = 0.5059830816304998\n","Episode = 35 | steps =  553 | episode_length = 16 | epsilon = 0.5034531662223473\n","Episode = 36 | steps =  579 | episode_length = 26 | epsilon = 0.5009359003912355\n","Episode = 37 | steps =  593 | episode_length = 14 | epsilon = 0.4984312208892793\n","Episode = 38 | steps =  602 | episode_length = 9 | epsilon = 0.4959390647848329\n","Episode = 39 | steps =  626 | episode_length = 24 | epsilon = 0.49345936946090874\n","Episode = 40 | steps =  646 | episode_length = 20 | epsilon = 0.49099207261360417\n","Episode = 41 | steps =  662 | episode_length = 16 | epsilon = 0.48853711225053614\n","Episode = 42 | steps =  711 | episode_length = 49 | epsilon = 0.4860944266892835\n","Episode = 43 | steps =  727 | episode_length = 16 | epsilon = 0.48366395455583705\n","Episode = 44 | steps =  753 | episode_length = 26 | epsilon = 0.4812456347830579\n","Episode = 45 | steps =  770 | episode_length = 17 | epsilon = 0.4788394066091426\n","Episode = 46 | steps =  793 | episode_length = 23 | epsilon = 0.4764452095760969\n","Episode = 47 | steps =  830 | episode_length = 37 | epsilon = 0.4740629835282164\n","Episode = 48 | steps =  850 | episode_length = 20 | epsilon = 0.47169266861057535\n","Episode = 49 | steps =  941 | episode_length = 91 | epsilon = 0.46933420526752245\n","Episode = 50 | steps =  957 | episode_length = 16 | epsilon = 0.46698753424118483\n","Episode = 51 | steps =  979 | episode_length = 22 | epsilon = 0.4646525965699789\n","Episode = 52 | steps =  1001 | episode_length = 22 | epsilon = 0.462329333587129\n","Episode = 53 | steps =  1013 | episode_length = 12 | epsilon = 0.46001768691919337\n","Episode = 54 | steps =  1037 | episode_length = 24 | epsilon = 0.4577175984845974\n","Episode = 55 | steps =  1055 | episode_length = 18 | epsilon = 0.45542901049217444\n","Episode = 56 | steps =  1064 | episode_length = 9 | epsilon = 0.4531518654397136\n","Episode = 57 | steps =  1082 | episode_length = 18 | epsilon = 0.450886106112515\n","Episode = 58 | steps =  1099 | episode_length = 17 | epsilon = 0.44863167558195244\n","Episode = 59 | steps =  1129 | episode_length = 30 | epsilon = 0.44638851720404266\n","Episode = 60 | steps =  1147 | episode_length = 18 | epsilon = 0.44415657461802244\n","Episode = 61 | steps =  1172 | episode_length = 25 | epsilon = 0.4419357917449323\n","Episode = 62 | steps =  1197 | episode_length = 25 | epsilon = 0.43972611278620766\n","Episode = 63 | steps =  1263 | episode_length = 66 | epsilon = 0.4375274822222766\n","Episode = 64 | steps =  1312 | episode_length = 49 | epsilon = 0.4353398448111652\n","Episode = 65 | steps =  1338 | episode_length = 26 | epsilon = 0.4331631455871094\n","Episode = 66 | steps =  1366 | episode_length = 28 | epsilon = 0.43099732985917383\n","Episode = 67 | steps =  1388 | episode_length = 22 | epsilon = 0.42884234320987796\n","Episode = 68 | steps =  1415 | episode_length = 27 | epsilon = 0.42669813149382857\n","Episode = 69 | steps =  1441 | episode_length = 26 | epsilon = 0.4245646408363594\n","Episode = 70 | steps =  1473 | episode_length = 32 | epsilon = 0.42244181763217764\n","Episode = 71 | steps =  1494 | episode_length = 21 | epsilon = 0.42032960854401674\n","Episode = 72 | steps =  1521 | episode_length = 27 | epsilon = 0.4182279605012967\n","Episode = 73 | steps =  1532 | episode_length = 11 | epsilon = 0.4161368206987902\n","Episode = 74 | steps =  1557 | episode_length = 25 | epsilon = 0.4140561365952963\n","Episode = 75 | steps =  1575 | episode_length = 18 | epsilon = 0.4119858559123198\n","Episode = 76 | steps =  1611 | episode_length = 36 | epsilon = 0.4099259266327582\n","Episode = 77 | steps =  1639 | episode_length = 28 | epsilon = 0.4078762969995944\n","Episode = 78 | steps =  1663 | episode_length = 24 | epsilon = 0.4058369155145964\n","Episode = 79 | steps =  1678 | episode_length = 15 | epsilon = 0.4038077309370234\n","Episode = 80 | steps =  1697 | episode_length = 19 | epsilon = 0.40178869228233827\n","Episode = 81 | steps =  1736 | episode_length = 39 | epsilon = 0.3997797488209266\n","Episode = 82 | steps =  1757 | episode_length = 21 | epsilon = 0.39778085007682196\n","Episode = 83 | steps =  1778 | episode_length = 21 | epsilon = 0.39579194582643784\n","Episode = 84 | steps =  1795 | episode_length = 17 | epsilon = 0.39381298609730564\n","Episode = 85 | steps =  1819 | episode_length = 24 | epsilon = 0.3918439211668191\n","Episode = 86 | steps =  1831 | episode_length = 12 | epsilon = 0.389884701560985\n","Episode = 87 | steps =  1841 | episode_length = 10 | epsilon = 0.38793527805318007\n","Episode = 88 | steps =  1859 | episode_length = 18 | epsilon = 0.3859956016629142\n","Episode = 89 | steps =  1889 | episode_length = 30 | epsilon = 0.3840656236545996\n","Episode = 90 | steps =  1904 | episode_length = 15 | epsilon = 0.38214529553632665\n","Episode = 91 | steps =  1923 | episode_length = 19 | epsilon = 0.380234569058645\n","Episode = 92 | steps =  1955 | episode_length = 32 | epsilon = 0.37833339621335177\n","Episode = 93 | steps =  1981 | episode_length = 26 | epsilon = 0.376441729232285\n","Episode = 94 | steps =  2009 | episode_length = 28 | epsilon = 0.3745595205861236\n","Episode = 95 | steps =  2038 | episode_length = 29 | epsilon = 0.37268672298319294\n","Episode = 96 | steps =  2057 | episode_length = 19 | epsilon = 0.37082328936827696\n","Episode = 97 | steps =  2085 | episode_length = 28 | epsilon = 0.36896917292143555\n","Episode = 98 | steps =  2117 | episode_length = 32 | epsilon = 0.36712432705682835\n","Episode = 99 | steps =  2141 | episode_length = 24 | epsilon = 0.3652887054215442\n","Episode = 100 | steps =  2157 | episode_length = 16 | epsilon = 0.3634622618944365\n","Episode = 101 | steps =  2185 | episode_length = 28 | epsilon = 0.3616449505849643\n","Episode = 102 | steps =  2200 | episode_length = 15 | epsilon = 0.3598367258320395\n","Episode = 103 | steps =  2233 | episode_length = 33 | epsilon = 0.3580375422028793\n","Episode = 104 | steps =  2261 | episode_length = 28 | epsilon = 0.3562473544918649\n","Episode = 105 | steps =  2293 | episode_length = 32 | epsilon = 0.35446611771940556\n","Episode = 106 | steps =  2311 | episode_length = 18 | epsilon = 0.3526937871308085\n","Episode = 107 | steps =  2343 | episode_length = 32 | epsilon = 0.35093031819515447\n","Episode = 108 | steps =  2367 | episode_length = 24 | epsilon = 0.3491756666041787\n","Episode = 109 | steps =  2399 | episode_length = 32 | epsilon = 0.3474297882711578\n","Episode = 110 | steps =  2427 | episode_length = 28 | epsilon = 0.345692639329802\n","Episode = 111 | steps =  2451 | episode_length = 24 | epsilon = 0.343964176133153\n","Episode = 112 | steps =  2471 | episode_length = 20 | epsilon = 0.34224435525248725\n","Episode = 113 | steps =  2500 | episode_length = 29 | epsilon = 0.3405331334762248\n","Episode = 114 | steps =  2531 | episode_length = 31 | epsilon = 0.3388304678088437\n","Episode = 115 | steps =  2556 | episode_length = 25 | epsilon = 0.33713631546979944\n","Episode = 116 | steps =  2588 | episode_length = 32 | epsilon = 0.33545063389245044\n","Episode = 117 | steps =  2599 | episode_length = 11 | epsilon = 0.3337733807229882\n","Episode = 118 | steps =  2627 | episode_length = 28 | epsilon = 0.3321045138193732\n","Episode = 119 | steps =  2651 | episode_length = 24 | epsilon = 0.33044399125027635\n","Episode = 120 | steps =  2676 | episode_length = 25 | epsilon = 0.328791771294025\n","Episode = 121 | steps =  2698 | episode_length = 22 | epsilon = 0.32714781243755486\n","Episode = 122 | steps =  2732 | episode_length = 34 | epsilon = 0.3255120733753671\n","Episode = 123 | steps =  2761 | episode_length = 29 | epsilon = 0.32388451300849025\n","Episode = 124 | steps =  2787 | episode_length = 26 | epsilon = 0.3222650904434478\n","Episode = 125 | steps =  2813 | episode_length = 26 | epsilon = 0.32065376499123055\n","Episode = 126 | steps =  2840 | episode_length = 27 | epsilon = 0.3190504961662744\n","Episode = 127 | steps =  2860 | episode_length = 20 | epsilon = 0.31745524368544303\n","Episode = 128 | steps =  2889 | episode_length = 29 | epsilon = 0.31586796746701584\n","Episode = 129 | steps =  2928 | episode_length = 39 | epsilon = 0.31428862762968074\n","Episode = 130 | steps =  2958 | episode_length = 30 | epsilon = 0.31271718449153235\n","Episode = 131 | steps =  2985 | episode_length = 27 | epsilon = 0.3111535985690747\n","Episode = 132 | steps =  3011 | episode_length = 26 | epsilon = 0.30959783057622936\n","Episode = 133 | steps =  3054 | episode_length = 43 | epsilon = 0.30804984142334824\n","Episode = 134 | steps =  3078 | episode_length = 24 | epsilon = 0.3065095922162315\n","Episode = 135 | steps =  3096 | episode_length = 18 | epsilon = 0.3049770442551503\n","Episode = 136 | steps =  3113 | episode_length = 17 | epsilon = 0.30345215903387457\n","Episode = 137 | steps =  3137 | episode_length = 24 | epsilon = 0.3019348982387052\n","Episode = 138 | steps =  3170 | episode_length = 33 | epsilon = 0.3004252237475116\n","Episode = 139 | steps =  3195 | episode_length = 25 | epsilon = 0.29892309762877406\n","Episode = 140 | steps =  3217 | episode_length = 22 | epsilon = 0.29742848214063017\n","Episode = 141 | steps =  3228 | episode_length = 11 | epsilon = 0.29594133972992703\n","Episode = 142 | steps =  3251 | episode_length = 23 | epsilon = 0.29446163303127737\n","Episode = 143 | steps =  3271 | episode_length = 20 | epsilon = 0.292989324866121\n","Episode = 144 | steps =  3296 | episode_length = 25 | epsilon = 0.2915243782417904\n","Episode = 145 | steps =  3318 | episode_length = 22 | epsilon = 0.29006675635058143\n","Episode = 146 | steps =  3328 | episode_length = 10 | epsilon = 0.2886164225688285\n","Episode = 147 | steps =  3347 | episode_length = 19 | epsilon = 0.28717334045598436\n","Episode = 148 | steps =  3371 | episode_length = 24 | epsilon = 0.28573747375370445\n","Episode = 149 | steps =  3409 | episode_length = 38 | epsilon = 0.28430878638493595\n","Episode = 150 | steps =  3433 | episode_length = 24 | epsilon = 0.2828872424530113\n","Episode = 151 | steps =  3454 | episode_length = 21 | epsilon = 0.2814728062407462\n","Episode = 152 | steps =  3482 | episode_length = 28 | epsilon = 0.2800654422095425\n","Episode = 153 | steps =  3502 | episode_length = 20 | epsilon = 0.2786651149984948\n","Episode = 154 | steps =  3530 | episode_length = 28 | epsilon = 0.2772717894235023\n","Episode = 155 | steps =  3556 | episode_length = 26 | epsilon = 0.27588543047638475\n","Episode = 156 | steps =  3577 | episode_length = 21 | epsilon = 0.2745060033240028\n","Episode = 157 | steps =  3590 | episode_length = 13 | epsilon = 0.2731334733073828\n","Episode = 158 | steps =  3624 | episode_length = 34 | epsilon = 0.2717678059408459\n","Episode = 159 | steps =  3655 | episode_length = 31 | epsilon = 0.27040896691114163\n","Episode = 160 | steps =  3682 | episode_length = 27 | epsilon = 0.26905692207658594\n","Episode = 161 | steps =  3704 | episode_length = 22 | epsilon = 0.267711637466203\n","Episode = 162 | steps =  3727 | episode_length = 23 | epsilon = 0.266373079278872\n","Episode = 163 | steps =  3753 | episode_length = 26 | epsilon = 0.26504121388247764\n","Episode = 164 | steps =  3772 | episode_length = 19 | epsilon = 0.26371600781306526\n","Episode = 165 | steps =  3802 | episode_length = 30 | epsilon = 0.26239742777399994\n","Episode = 166 | steps =  3823 | episode_length = 21 | epsilon = 0.26108544063512995\n","Episode = 167 | steps =  3841 | episode_length = 18 | epsilon = 0.2597800134319543\n","Episode = 168 | steps =  3854 | episode_length = 13 | epsilon = 0.2584811133647945\n","Episode = 169 | steps =  3866 | episode_length = 12 | epsilon = 0.25718870779797054\n","Episode = 170 | steps =  3887 | episode_length = 21 | epsilon = 0.2559027642589807\n","Episode = 171 | steps =  3910 | episode_length = 23 | epsilon = 0.25462325043768574\n","Episode = 172 | steps =  3932 | episode_length = 22 | epsilon = 0.2533501341854973\n","Episode = 173 | steps =  3948 | episode_length = 16 | epsilon = 0.2520833835145698\n","Episode = 174 | steps =  3957 | episode_length = 9 | epsilon = 0.250822966596997\n","Episode = 175 | steps =  3968 | episode_length = 11 | epsilon = 0.249568851764012\n","Episode = 176 | steps =  3990 | episode_length = 22 | epsilon = 0.24832100750519193\n","Episode = 177 | steps =  4009 | episode_length = 19 | epsilon = 0.24707940246766597\n","Episode = 178 | steps =  4028 | episode_length = 19 | epsilon = 0.24584400545532764\n","Episode = 179 | steps =  4037 | episode_length = 9 | epsilon = 0.244614785428051\n","Episode = 180 | steps =  4050 | episode_length = 13 | epsilon = 0.24339171150091077\n","Episode = 181 | steps =  4061 | episode_length = 11 | epsilon = 0.2421747529434062\n","Episode = 182 | steps =  4075 | episode_length = 14 | epsilon = 0.24096387917868917\n","Episode = 183 | steps =  4089 | episode_length = 14 | epsilon = 0.2397590597827957\n","Episode = 184 | steps =  4101 | episode_length = 12 | epsilon = 0.23856026448388173\n","Episode = 185 | steps =  4112 | episode_length = 11 | epsilon = 0.23736746316146232\n","Episode = 186 | steps =  4123 | episode_length = 11 | epsilon = 0.236180625845655\n","Episode = 187 | steps =  4134 | episode_length = 11 | epsilon = 0.23499972271642672\n","Episode = 188 | steps =  4149 | episode_length = 15 | epsilon = 0.23382472410284458\n","Episode = 189 | steps =  4162 | episode_length = 13 | epsilon = 0.23265560048233036\n","Episode = 190 | steps =  4182 | episode_length = 20 | epsilon = 0.23149232247991872\n","Episode = 191 | steps =  4193 | episode_length = 11 | epsilon = 0.23033486086751911\n","Episode = 192 | steps =  4212 | episode_length = 19 | epsilon = 0.2291831865631815\n","Episode = 193 | steps =  4224 | episode_length = 12 | epsilon = 0.2280372706303656\n","Episode = 194 | steps =  4236 | episode_length = 12 | epsilon = 0.22689708427721378\n","Episode = 195 | steps =  4246 | episode_length = 10 | epsilon = 0.2257625988558277\n","Episode = 196 | steps =  4258 | episode_length = 12 | epsilon = 0.22463378586154856\n","Episode = 197 | steps =  4274 | episode_length = 16 | epsilon = 0.2235106169322408\n","Episode = 198 | steps =  4286 | episode_length = 12 | epsilon = 0.2223930638475796\n","Episode = 199 | steps =  4300 | episode_length = 14 | epsilon = 0.22128109852834169\n","Episode = 200 | steps =  4328 | episode_length = 28 | epsilon = 0.22017469303569998\n","Episode = 201 | steps =  4343 | episode_length = 15 | epsilon = 0.21907381957052147\n","Episode = 202 | steps =  4354 | episode_length = 11 | epsilon = 0.21797845047266887\n","Episode = 203 | steps =  4364 | episode_length = 10 | epsilon = 0.21688855822030553\n","Episode = 204 | steps =  4380 | episode_length = 16 | epsilon = 0.215804115429204\n","Episode = 205 | steps =  4390 | episode_length = 10 | epsilon = 0.21472509485205798\n","Episode = 206 | steps =  4401 | episode_length = 11 | epsilon = 0.2136514693777977\n","Episode = 207 | steps =  4410 | episode_length = 9 | epsilon = 0.2125832120309087\n","Episode = 208 | steps =  4425 | episode_length = 15 | epsilon = 0.21152029597075417\n","Episode = 209 | steps =  4437 | episode_length = 12 | epsilon = 0.2104626944909004\n","Episode = 210 | steps =  4451 | episode_length = 14 | epsilon = 0.20941038101844592\n","Episode = 211 | steps =  4461 | episode_length = 10 | epsilon = 0.2083633291133537\n","Episode = 212 | steps =  4472 | episode_length = 11 | epsilon = 0.20732151246778693\n","Episode = 213 | steps =  4488 | episode_length = 16 | epsilon = 0.20628490490544799\n","Episode = 214 | steps =  4514 | episode_length = 26 | epsilon = 0.20525348038092076\n","Episode = 215 | steps =  4526 | episode_length = 12 | epsilon = 0.20422721297901614\n","Episode = 216 | steps =  4538 | episode_length = 12 | epsilon = 0.20320607691412104\n","Episode = 217 | steps =  4552 | episode_length = 14 | epsilon = 0.20219004652955044\n","Episode = 218 | steps =  4562 | episode_length = 10 | epsilon = 0.2011790962969027\n","Episode = 219 | steps =  4573 | episode_length = 11 | epsilon = 0.20017320081541817\n","Episode = 220 | steps =  4584 | episode_length = 11 | epsilon = 0.19917233481134108\n","Episode = 221 | steps =  4595 | episode_length = 11 | epsilon = 0.19817647313728437\n","Episode = 222 | steps =  4608 | episode_length = 13 | epsilon = 0.19718559077159795\n","Episode = 223 | steps =  4619 | episode_length = 11 | epsilon = 0.19619966281773996\n","Episode = 224 | steps =  4630 | episode_length = 11 | epsilon = 0.19521866450365125\n","Episode = 225 | steps =  4639 | episode_length = 9 | epsilon = 0.194242571181133\n","Episode = 226 | steps =  4648 | episode_length = 9 | epsilon = 0.19327135832522732\n","Episode = 227 | steps =  4659 | episode_length = 11 | epsilon = 0.1923050015336012\n","Episode = 228 | steps =  4669 | episode_length = 10 | epsilon = 0.1913434765259332\n","Episode = 229 | steps =  4680 | episode_length = 11 | epsilon = 0.19038675914330352\n","Episode = 230 | steps =  4693 | episode_length = 13 | epsilon = 0.18943482534758702\n","Episode = 231 | steps =  4702 | episode_length = 9 | epsilon = 0.18848765122084907\n","Episode = 232 | steps =  4712 | episode_length = 10 | epsilon = 0.18754521296474483\n","Episode = 233 | steps =  4722 | episode_length = 10 | epsilon = 0.1866074868999211\n","Episode = 234 | steps =  4732 | episode_length = 10 | epsilon = 0.1856744494654215\n","Episode = 235 | steps =  4748 | episode_length = 16 | epsilon = 0.18474607721809438\n","Episode = 236 | steps =  4774 | episode_length = 26 | epsilon = 0.1838223468320039\n","Episode = 237 | steps =  4803 | episode_length = 29 | epsilon = 0.1829032350978439\n","Episode = 238 | steps =  4815 | episode_length = 12 | epsilon = 0.1819887189223547\n","Episode = 239 | steps =  4825 | episode_length = 10 | epsilon = 0.1810787753277429\n","Episode = 240 | steps =  4837 | episode_length = 12 | epsilon = 0.1801733814511042\n","Episode = 241 | steps =  4848 | episode_length = 11 | epsilon = 0.17927251454384868\n","Episode = 242 | steps =  4873 | episode_length = 25 | epsilon = 0.17837615197112944\n","Episode = 243 | steps =  4883 | episode_length = 10 | epsilon = 0.1774842712112738\n","Episode = 244 | steps =  4910 | episode_length = 27 | epsilon = 0.17659684985521742\n","Episode = 245 | steps =  4920 | episode_length = 10 | epsilon = 0.17571386560594135\n","Episode = 246 | steps =  4953 | episode_length = 33 | epsilon = 0.17483529627791164\n","Episode = 247 | steps =  4979 | episode_length = 26 | epsilon = 0.17396111979652207\n","Episode = 248 | steps =  5005 | episode_length = 26 | epsilon = 0.17309131419753945\n","Episode = 249 | steps =  5015 | episode_length = 10 | epsilon = 0.17222585762655176\n","Episode = 250 | steps =  5043 | episode_length = 28 | epsilon = 0.171364728338419\n","Episode = 251 | steps =  5054 | episode_length = 11 | epsilon = 0.17050790469672691\n","Episode = 252 | steps =  5080 | episode_length = 26 | epsilon = 0.16965536517324328\n","Episode = 253 | steps =  5090 | episode_length = 10 | epsilon = 0.16880708834737707\n","Episode = 254 | steps =  5120 | episode_length = 30 | epsilon = 0.16796305290564018\n","Episode = 255 | steps =  5149 | episode_length = 29 | epsilon = 0.16712323764111198\n","Episode = 256 | steps =  5160 | episode_length = 11 | epsilon = 0.1662876214529064\n","Episode = 257 | steps =  5172 | episode_length = 12 | epsilon = 0.16545618334564188\n","Episode = 258 | steps =  5185 | episode_length = 13 | epsilon = 0.16462890242891368\n","Episode = 259 | steps =  5196 | episode_length = 11 | epsilon = 0.1638057579167691\n","Episode = 260 | steps =  5207 | episode_length = 11 | epsilon = 0.16298672912718526\n","Episode = 261 | steps =  5218 | episode_length = 11 | epsilon = 0.16217179548154934\n","Episode = 262 | steps =  5228 | episode_length = 10 | epsilon = 0.1613609365041416\n","Episode = 263 | steps =  5238 | episode_length = 10 | epsilon = 0.1605541318216209\n","Episode = 264 | steps =  5263 | episode_length = 25 | epsilon = 0.1597513611625128\n","Episode = 265 | steps =  5288 | episode_length = 25 | epsilon = 0.15895260435670022\n","Episode = 266 | steps =  5317 | episode_length = 29 | epsilon = 0.15815784133491673\n","Episode = 267 | steps =  5328 | episode_length = 11 | epsilon = 0.15736705212824215\n","Episode = 268 | steps =  5341 | episode_length = 13 | epsilon = 0.15658021686760093\n","Episode = 269 | steps =  5358 | episode_length = 17 | epsilon = 0.15579731578326292\n","Episode = 270 | steps =  5385 | episode_length = 27 | epsilon = 0.1550183292043466\n","Episode = 271 | steps =  5412 | episode_length = 27 | epsilon = 0.15424323755832486\n","Episode = 272 | steps =  5423 | episode_length = 11 | epsilon = 0.15347202137053323\n","Episode = 273 | steps =  5433 | episode_length = 10 | epsilon = 0.15270466126368057\n","Episode = 274 | steps =  5460 | episode_length = 27 | epsilon = 0.15194113795736217\n","Episode = 275 | steps =  5472 | episode_length = 12 | epsilon = 0.15118143226757536\n","Episode = 276 | steps =  5483 | episode_length = 11 | epsilon = 0.15042552510623747\n","Episode = 277 | steps =  5493 | episode_length = 10 | epsilon = 0.14967339748070627\n","Episode = 278 | steps =  5503 | episode_length = 10 | epsilon = 0.14892503049330275\n","Episode = 279 | steps =  5513 | episode_length = 10 | epsilon = 0.14818040534083624\n","Episode = 280 | steps =  5523 | episode_length = 10 | epsilon = 0.14743950331413205\n","Episode = 281 | steps =  5533 | episode_length = 10 | epsilon = 0.14670230579756138\n","Episode = 282 | steps =  5547 | episode_length = 14 | epsilon = 0.14596879426857357\n","Episode = 283 | steps =  5558 | episode_length = 11 | epsilon = 0.1452389502972307\n","Episode = 284 | steps =  5567 | episode_length = 9 | epsilon = 0.14451275554574453\n","Episode = 285 | steps =  5578 | episode_length = 11 | epsilon = 0.1437901917680158\n","Episode = 286 | steps =  5589 | episode_length = 11 | epsilon = 0.14307124080917571\n","Episode = 287 | steps =  5600 | episode_length = 11 | epsilon = 0.14235588460512982\n","Episode = 288 | steps =  5611 | episode_length = 11 | epsilon = 0.14164410518210419\n","Episode = 289 | steps =  5621 | episode_length = 10 | epsilon = 0.14093588465619367\n","Episode = 290 | steps =  5630 | episode_length = 9 | epsilon = 0.1402312052329127\n","Episode = 291 | steps =  5639 | episode_length = 9 | epsilon = 0.13953004920674814\n","Episode = 292 | steps =  5649 | episode_length = 10 | epsilon = 0.1388323989607144\n","Episode = 293 | steps =  5658 | episode_length = 9 | epsilon = 0.13813823696591082\n","Episode = 294 | steps =  5668 | episode_length = 10 | epsilon = 0.13744754578108126\n","Episode = 295 | steps =  5683 | episode_length = 15 | epsilon = 0.13676030805217584\n","Episode = 296 | steps =  5694 | episode_length = 11 | epsilon = 0.13607650651191497\n","Episode = 297 | steps =  5706 | episode_length = 12 | epsilon = 0.13539612397935538\n","Episode = 298 | steps =  5736 | episode_length = 30 | epsilon = 0.13471914335945862\n","Episode = 299 | steps =  5746 | episode_length = 10 | epsilon = 0.13404554764266133\n","Episode = 300 | steps =  5757 | episode_length = 11 | epsilon = 0.133375319904448\n","Episode = 301 | steps =  5769 | episode_length = 12 | epsilon = 0.13270844330492576\n","Episode = 302 | steps =  5781 | episode_length = 12 | epsilon = 0.13204490108840114\n","Episode = 303 | steps =  5792 | episode_length = 11 | epsilon = 0.13138467658295913\n","Episode = 304 | steps =  5802 | episode_length = 10 | epsilon = 0.13072775320004434\n","Episode = 305 | steps =  5813 | episode_length = 11 | epsilon = 0.13007411443404412\n","Episode = 306 | steps =  5823 | episode_length = 10 | epsilon = 0.1294237438618739\n","Episode = 307 | steps =  5834 | episode_length = 11 | epsilon = 0.12877662514256455\n","Episode = 308 | steps =  5846 | episode_length = 12 | epsilon = 0.12813274201685174\n","Episode = 309 | steps =  5858 | episode_length = 12 | epsilon = 0.12749207830676748\n","Episode = 310 | steps =  5869 | episode_length = 11 | epsilon = 0.12685461791523364\n","Episode = 311 | steps =  5883 | episode_length = 14 | epsilon = 0.12622034482565747\n","Episode = 312 | steps =  5893 | episode_length = 10 | epsilon = 0.1255892431015292\n","Episode = 313 | steps =  5916 | episode_length = 23 | epsilon = 0.12496129688602155\n","Episode = 314 | steps =  5960 | episode_length = 44 | epsilon = 0.12433649040159143\n","Episode = 315 | steps =  5972 | episode_length = 12 | epsilon = 0.12371480794958348\n","Episode = 316 | steps =  6004 | episode_length = 32 | epsilon = 0.12309623390983555\n","Episode = 317 | steps =  6015 | episode_length = 11 | epsilon = 0.12248075274028637\n","Episode = 318 | steps =  6056 | episode_length = 41 | epsilon = 0.12186834897658494\n","Episode = 319 | steps =  6081 | episode_length = 25 | epsilon = 0.12125900723170202\n","Episode = 320 | steps =  6127 | episode_length = 46 | epsilon = 0.12065271219554351\n","Episode = 321 | steps =  6178 | episode_length = 51 | epsilon = 0.12004944863456579\n","Episode = 322 | steps =  6190 | episode_length = 12 | epsilon = 0.11944920139139296\n","Episode = 323 | steps =  6221 | episode_length = 31 | epsilon = 0.118851955384436\n","Episode = 324 | steps =  6259 | episode_length = 38 | epsilon = 0.11825769560751383\n","Episode = 325 | steps =  6295 | episode_length = 36 | epsilon = 0.11766640712947626\n","Episode = 326 | steps =  6333 | episode_length = 38 | epsilon = 0.11707807509382888\n","Episode = 327 | steps =  6377 | episode_length = 44 | epsilon = 0.11649268471835973\n","Episode = 328 | steps =  6392 | episode_length = 15 | epsilon = 0.11591022129476793\n","Episode = 329 | steps =  6447 | episode_length = 55 | epsilon = 0.11533067018829408\n","Episode = 330 | steps =  6503 | episode_length = 56 | epsilon = 0.11475401683735262\n","Episode = 331 | steps =  6550 | episode_length = 47 | epsilon = 0.11418024675316585\n","Episode = 332 | steps =  6603 | episode_length = 53 | epsilon = 0.11360934551940002\n","Episode = 333 | steps =  6648 | episode_length = 45 | epsilon = 0.11304129879180302\n","Episode = 334 | steps =  6659 | episode_length = 11 | epsilon = 0.11247609229784401\n","Episode = 335 | steps =  6717 | episode_length = 58 | epsilon = 0.11191371183635479\n","Episode = 336 | steps =  6776 | episode_length = 59 | epsilon = 0.11135414327717301\n","Episode = 337 | steps =  6831 | episode_length = 55 | epsilon = 0.11079737256078714\n","Episode = 338 | steps =  6914 | episode_length = 83 | epsilon = 0.11024338569798321\n","Episode = 339 | steps =  7005 | episode_length = 91 | epsilon = 0.1096921687694933\n","Episode = 340 | steps =  7082 | episode_length = 77 | epsilon = 0.10914370792564583\n","Episode = 341 | steps =  7113 | episode_length = 31 | epsilon = 0.1085979893860176\n","Episode = 342 | steps =  7141 | episode_length = 28 | epsilon = 0.10805499943908752\n","Episode = 343 | steps =  7165 | episode_length = 24 | epsilon = 0.10751472444189207\n","Episode = 344 | steps =  7185 | episode_length = 20 | epsilon = 0.1069771508196826\n","Episode = 345 | steps =  7203 | episode_length = 18 | epsilon = 0.1064422650655842\n","Episode = 346 | steps =  7229 | episode_length = 26 | epsilon = 0.10591005374025628\n","Episode = 347 | steps =  7247 | episode_length = 18 | epsilon = 0.10538050347155499\n","Episode = 348 | steps =  7265 | episode_length = 18 | epsilon = 0.10485360095419721\n","Episode = 349 | steps =  7288 | episode_length = 23 | epsilon = 0.10432933294942623\n","Episode = 350 | steps =  7309 | episode_length = 21 | epsilon = 0.1038076862846791\n","Episode = 351 | steps =  7338 | episode_length = 29 | epsilon = 0.1032886478532557\n","Episode = 352 | steps =  7360 | episode_length = 22 | epsilon = 0.10277220461398942\n","Episode = 353 | steps =  7389 | episode_length = 29 | epsilon = 0.10225834359091947\n","Episode = 354 | steps =  7416 | episode_length = 27 | epsilon = 0.10174705187296487\n","Episode = 355 | steps =  7437 | episode_length = 21 | epsilon = 0.10123831661360004\n","Episode = 356 | steps =  7462 | episode_length = 25 | epsilon = 0.10073212503053204\n","Episode = 357 | steps =  7485 | episode_length = 23 | epsilon = 0.10022846440537939\n","Episode = 358 | steps =  7503 | episode_length = 18 | epsilon = 0.09972732208335248\n","Episode = 359 | steps =  7530 | episode_length = 27 | epsilon = 0.09972732208335248\n","Episode = 360 | steps =  7552 | episode_length = 22 | epsilon = 0.09972732208335248\n","Episode = 361 | steps =  7570 | episode_length = 18 | epsilon = 0.09972732208335248\n","Episode = 362 | steps =  7581 | episode_length = 11 | epsilon = 0.09972732208335248\n","Episode = 363 | steps =  7600 | episode_length = 19 | epsilon = 0.09972732208335248\n","Episode = 364 | steps =  7623 | episode_length = 23 | epsilon = 0.09972732208335248\n","Episode = 365 | steps =  7642 | episode_length = 19 | epsilon = 0.09972732208335248\n","Episode = 366 | steps =  7664 | episode_length = 22 | epsilon = 0.09972732208335248\n","Episode = 367 | steps =  7686 | episode_length = 22 | epsilon = 0.09972732208335248\n","Episode = 368 | steps =  7711 | episode_length = 25 | epsilon = 0.09972732208335248\n","Episode = 369 | steps =  7739 | episode_length = 28 | epsilon = 0.09972732208335248\n","Episode = 370 | steps =  7756 | episode_length = 17 | epsilon = 0.09972732208335248\n","Episode = 371 | steps =  7776 | episode_length = 20 | epsilon = 0.09972732208335248\n","Episode = 372 | steps =  7794 | episode_length = 18 | epsilon = 0.09972732208335248\n","Episode = 373 | steps =  7816 | episode_length = 22 | epsilon = 0.09972732208335248\n","Episode = 374 | steps =  7838 | episode_length = 22 | epsilon = 0.09972732208335248\n","Episode = 375 | steps =  7856 | episode_length = 18 | epsilon = 0.09972732208335248\n","Episode = 376 | steps =  7871 | episode_length = 15 | epsilon = 0.09972732208335248\n","Episode = 377 | steps =  7890 | episode_length = 19 | epsilon = 0.09972732208335248\n","Episode = 378 | steps =  7908 | episode_length = 18 | epsilon = 0.09972732208335248\n","Episode = 379 | steps =  7932 | episode_length = 24 | epsilon = 0.09972732208335248\n","Episode = 380 | steps =  7958 | episode_length = 26 | epsilon = 0.09972732208335248\n","Episode = 381 | steps =  7978 | episode_length = 20 | epsilon = 0.09972732208335248\n","Episode = 382 | steps =  8005 | episode_length = 27 | epsilon = 0.09972732208335248\n"]}]},{"cell_type":"code","source":["progress_limit_buffer, loss_history_limit_buffer, _ = run_dqn(8000, use_target_network=True, max_replay_buffer_size=500)"],"metadata":{"colab":{"base_uri":"https://localhost:8080/"},"id":"vlCoSUwlLQA7","executionInfo":{"status":"ok","timestamp":1680283324042,"user_tz":-120,"elapsed":374007,"user":{"displayName":"Michał Janik","userId":"10483929333441523386"}},"outputId":"c9d955f0-c58f-4591-ef91-04aff9cf95e4"},"execution_count":18,"outputs":[{"output_type":"stream","name":"stdout","text":["Episode = 1 | steps =  10 | episode_length = 10 | epsilon = 0.597\n","Episode = 2 | steps =  31 | episode_length = 21 | epsilon = 0.594015\n","Episode = 3 | steps =  41 | episode_length = 10 | epsilon = 0.5910449249999999\n","Episode = 4 | steps =  50 | episode_length = 9 | epsilon = 0.588089700375\n","Episode = 5 | steps =  62 | episode_length = 12 | epsilon = 0.5851492518731249\n","Episode = 6 | steps =  78 | episode_length = 16 | epsilon = 0.5822235056137594\n","Episode = 7 | steps =  91 | episode_length = 13 | epsilon = 0.5793123880856905\n","Episode = 8 | steps =  103 | episode_length = 12 | epsilon = 0.5764158261452621\n","Episode = 9 | steps =  132 | episode_length = 29 | epsilon = 0.5735337470145357\n","Episode = 10 | steps =  152 | episode_length = 20 | epsilon = 0.570666078279463\n","Episode = 11 | steps =  177 | episode_length = 25 | epsilon = 0.5678127478880657\n","Episode = 12 | steps =  196 | episode_length = 19 | epsilon = 0.5649736841486254\n","Episode = 13 | steps =  206 | episode_length = 10 | epsilon = 0.5621488157278822\n","Episode = 14 | steps =  217 | episode_length = 11 | epsilon = 0.5593380716492428\n","Episode = 15 | steps =  238 | episode_length = 21 | epsilon = 0.5565413812909965\n","Episode = 16 | steps =  247 | episode_length = 9 | epsilon = 0.5537586743845415\n","Episode = 17 | steps =  256 | episode_length = 9 | epsilon = 0.5509898810126188\n","Episode = 18 | steps =  267 | episode_length = 11 | epsilon = 0.5482349316075558\n","Episode = 19 | steps =  277 | episode_length = 10 | epsilon = 0.545493756949518\n","Episode = 20 | steps =  289 | episode_length = 12 | epsilon = 0.5427662881647703\n","Episode = 21 | steps =  299 | episode_length = 10 | epsilon = 0.5400524567239465\n","Episode = 22 | steps =  313 | episode_length = 14 | epsilon = 0.5373521944403268\n","Episode = 23 | steps =  324 | episode_length = 11 | epsilon = 0.5346654334681251\n","Episode = 24 | steps =  336 | episode_length = 12 | epsilon = 0.5319921063007845\n","Episode = 25 | steps =  353 | episode_length = 17 | epsilon = 0.5293321457692806\n","Episode = 26 | steps =  370 | episode_length = 17 | epsilon = 0.5266854850404341\n","Episode = 27 | steps =  386 | episode_length = 16 | epsilon = 0.5240520576152319\n","Episode = 28 | steps =  400 | episode_length = 14 | epsilon = 0.5214317973271557\n","Episode = 29 | steps =  412 | episode_length = 12 | epsilon = 0.51882463834052\n","Episode = 30 | steps =  423 | episode_length = 11 | epsilon = 0.5162305151488173\n","Episode = 31 | steps =  433 | episode_length = 10 | epsilon = 0.5136493625730733\n","Episode = 32 | steps =  449 | episode_length = 16 | epsilon = 0.5110811157602079\n","Episode = 33 | steps =  463 | episode_length = 14 | epsilon = 0.5085257101814068\n","Episode = 34 | steps =  472 | episode_length = 9 | epsilon = 0.5059830816304998\n","Episode = 35 | steps =  484 | episode_length = 12 | epsilon = 0.5034531662223473\n","Episode = 36 | steps =  497 | episode_length = 13 | epsilon = 0.5009359003912355\n","Episode = 37 | steps =  508 | episode_length = 11 | epsilon = 0.4984312208892793\n","Episode = 38 | steps =  518 | episode_length = 10 | epsilon = 0.4959390647848329\n","Episode = 39 | steps =  531 | episode_length = 13 | epsilon = 0.49345936946090874\n","Episode = 40 | steps =  552 | episode_length = 21 | epsilon = 0.49099207261360417\n","Episode = 41 | steps =  565 | episode_length = 13 | epsilon = 0.48853711225053614\n","Episode = 42 | steps =  576 | episode_length = 11 | epsilon = 0.4860944266892835\n","Episode = 43 | steps =  587 | episode_length = 11 | epsilon = 0.48366395455583705\n","Episode = 44 | steps =  597 | episode_length = 10 | epsilon = 0.4812456347830579\n","Episode = 45 | steps =  606 | episode_length = 9 | epsilon = 0.4788394066091426\n","Episode = 46 | steps =  620 | episode_length = 14 | epsilon = 0.4764452095760969\n","Episode = 47 | steps =  629 | episode_length = 9 | epsilon = 0.4740629835282164\n","Episode = 48 | steps =  642 | episode_length = 13 | epsilon = 0.47169266861057535\n","Episode = 49 | steps =  658 | episode_length = 16 | epsilon = 0.46933420526752245\n","Episode = 50 | steps =  670 | episode_length = 12 | epsilon = 0.46698753424118483\n","Episode = 51 | steps =  680 | episode_length = 10 | epsilon = 0.4646525965699789\n","Episode = 52 | steps =  691 | episode_length = 11 | epsilon = 0.462329333587129\n","Episode = 53 | steps =  701 | episode_length = 10 | epsilon = 0.46001768691919337\n","Episode = 54 | steps =  712 | episode_length = 11 | epsilon = 0.4577175984845974\n","Episode = 55 | steps =  723 | episode_length = 11 | epsilon = 0.45542901049217444\n","Episode = 56 | steps =  733 | episode_length = 10 | epsilon = 0.4531518654397136\n","Episode = 57 | steps =  745 | episode_length = 12 | epsilon = 0.450886106112515\n","Episode = 58 | steps =  766 | episode_length = 21 | epsilon = 0.44863167558195244\n","Episode = 59 | steps =  780 | episode_length = 14 | epsilon = 0.44638851720404266\n","Episode = 60 | steps =  805 | episode_length = 25 | epsilon = 0.44415657461802244\n","Episode = 61 | steps =  815 | episode_length = 10 | epsilon = 0.4419357917449323\n","Episode = 62 | steps =  826 | episode_length = 11 | epsilon = 0.43972611278620766\n","Episode = 63 | steps =  836 | episode_length = 10 | epsilon = 0.4375274822222766\n","Episode = 64 | steps =  845 | episode_length = 9 | epsilon = 0.4353398448111652\n","Episode = 65 | steps =  857 | episode_length = 12 | epsilon = 0.4331631455871094\n","Episode = 66 | steps =  866 | episode_length = 9 | epsilon = 0.43099732985917383\n","Episode = 67 | steps =  888 | episode_length = 22 | epsilon = 0.42884234320987796\n","Episode = 68 | steps =  922 | episode_length = 34 | epsilon = 0.42669813149382857\n","Episode = 69 | steps =  932 | episode_length = 10 | epsilon = 0.4245646408363594\n","Episode = 70 | steps =  947 | episode_length = 15 | epsilon = 0.42244181763217764\n","Episode = 71 | steps =  962 | episode_length = 15 | epsilon = 0.42032960854401674\n","Episode = 72 | steps =  974 | episode_length = 12 | epsilon = 0.4182279605012967\n","Episode = 73 | steps =  986 | episode_length = 12 | epsilon = 0.4161368206987902\n","Episode = 74 | steps =  999 | episode_length = 13 | epsilon = 0.4140561365952963\n","Episode = 75 | steps =  1013 | episode_length = 14 | epsilon = 0.4119858559123198\n","Episode = 76 | steps =  1026 | episode_length = 13 | epsilon = 0.4099259266327582\n","Episode = 77 | steps =  1051 | episode_length = 25 | epsilon = 0.4078762969995944\n","Episode = 78 | steps =  1060 | episode_length = 9 | epsilon = 0.4058369155145964\n","Episode = 79 | steps =  1069 | episode_length = 9 | epsilon = 0.4038077309370234\n","Episode = 80 | steps =  1090 | episode_length = 21 | epsilon = 0.40178869228233827\n","Episode = 81 | steps =  1116 | episode_length = 26 | epsilon = 0.3997797488209266\n","Episode = 82 | steps =  1150 | episode_length = 34 | epsilon = 0.39778085007682196\n","Episode = 83 | steps =  1173 | episode_length = 23 | epsilon = 0.39579194582643784\n","Episode = 84 | steps =  1223 | episode_length = 50 | epsilon = 0.39381298609730564\n","Episode = 85 | steps =  1262 | episode_length = 39 | epsilon = 0.3918439211668191\n","Episode = 86 | steps =  1322 | episode_length = 60 | epsilon = 0.389884701560985\n","Episode = 87 | steps =  1350 | episode_length = 28 | epsilon = 0.38793527805318007\n","Episode = 88 | steps =  1371 | episode_length = 21 | epsilon = 0.3859956016629142\n","Episode = 89 | steps =  1403 | episode_length = 32 | epsilon = 0.3840656236545996\n","Episode = 90 | steps =  1442 | episode_length = 39 | epsilon = 0.38214529553632665\n","Episode = 91 | steps =  1571 | episode_length = 129 | epsilon = 0.380234569058645\n","Episode = 92 | steps =  1602 | episode_length = 31 | epsilon = 0.37833339621335177\n","Episode = 93 | steps =  1644 | episode_length = 42 | epsilon = 0.376441729232285\n","Episode = 94 | steps =  1689 | episode_length = 45 | epsilon = 0.3745595205861236\n","Episode = 95 | steps =  1718 | episode_length = 29 | epsilon = 0.37268672298319294\n","Episode = 96 | steps =  1779 | episode_length = 61 | epsilon = 0.37082328936827696\n","Episode = 97 | steps =  1818 | episode_length = 39 | epsilon = 0.36896917292143555\n","Episode = 98 | steps =  1843 | episode_length = 25 | epsilon = 0.36712432705682835\n","Episode = 99 | steps =  1892 | episode_length = 49 | epsilon = 0.3652887054215442\n","Episode = 100 | steps =  1938 | episode_length = 46 | epsilon = 0.3634622618944365\n","Episode = 101 | steps =  1980 | episode_length = 42 | epsilon = 0.3616449505849643\n","Episode = 102 | steps =  2016 | episode_length = 36 | epsilon = 0.3598367258320395\n","Episode = 103 | steps =  2069 | episode_length = 53 | epsilon = 0.3580375422028793\n","Episode = 104 | steps =  2121 | episode_length = 52 | epsilon = 0.3562473544918649\n","Episode = 105 | steps =  2196 | episode_length = 75 | epsilon = 0.35446611771940556\n","Episode = 106 | steps =  2235 | episode_length = 39 | epsilon = 0.3526937871308085\n","Episode = 107 | steps =  2297 | episode_length = 62 | epsilon = 0.35093031819515447\n","Episode = 108 | steps =  2328 | episode_length = 31 | epsilon = 0.3491756666041787\n","Episode = 109 | steps =  2381 | episode_length = 53 | epsilon = 0.3474297882711578\n","Episode = 110 | steps =  2430 | episode_length = 49 | epsilon = 0.345692639329802\n","Episode = 111 | steps =  2476 | episode_length = 46 | epsilon = 0.343964176133153\n","Episode = 112 | steps =  2508 | episode_length = 32 | epsilon = 0.34224435525248725\n","Episode = 113 | steps =  2539 | episode_length = 31 | epsilon = 0.3405331334762248\n","Episode = 114 | steps =  2592 | episode_length = 53 | epsilon = 0.3388304678088437\n","Episode = 115 | steps =  2639 | episode_length = 47 | epsilon = 0.33713631546979944\n","Episode = 116 | steps =  2674 | episode_length = 35 | epsilon = 0.33545063389245044\n","Episode = 117 | steps =  2706 | episode_length = 32 | epsilon = 0.3337733807229882\n","Episode = 118 | steps =  2742 | episode_length = 36 | epsilon = 0.3321045138193732\n","Episode = 119 | steps =  2779 | episode_length = 37 | epsilon = 0.33044399125027635\n","Episode = 120 | steps =  2812 | episode_length = 33 | epsilon = 0.328791771294025\n","Episode = 121 | steps =  2842 | episode_length = 30 | epsilon = 0.32714781243755486\n","Episode = 122 | steps =  2880 | episode_length = 38 | epsilon = 0.3255120733753671\n","Episode = 123 | steps =  2924 | episode_length = 44 | epsilon = 0.32388451300849025\n","Episode = 124 | steps =  2960 | episode_length = 36 | epsilon = 0.3222650904434478\n","Episode = 125 | steps =  2990 | episode_length = 30 | epsilon = 0.32065376499123055\n","Episode = 126 | steps =  3021 | episode_length = 31 | epsilon = 0.3190504961662744\n","Episode = 127 | steps =  3082 | episode_length = 61 | epsilon = 0.31745524368544303\n","Episode = 128 | steps =  3113 | episode_length = 31 | epsilon = 0.31586796746701584\n","Episode = 129 | steps =  3147 | episode_length = 34 | epsilon = 0.31428862762968074\n","Episode = 130 | steps =  3201 | episode_length = 54 | epsilon = 0.31271718449153235\n","Episode = 131 | steps =  3231 | episode_length = 30 | epsilon = 0.3111535985690747\n","Episode = 132 | steps =  3272 | episode_length = 41 | epsilon = 0.30959783057622936\n","Episode = 133 | steps =  3316 | episode_length = 44 | epsilon = 0.30804984142334824\n","Episode = 134 | steps =  3364 | episode_length = 48 | epsilon = 0.3065095922162315\n","Episode = 135 | steps =  3404 | episode_length = 40 | epsilon = 0.3049770442551503\n","Episode = 136 | steps =  3449 | episode_length = 45 | epsilon = 0.30345215903387457\n","Episode = 137 | steps =  3479 | episode_length = 30 | epsilon = 0.3019348982387052\n","Episode = 138 | steps =  3514 | episode_length = 35 | epsilon = 0.3004252237475116\n","Episode = 139 | steps =  3526 | episode_length = 12 | epsilon = 0.29892309762877406\n","Episode = 140 | steps =  3572 | episode_length = 46 | epsilon = 0.29742848214063017\n","Episode = 141 | steps =  3622 | episode_length = 50 | epsilon = 0.29594133972992703\n","Episode = 142 | steps =  3676 | episode_length = 54 | epsilon = 0.29446163303127737\n","Episode = 143 | steps =  3724 | episode_length = 48 | epsilon = 0.292989324866121\n","Episode = 144 | steps =  3757 | episode_length = 33 | epsilon = 0.2915243782417904\n","Episode = 145 | steps =  3794 | episode_length = 37 | epsilon = 0.29006675635058143\n","Episode = 146 | steps =  3836 | episode_length = 42 | epsilon = 0.2886164225688285\n","Episode = 147 | steps =  3857 | episode_length = 21 | epsilon = 0.28717334045598436\n","Episode = 148 | steps =  3891 | episode_length = 34 | epsilon = 0.28573747375370445\n","Episode = 149 | steps =  3932 | episode_length = 41 | epsilon = 0.28430878638493595\n","Episode = 150 | steps =  3955 | episode_length = 23 | epsilon = 0.2828872424530113\n","Episode = 151 | steps =  3997 | episode_length = 42 | epsilon = 0.2814728062407462\n","Episode = 152 | steps =  4029 | episode_length = 32 | epsilon = 0.2800654422095425\n","Episode = 153 | steps =  4059 | episode_length = 30 | epsilon = 0.2786651149984948\n","Episode = 154 | steps =  4090 | episode_length = 31 | epsilon = 0.2772717894235023\n","Episode = 155 | steps =  4146 | episode_length = 56 | epsilon = 0.27588543047638475\n","Episode = 156 | steps =  4194 | episode_length = 48 | epsilon = 0.2745060033240028\n","Episode = 157 | steps =  4222 | episode_length = 28 | epsilon = 0.2731334733073828\n","Episode = 158 | steps =  4249 | episode_length = 27 | epsilon = 0.2717678059408459\n","Episode = 159 | steps =  4280 | episode_length = 31 | epsilon = 0.27040896691114163\n","Episode = 160 | steps =  4351 | episode_length = 71 | epsilon = 0.26905692207658594\n","Episode = 161 | steps =  4394 | episode_length = 43 | epsilon = 0.267711637466203\n","Episode = 162 | steps =  4442 | episode_length = 48 | epsilon = 0.266373079278872\n","Episode = 163 | steps =  4532 | episode_length = 90 | epsilon = 0.26504121388247764\n","Episode = 164 | steps =  4622 | episode_length = 90 | epsilon = 0.26371600781306526\n","Episode = 165 | steps =  4651 | episode_length = 29 | epsilon = 0.26239742777399994\n","Episode = 166 | steps =  4690 | episode_length = 39 | epsilon = 0.26108544063512995\n","Episode = 167 | steps =  4736 | episode_length = 46 | epsilon = 0.2597800134319543\n","Episode = 168 | steps =  4780 | episode_length = 44 | epsilon = 0.2584811133647945\n","Episode = 169 | steps =  4827 | episode_length = 47 | epsilon = 0.25718870779797054\n","Episode = 170 | steps =  4859 | episode_length = 32 | epsilon = 0.2559027642589807\n","Episode = 171 | steps =  4911 | episode_length = 52 | epsilon = 0.25462325043768574\n","Episode = 172 | steps =  4936 | episode_length = 25 | epsilon = 0.2533501341854973\n","Episode = 173 | steps =  4972 | episode_length = 36 | epsilon = 0.2520833835145698\n","Episode = 174 | steps =  5013 | episode_length = 41 | epsilon = 0.250822966596997\n","Episode = 175 | steps =  5048 | episode_length = 35 | epsilon = 0.249568851764012\n","Episode = 176 | steps =  5095 | episode_length = 47 | epsilon = 0.24832100750519193\n","Episode = 177 | steps =  5118 | episode_length = 23 | epsilon = 0.24707940246766597\n","Episode = 178 | steps =  5143 | episode_length = 25 | epsilon = 0.24584400545532764\n","Episode = 179 | steps =  5193 | episode_length = 50 | epsilon = 0.244614785428051\n","Episode = 180 | steps =  5245 | episode_length = 52 | epsilon = 0.24339171150091077\n","Episode = 181 | steps =  5278 | episode_length = 33 | epsilon = 0.2421747529434062\n","Episode = 182 | steps =  5305 | episode_length = 27 | epsilon = 0.24096387917868917\n","Episode = 183 | steps =  5332 | episode_length = 27 | epsilon = 0.2397590597827957\n","Episode = 184 | steps =  5364 | episode_length = 32 | epsilon = 0.23856026448388173\n","Episode = 185 | steps =  5406 | episode_length = 42 | epsilon = 0.23736746316146232\n","Episode = 186 | steps =  5427 | episode_length = 21 | epsilon = 0.236180625845655\n","Episode = 187 | steps =  5456 | episode_length = 29 | epsilon = 0.23499972271642672\n","Episode = 188 | steps =  5475 | episode_length = 19 | epsilon = 0.23382472410284458\n","Episode = 189 | steps =  5518 | episode_length = 43 | epsilon = 0.23265560048233036\n","Episode = 190 | steps =  5541 | episode_length = 23 | epsilon = 0.23149232247991872\n","Episode = 191 | steps =  5568 | episode_length = 27 | epsilon = 0.23033486086751911\n","Episode = 192 | steps =  5607 | episode_length = 39 | epsilon = 0.2291831865631815\n","Episode = 193 | steps =  5643 | episode_length = 36 | epsilon = 0.2280372706303656\n","Episode = 194 | steps =  5676 | episode_length = 33 | epsilon = 0.22689708427721378\n","Episode = 195 | steps =  5698 | episode_length = 22 | epsilon = 0.2257625988558277\n","Episode = 196 | steps =  5726 | episode_length = 28 | epsilon = 0.22463378586154856\n","Episode = 197 | steps =  5748 | episode_length = 22 | epsilon = 0.2235106169322408\n","Episode = 198 | steps =  5771 | episode_length = 23 | epsilon = 0.2223930638475796\n","Episode = 199 | steps =  5801 | episode_length = 30 | epsilon = 0.22128109852834169\n","Episode = 200 | steps =  5826 | episode_length = 25 | epsilon = 0.22017469303569998\n","Episode = 201 | steps =  5859 | episode_length = 33 | epsilon = 0.21907381957052147\n","Episode = 202 | steps =  5880 | episode_length = 21 | epsilon = 0.21797845047266887\n","Episode = 203 | steps =  5918 | episode_length = 38 | epsilon = 0.21688855822030553\n","Episode = 204 | steps =  5949 | episode_length = 31 | epsilon = 0.215804115429204\n","Episode = 205 | steps =  5991 | episode_length = 42 | epsilon = 0.21472509485205798\n","Episode = 206 | steps =  6015 | episode_length = 24 | epsilon = 0.2136514693777977\n","Episode = 207 | steps =  6051 | episode_length = 36 | epsilon = 0.2125832120309087\n","Episode = 208 | steps =  6070 | episode_length = 19 | epsilon = 0.21152029597075417\n","Episode = 209 | steps =  6110 | episode_length = 40 | epsilon = 0.2104626944909004\n","Episode = 210 | steps =  6137 | episode_length = 27 | epsilon = 0.20941038101844592\n","Episode = 211 | steps =  6172 | episode_length = 35 | epsilon = 0.2083633291133537\n","Episode = 212 | steps =  6196 | episode_length = 24 | epsilon = 0.20732151246778693\n","Episode = 213 | steps =  6227 | episode_length = 31 | epsilon = 0.20628490490544799\n","Episode = 214 | steps =  6254 | episode_length = 27 | epsilon = 0.20525348038092076\n","Episode = 215 | steps =  6275 | episode_length = 21 | epsilon = 0.20422721297901614\n","Episode = 216 | steps =  6297 | episode_length = 22 | epsilon = 0.20320607691412104\n","Episode = 217 | steps =  6321 | episode_length = 24 | epsilon = 0.20219004652955044\n","Episode = 218 | steps =  6338 | episode_length = 17 | epsilon = 0.2011790962969027\n","Episode = 219 | steps =  6361 | episode_length = 23 | epsilon = 0.20017320081541817\n","Episode = 220 | steps =  6385 | episode_length = 24 | epsilon = 0.19917233481134108\n","Episode = 221 | steps =  6400 | episode_length = 15 | epsilon = 0.19817647313728437\n","Episode = 222 | steps =  6426 | episode_length = 26 | epsilon = 0.19718559077159795\n","Episode = 223 | steps =  6454 | episode_length = 28 | epsilon = 0.19619966281773996\n","Episode = 224 | steps =  6473 | episode_length = 19 | epsilon = 0.19521866450365125\n","Episode = 225 | steps =  6505 | episode_length = 32 | epsilon = 0.194242571181133\n","Episode = 226 | steps =  6534 | episode_length = 29 | epsilon = 0.19327135832522732\n","Episode = 227 | steps =  6568 | episode_length = 34 | epsilon = 0.1923050015336012\n","Episode = 228 | steps =  6600 | episode_length = 32 | epsilon = 0.1913434765259332\n","Episode = 229 | steps =  6625 | episode_length = 25 | epsilon = 0.19038675914330352\n","Episode = 230 | steps =  6652 | episode_length = 27 | epsilon = 0.18943482534758702\n","Episode = 231 | steps =  6689 | episode_length = 37 | epsilon = 0.18848765122084907\n","Episode = 232 | steps =  6738 | episode_length = 49 | epsilon = 0.18754521296474483\n","Episode = 233 | steps =  6768 | episode_length = 30 | epsilon = 0.1866074868999211\n","Episode = 234 | steps =  6811 | episode_length = 43 | epsilon = 0.1856744494654215\n","Episode = 235 | steps =  6865 | episode_length = 54 | epsilon = 0.18474607721809438\n","Episode = 236 | steps =  6917 | episode_length = 52 | epsilon = 0.1838223468320039\n","Episode = 237 | steps =  6953 | episode_length = 36 | epsilon = 0.1829032350978439\n","Episode = 238 | steps =  7003 | episode_length = 50 | epsilon = 0.1819887189223547\n","Episode = 239 | steps =  7038 | episode_length = 35 | epsilon = 0.1810787753277429\n","Episode = 240 | steps =  7080 | episode_length = 42 | epsilon = 0.1801733814511042\n","Episode = 241 | steps =  7115 | episode_length = 35 | epsilon = 0.17927251454384868\n","Episode = 242 | steps =  7148 | episode_length = 33 | epsilon = 0.17837615197112944\n","Episode = 243 | steps =  7185 | episode_length = 37 | epsilon = 0.1774842712112738\n","Episode = 244 | steps =  7219 | episode_length = 34 | epsilon = 0.17659684985521742\n","Episode = 245 | steps =  7269 | episode_length = 50 | epsilon = 0.17571386560594135\n","Episode = 246 | steps =  7319 | episode_length = 50 | epsilon = 0.17483529627791164\n","Episode = 247 | steps =  7367 | episode_length = 48 | epsilon = 0.17396111979652207\n","Episode = 248 | steps =  7406 | episode_length = 39 | epsilon = 0.17309131419753945\n","Episode = 249 | steps =  7479 | episode_length = 73 | epsilon = 0.17222585762655176\n","Episode = 250 | steps =  7510 | episode_length = 31 | epsilon = 0.171364728338419\n","Episode = 251 | steps =  7564 | episode_length = 54 | epsilon = 0.17050790469672691\n","Episode = 252 | steps =  7631 | episode_length = 67 | epsilon = 0.16965536517324328\n","Episode = 253 | steps =  7701 | episode_length = 70 | epsilon = 0.16880708834737707\n","Episode = 254 | steps =  7767 | episode_length = 66 | epsilon = 0.16796305290564018\n","Episode = 255 | steps =  7817 | episode_length = 50 | epsilon = 0.16712323764111198\n","Episode = 256 | steps =  7900 | episode_length = 83 | epsilon = 0.1662876214529064\n","Episode = 257 | steps =  8016 | episode_length = 116 | epsilon = 0.16545618334564188\n"]}]},{"cell_type":"code","source":["progress_no_target_limit_buffer, loss_history_no_target_limit_buffer, _ = run_dqn(8000, use_target_network=False, max_replay_buffer_size=500)\n"],"metadata":{"colab":{"base_uri":"https://localhost:8080/"},"id":"sbhTRB4iLR4D","executionInfo":{"status":"ok","timestamp":1680284123124,"user_tz":-120,"elapsed":378217,"user":{"displayName":"Michał Janik","userId":"10483929333441523386"}},"outputId":"fd270342-a58a-4f36-b99b-0d509d21414b"},"execution_count":21,"outputs":[{"output_type":"stream","name":"stdout","text":["Episode = 1 | steps =  9 | episode_length = 9 | epsilon = 0.597\n","Episode = 2 | steps =  21 | episode_length = 12 | epsilon = 0.594015\n","Episode = 3 | steps =  41 | episode_length = 20 | epsilon = 0.5910449249999999\n","Episode = 4 | steps =  68 | episode_length = 27 | epsilon = 0.588089700375\n","Episode = 5 | steps =  82 | episode_length = 14 | epsilon = 0.5851492518731249\n","Episode = 6 | steps =  96 | episode_length = 14 | epsilon = 0.5822235056137594\n","Episode = 7 | steps =  121 | episode_length = 25 | epsilon = 0.5793123880856905\n","Episode = 8 | steps =  155 | episode_length = 34 | epsilon = 0.5764158261452621\n","Episode = 9 | steps =  170 | episode_length = 15 | epsilon = 0.5735337470145357\n","Episode = 10 | steps =  186 | episode_length = 16 | epsilon = 0.570666078279463\n","Episode = 11 | steps =  204 | episode_length = 18 | epsilon = 0.5678127478880657\n","Episode = 12 | steps =  226 | episode_length = 22 | epsilon = 0.5649736841486254\n","Episode = 13 | steps =  241 | episode_length = 15 | epsilon = 0.5621488157278822\n","Episode = 14 | steps =  255 | episode_length = 14 | epsilon = 0.5593380716492428\n","Episode = 15 | steps =  267 | episode_length = 12 | epsilon = 0.5565413812909965\n","Episode = 16 | steps =  276 | episode_length = 9 | epsilon = 0.5537586743845415\n","Episode = 17 | steps =  287 | episode_length = 11 | epsilon = 0.5509898810126188\n","Episode = 18 | steps =  303 | episode_length = 16 | epsilon = 0.5482349316075558\n","Episode = 19 | steps =  312 | episode_length = 9 | epsilon = 0.545493756949518\n","Episode = 20 | steps =  324 | episode_length = 12 | epsilon = 0.5427662881647703\n","Episode = 21 | steps =  334 | episode_length = 10 | epsilon = 0.5400524567239465\n","Episode = 22 | steps =  347 | episode_length = 13 | epsilon = 0.5373521944403268\n","Episode = 23 | steps =  362 | episode_length = 15 | epsilon = 0.5346654334681251\n","Episode = 24 | steps =  390 | episode_length = 28 | epsilon = 0.5319921063007845\n","Episode = 25 | steps =  408 | episode_length = 18 | epsilon = 0.5293321457692806\n","Episode = 26 | steps =  421 | episode_length = 13 | epsilon = 0.5266854850404341\n","Episode = 27 | steps =  445 | episode_length = 24 | epsilon = 0.5240520576152319\n","Episode = 28 | steps =  462 | episode_length = 17 | epsilon = 0.5214317973271557\n","Episode = 29 | steps =  474 | episode_length = 12 | epsilon = 0.51882463834052\n","Episode = 30 | steps =  500 | episode_length = 26 | epsilon = 0.5162305151488173\n","Episode = 31 | steps =  511 | episode_length = 11 | epsilon = 0.5136493625730733\n","Episode = 32 | steps =  527 | episode_length = 16 | epsilon = 0.5110811157602079\n","Episode = 33 | steps =  544 | episode_length = 17 | epsilon = 0.5085257101814068\n","Episode = 34 | steps =  555 | episode_length = 11 | epsilon = 0.5059830816304998\n","Episode = 35 | steps =  567 | episode_length = 12 | epsilon = 0.5034531662223473\n","Episode = 36 | steps =  575 | episode_length = 8 | epsilon = 0.5009359003912355\n","Episode = 37 | steps =  590 | episode_length = 15 | epsilon = 0.4984312208892793\n","Episode = 38 | steps =  602 | episode_length = 12 | epsilon = 0.4959390647848329\n","Episode = 39 | steps =  612 | episode_length = 10 | epsilon = 0.49345936946090874\n","Episode = 40 | steps =  623 | episode_length = 11 | epsilon = 0.49099207261360417\n","Episode = 41 | steps =  634 | episode_length = 11 | epsilon = 0.48853711225053614\n","Episode = 42 | steps =  645 | episode_length = 11 | epsilon = 0.4860944266892835\n","Episode = 43 | steps =  659 | episode_length = 14 | epsilon = 0.48366395455583705\n","Episode = 44 | steps =  679 | episode_length = 20 | epsilon = 0.4812456347830579\n","Episode = 45 | steps =  689 | episode_length = 10 | epsilon = 0.4788394066091426\n","Episode = 46 | steps =  700 | episode_length = 11 | epsilon = 0.4764452095760969\n","Episode = 47 | steps =  713 | episode_length = 13 | epsilon = 0.4740629835282164\n","Episode = 48 | steps =  722 | episode_length = 9 | epsilon = 0.47169266861057535\n","Episode = 49 | steps =  731 | episode_length = 9 | epsilon = 0.46933420526752245\n","Episode = 50 | steps =  740 | episode_length = 9 | epsilon = 0.46698753424118483\n","Episode = 51 | steps =  750 | episode_length = 10 | epsilon = 0.4646525965699789\n","Episode = 52 | steps =  759 | episode_length = 9 | epsilon = 0.462329333587129\n","Episode = 53 | steps =  774 | episode_length = 15 | epsilon = 0.46001768691919337\n","Episode = 54 | steps =  785 | episode_length = 11 | epsilon = 0.4577175984845974\n","Episode = 55 | steps =  803 | episode_length = 18 | epsilon = 0.45542901049217444\n","Episode = 56 | steps =  816 | episode_length = 13 | epsilon = 0.4531518654397136\n","Episode = 57 | steps =  829 | episode_length = 13 | epsilon = 0.450886106112515\n","Episode = 58 | steps =  848 | episode_length = 19 | epsilon = 0.44863167558195244\n","Episode = 59 | steps =  864 | episode_length = 16 | epsilon = 0.44638851720404266\n","Episode = 60 | steps =  876 | episode_length = 12 | epsilon = 0.44415657461802244\n","Episode = 61 | steps =  888 | episode_length = 12 | epsilon = 0.4419357917449323\n","Episode = 62 | steps =  898 | episode_length = 10 | epsilon = 0.43972611278620766\n","Episode = 63 | steps =  908 | episode_length = 10 | epsilon = 0.4375274822222766\n","Episode = 64 | steps =  925 | episode_length = 17 | epsilon = 0.4353398448111652\n","Episode = 65 | steps =  934 | episode_length = 9 | epsilon = 0.4331631455871094\n","Episode = 66 | steps =  943 | episode_length = 9 | epsilon = 0.43099732985917383\n","Episode = 67 | steps =  956 | episode_length = 13 | epsilon = 0.42884234320987796\n","Episode = 68 | steps =  970 | episode_length = 14 | epsilon = 0.42669813149382857\n","Episode = 69 | steps =  981 | episode_length = 11 | epsilon = 0.4245646408363594\n","Episode = 70 | steps =  998 | episode_length = 17 | epsilon = 0.42244181763217764\n","Episode = 71 | steps =  1008 | episode_length = 10 | epsilon = 0.42032960854401674\n","Episode = 72 | steps =  1023 | episode_length = 15 | epsilon = 0.4182279605012967\n","Episode = 73 | steps =  1048 | episode_length = 25 | epsilon = 0.4161368206987902\n","Episode = 74 | steps =  1060 | episode_length = 12 | epsilon = 0.4140561365952963\n","Episode = 75 | steps =  1072 | episode_length = 12 | epsilon = 0.4119858559123198\n","Episode = 76 | steps =  1081 | episode_length = 9 | epsilon = 0.4099259266327582\n","Episode = 77 | steps =  1100 | episode_length = 19 | epsilon = 0.4078762969995944\n","Episode = 78 | steps =  1112 | episode_length = 12 | epsilon = 0.4058369155145964\n","Episode = 79 | steps =  1142 | episode_length = 30 | epsilon = 0.4038077309370234\n","Episode = 80 | steps =  1173 | episode_length = 31 | epsilon = 0.40178869228233827\n","Episode = 81 | steps =  1188 | episode_length = 15 | epsilon = 0.3997797488209266\n","Episode = 82 | steps =  1234 | episode_length = 46 | epsilon = 0.39778085007682196\n","Episode = 83 | steps =  1263 | episode_length = 29 | epsilon = 0.39579194582643784\n","Episode = 84 | steps =  1298 | episode_length = 35 | epsilon = 0.39381298609730564\n","Episode = 85 | steps =  1311 | episode_length = 13 | epsilon = 0.3918439211668191\n","Episode = 86 | steps =  1328 | episode_length = 17 | epsilon = 0.389884701560985\n","Episode = 87 | steps =  1384 | episode_length = 56 | epsilon = 0.38793527805318007\n","Episode = 88 | steps =  1412 | episode_length = 28 | epsilon = 0.3859956016629142\n","Episode = 89 | steps =  1429 | episode_length = 17 | epsilon = 0.3840656236545996\n","Episode = 90 | steps =  1439 | episode_length = 10 | epsilon = 0.38214529553632665\n","Episode = 91 | steps =  1458 | episode_length = 19 | epsilon = 0.380234569058645\n","Episode = 92 | steps =  1487 | episode_length = 29 | epsilon = 0.37833339621335177\n","Episode = 93 | steps =  1508 | episode_length = 21 | epsilon = 0.376441729232285\n","Episode = 94 | steps =  1522 | episode_length = 14 | epsilon = 0.3745595205861236\n","Episode = 95 | steps =  1554 | episode_length = 32 | epsilon = 0.37268672298319294\n","Episode = 96 | steps =  1566 | episode_length = 12 | epsilon = 0.37082328936827696\n","Episode = 97 | steps =  1602 | episode_length = 36 | epsilon = 0.36896917292143555\n","Episode = 98 | steps =  1618 | episode_length = 16 | epsilon = 0.36712432705682835\n","Episode = 99 | steps =  1648 | episode_length = 30 | epsilon = 0.3652887054215442\n","Episode = 100 | steps =  1660 | episode_length = 12 | epsilon = 0.3634622618944365\n","Episode = 101 | steps =  1698 | episode_length = 38 | epsilon = 0.3616449505849643\n","Episode = 102 | steps =  1748 | episode_length = 50 | epsilon = 0.3598367258320395\n","Episode = 103 | steps =  1800 | episode_length = 52 | epsilon = 0.3580375422028793\n","Episode = 104 | steps =  1850 | episode_length = 50 | epsilon = 0.3562473544918649\n","Episode = 105 | steps =  1878 | episode_length = 28 | epsilon = 0.35446611771940556\n","Episode = 106 | steps =  1925 | episode_length = 47 | epsilon = 0.3526937871308085\n","Episode = 107 | steps =  1956 | episode_length = 31 | epsilon = 0.35093031819515447\n","Episode = 108 | steps =  2048 | episode_length = 92 | epsilon = 0.3491756666041787\n","Episode = 109 | steps =  2173 | episode_length = 125 | epsilon = 0.3474297882711578\n","Episode = 110 | steps =  2227 | episode_length = 54 | epsilon = 0.345692639329802\n","Episode = 111 | steps =  2317 | episode_length = 90 | epsilon = 0.343964176133153\n","Episode = 112 | steps =  2380 | episode_length = 63 | epsilon = 0.34224435525248725\n","Episode = 113 | steps =  2390 | episode_length = 10 | epsilon = 0.3405331334762248\n","Episode = 114 | steps =  2400 | episode_length = 10 | epsilon = 0.3388304678088437\n","Episode = 115 | steps =  2411 | episode_length = 11 | epsilon = 0.33713631546979944\n","Episode = 116 | steps =  2421 | episode_length = 10 | epsilon = 0.33545063389245044\n","Episode = 117 | steps =  2434 | episode_length = 13 | epsilon = 0.3337733807229882\n","Episode = 118 | steps =  2457 | episode_length = 23 | epsilon = 0.3321045138193732\n","Episode = 119 | steps =  2474 | episode_length = 17 | epsilon = 0.33044399125027635\n","Episode = 120 | steps =  2482 | episode_length = 8 | epsilon = 0.328791771294025\n","Episode = 121 | steps =  2490 | episode_length = 8 | epsilon = 0.32714781243755486\n","Episode = 122 | steps =  2500 | episode_length = 10 | epsilon = 0.3255120733753671\n","Episode = 123 | steps =  2511 | episode_length = 11 | epsilon = 0.32388451300849025\n","Episode = 124 | steps =  2522 | episode_length = 11 | epsilon = 0.3222650904434478\n","Episode = 125 | steps =  2532 | episode_length = 10 | epsilon = 0.32065376499123055\n","Episode = 126 | steps =  2541 | episode_length = 9 | epsilon = 0.3190504961662744\n","Episode = 127 | steps =  2563 | episode_length = 22 | epsilon = 0.31745524368544303\n","Episode = 128 | steps =  2575 | episode_length = 12 | epsilon = 0.31586796746701584\n","Episode = 129 | steps =  2586 | episode_length = 11 | epsilon = 0.31428862762968074\n","Episode = 130 | steps =  2599 | episode_length = 13 | epsilon = 0.31271718449153235\n","Episode = 131 | steps =  2609 | episode_length = 10 | epsilon = 0.3111535985690747\n","Episode = 132 | steps =  2621 | episode_length = 12 | epsilon = 0.30959783057622936\n","Episode = 133 | steps =  2631 | episode_length = 10 | epsilon = 0.30804984142334824\n","Episode = 134 | steps =  2641 | episode_length = 10 | epsilon = 0.3065095922162315\n","Episode = 135 | steps =  2655 | episode_length = 14 | epsilon = 0.3049770442551503\n","Episode = 136 | steps =  2665 | episode_length = 10 | epsilon = 0.30345215903387457\n","Episode = 137 | steps =  2680 | episode_length = 15 | epsilon = 0.3019348982387052\n","Episode = 138 | steps =  2694 | episode_length = 14 | epsilon = 0.3004252237475116\n","Episode = 139 | steps =  2704 | episode_length = 10 | epsilon = 0.29892309762877406\n","Episode = 140 | steps =  2719 | episode_length = 15 | epsilon = 0.29742848214063017\n","Episode = 141 | steps =  2729 | episode_length = 10 | epsilon = 0.29594133972992703\n","Episode = 142 | steps =  2739 | episode_length = 10 | epsilon = 0.29446163303127737\n","Episode = 143 | steps =  2750 | episode_length = 11 | epsilon = 0.292989324866121\n","Episode = 144 | steps =  2760 | episode_length = 10 | epsilon = 0.2915243782417904\n","Episode = 145 | steps =  2771 | episode_length = 11 | epsilon = 0.29006675635058143\n","Episode = 146 | steps =  2779 | episode_length = 8 | epsilon = 0.2886164225688285\n","Episode = 147 | steps =  2791 | episode_length = 12 | epsilon = 0.28717334045598436\n","Episode = 148 | steps =  2803 | episode_length = 12 | epsilon = 0.28573747375370445\n","Episode = 149 | steps =  2813 | episode_length = 10 | epsilon = 0.28430878638493595\n","Episode = 150 | steps =  2824 | episode_length = 11 | epsilon = 0.2828872424530113\n","Episode = 151 | steps =  2836 | episode_length = 12 | epsilon = 0.2814728062407462\n","Episode = 152 | steps =  2847 | episode_length = 11 | epsilon = 0.2800654422095425\n","Episode = 153 | steps =  2857 | episode_length = 10 | epsilon = 0.2786651149984948\n","Episode = 154 | steps =  2866 | episode_length = 9 | epsilon = 0.2772717894235023\n","Episode = 155 | steps =  2885 | episode_length = 19 | epsilon = 0.27588543047638475\n","Episode = 156 | steps =  2897 | episode_length = 12 | epsilon = 0.2745060033240028\n","Episode = 157 | steps =  2908 | episode_length = 11 | epsilon = 0.2731334733073828\n","Episode = 158 | steps =  2922 | episode_length = 14 | epsilon = 0.2717678059408459\n","Episode = 159 | steps =  2933 | episode_length = 11 | epsilon = 0.27040896691114163\n","Episode = 160 | steps =  2942 | episode_length = 9 | epsilon = 0.26905692207658594\n","Episode = 161 | steps =  2955 | episode_length = 13 | epsilon = 0.267711637466203\n","Episode = 162 | steps =  2965 | episode_length = 10 | epsilon = 0.266373079278872\n","Episode = 163 | steps =  2976 | episode_length = 11 | epsilon = 0.26504121388247764\n","Episode = 164 | steps =  2988 | episode_length = 12 | epsilon = 0.26371600781306526\n","Episode = 165 | steps =  2999 | episode_length = 11 | epsilon = 0.26239742777399994\n","Episode = 166 | steps =  3009 | episode_length = 10 | epsilon = 0.26108544063512995\n","Episode = 167 | steps =  3019 | episode_length = 10 | epsilon = 0.2597800134319543\n","Episode = 168 | steps =  3030 | episode_length = 11 | epsilon = 0.2584811133647945\n","Episode = 169 | steps =  3043 | episode_length = 13 | epsilon = 0.25718870779797054\n","Episode = 170 | steps =  3053 | episode_length = 10 | epsilon = 0.2559027642589807\n","Episode = 171 | steps =  3066 | episode_length = 13 | epsilon = 0.25462325043768574\n","Episode = 172 | steps =  3079 | episode_length = 13 | epsilon = 0.2533501341854973\n","Episode = 173 | steps =  3096 | episode_length = 17 | epsilon = 0.2520833835145698\n","Episode = 174 | steps =  3107 | episode_length = 11 | epsilon = 0.250822966596997\n","Episode = 175 | steps =  3126 | episode_length = 19 | epsilon = 0.249568851764012\n","Episode = 176 | steps =  3147 | episode_length = 21 | epsilon = 0.24832100750519193\n","Episode = 177 | steps =  3158 | episode_length = 11 | epsilon = 0.24707940246766597\n","Episode = 178 | steps =  3166 | episode_length = 8 | epsilon = 0.24584400545532764\n","Episode = 179 | steps =  3178 | episode_length = 12 | epsilon = 0.244614785428051\n","Episode = 180 | steps =  3189 | episode_length = 11 | epsilon = 0.24339171150091077\n","Episode = 181 | steps =  3201 | episode_length = 12 | epsilon = 0.2421747529434062\n","Episode = 182 | steps =  3211 | episode_length = 10 | epsilon = 0.24096387917868917\n","Episode = 183 | steps =  3220 | episode_length = 9 | epsilon = 0.2397590597827957\n","Episode = 184 | steps =  3229 | episode_length = 9 | epsilon = 0.23856026448388173\n","Episode = 185 | steps =  3240 | episode_length = 11 | epsilon = 0.23736746316146232\n","Episode = 186 | steps =  3250 | episode_length = 10 | epsilon = 0.236180625845655\n","Episode = 187 | steps =  3261 | episode_length = 11 | epsilon = 0.23499972271642672\n","Episode = 188 | steps =  3272 | episode_length = 11 | epsilon = 0.23382472410284458\n","Episode = 189 | steps =  3283 | episode_length = 11 | epsilon = 0.23265560048233036\n","Episode = 190 | steps =  3295 | episode_length = 12 | epsilon = 0.23149232247991872\n","Episode = 191 | steps =  3304 | episode_length = 9 | epsilon = 0.23033486086751911\n","Episode = 192 | steps =  3319 | episode_length = 15 | epsilon = 0.2291831865631815\n","Episode = 193 | steps =  3331 | episode_length = 12 | epsilon = 0.2280372706303656\n","Episode = 194 | steps =  3340 | episode_length = 9 | epsilon = 0.22689708427721378\n","Episode = 195 | steps =  3351 | episode_length = 11 | epsilon = 0.2257625988558277\n","Episode = 196 | steps =  3363 | episode_length = 12 | epsilon = 0.22463378586154856\n","Episode = 197 | steps =  3371 | episode_length = 8 | epsilon = 0.2235106169322408\n","Episode = 198 | steps =  3381 | episode_length = 10 | epsilon = 0.2223930638475796\n","Episode = 199 | steps =  3393 | episode_length = 12 | epsilon = 0.22128109852834169\n","Episode = 200 | steps =  3402 | episode_length = 9 | epsilon = 0.22017469303569998\n","Episode = 201 | steps =  3411 | episode_length = 9 | epsilon = 0.21907381957052147\n","Episode = 202 | steps =  3421 | episode_length = 10 | epsilon = 0.21797845047266887\n","Episode = 203 | steps =  3431 | episode_length = 10 | epsilon = 0.21688855822030553\n","Episode = 204 | steps =  3444 | episode_length = 13 | epsilon = 0.215804115429204\n","Episode = 205 | steps =  3455 | episode_length = 11 | epsilon = 0.21472509485205798\n","Episode = 206 | steps =  3463 | episode_length = 8 | epsilon = 0.2136514693777977\n","Episode = 207 | steps =  3472 | episode_length = 9 | epsilon = 0.2125832120309087\n","Episode = 208 | steps =  3482 | episode_length = 10 | epsilon = 0.21152029597075417\n","Episode = 209 | steps =  3492 | episode_length = 10 | epsilon = 0.2104626944909004\n","Episode = 210 | steps =  3502 | episode_length = 10 | epsilon = 0.20941038101844592\n","Episode = 211 | steps =  3511 | episode_length = 9 | epsilon = 0.2083633291133537\n","Episode = 212 | steps =  3521 | episode_length = 10 | epsilon = 0.20732151246778693\n","Episode = 213 | steps =  3530 | episode_length = 9 | epsilon = 0.20628490490544799\n","Episode = 214 | steps =  3540 | episode_length = 10 | epsilon = 0.20525348038092076\n","Episode = 215 | steps =  3550 | episode_length = 10 | epsilon = 0.20422721297901614\n","Episode = 216 | steps =  3560 | episode_length = 10 | epsilon = 0.20320607691412104\n","Episode = 217 | steps =  3576 | episode_length = 16 | epsilon = 0.20219004652955044\n","Episode = 218 | steps =  3586 | episode_length = 10 | epsilon = 0.2011790962969027\n","Episode = 219 | steps =  3596 | episode_length = 10 | epsilon = 0.20017320081541817\n","Episode = 220 | steps =  3607 | episode_length = 11 | epsilon = 0.19917233481134108\n","Episode = 221 | steps =  3616 | episode_length = 9 | epsilon = 0.19817647313728437\n","Episode = 222 | steps =  3627 | episode_length = 11 | epsilon = 0.19718559077159795\n","Episode = 223 | steps =  3636 | episode_length = 9 | epsilon = 0.19619966281773996\n","Episode = 224 | steps =  3646 | episode_length = 10 | epsilon = 0.19521866450365125\n","Episode = 225 | steps =  3655 | episode_length = 9 | epsilon = 0.194242571181133\n","Episode = 226 | steps =  3665 | episode_length = 10 | epsilon = 0.19327135832522732\n","Episode = 227 | steps =  3677 | episode_length = 12 | epsilon = 0.1923050015336012\n","Episode = 228 | steps =  3689 | episode_length = 12 | epsilon = 0.1913434765259332\n","Episode = 229 | steps =  3698 | episode_length = 9 | epsilon = 0.19038675914330352\n","Episode = 230 | steps =  3710 | episode_length = 12 | epsilon = 0.18943482534758702\n","Episode = 231 | steps =  3720 | episode_length = 10 | epsilon = 0.18848765122084907\n","Episode = 232 | steps =  3729 | episode_length = 9 | epsilon = 0.18754521296474483\n","Episode = 233 | steps =  3737 | episode_length = 8 | epsilon = 0.1866074868999211\n","Episode = 234 | steps =  3747 | episode_length = 10 | epsilon = 0.1856744494654215\n","Episode = 235 | steps =  3756 | episode_length = 9 | epsilon = 0.18474607721809438\n","Episode = 236 | steps =  3768 | episode_length = 12 | epsilon = 0.1838223468320039\n","Episode = 237 | steps =  3776 | episode_length = 8 | epsilon = 0.1829032350978439\n","Episode = 238 | steps =  3787 | episode_length = 11 | epsilon = 0.1819887189223547\n","Episode = 239 | steps =  3796 | episode_length = 9 | epsilon = 0.1810787753277429\n","Episode = 240 | steps =  3805 | episode_length = 9 | epsilon = 0.1801733814511042\n","Episode = 241 | steps =  3815 | episode_length = 10 | epsilon = 0.17927251454384868\n","Episode = 242 | steps =  3825 | episode_length = 10 | epsilon = 0.17837615197112944\n","Episode = 243 | steps =  3834 | episode_length = 9 | epsilon = 0.1774842712112738\n","Episode = 244 | steps =  3844 | episode_length = 10 | epsilon = 0.17659684985521742\n","Episode = 245 | steps =  3854 | episode_length = 10 | epsilon = 0.17571386560594135\n","Episode = 246 | steps =  3863 | episode_length = 9 | epsilon = 0.17483529627791164\n","Episode = 247 | steps =  3874 | episode_length = 11 | epsilon = 0.17396111979652207\n","Episode = 248 | steps =  3887 | episode_length = 13 | epsilon = 0.17309131419753945\n","Episode = 249 | steps =  3898 | episode_length = 11 | epsilon = 0.17222585762655176\n","Episode = 250 | steps =  3910 | episode_length = 12 | epsilon = 0.171364728338419\n","Episode = 251 | steps =  3922 | episode_length = 12 | epsilon = 0.17050790469672691\n","Episode = 252 | steps =  3931 | episode_length = 9 | epsilon = 0.16965536517324328\n","Episode = 253 | steps =  3943 | episode_length = 12 | epsilon = 0.16880708834737707\n","Episode = 254 | steps =  3955 | episode_length = 12 | epsilon = 0.16796305290564018\n","Episode = 255 | steps =  3967 | episode_length = 12 | epsilon = 0.16712323764111198\n","Episode = 256 | steps =  3979 | episode_length = 12 | epsilon = 0.1662876214529064\n","Episode = 257 | steps =  3991 | episode_length = 12 | epsilon = 0.16545618334564188\n","Episode = 258 | steps =  4001 | episode_length = 10 | epsilon = 0.16462890242891368\n","Episode = 259 | steps =  4014 | episode_length = 13 | epsilon = 0.1638057579167691\n","Episode = 260 | steps =  4025 | episode_length = 11 | epsilon = 0.16298672912718526\n","Episode = 261 | steps =  4041 | episode_length = 16 | epsilon = 0.16217179548154934\n","Episode = 262 | steps =  4053 | episode_length = 12 | epsilon = 0.1613609365041416\n","Episode = 263 | steps =  4065 | episode_length = 12 | epsilon = 0.1605541318216209\n","Episode = 264 | steps =  4075 | episode_length = 10 | epsilon = 0.1597513611625128\n","Episode = 265 | steps =  4090 | episode_length = 15 | epsilon = 0.15895260435670022\n","Episode = 266 | steps =  4102 | episode_length = 12 | epsilon = 0.15815784133491673\n","Episode = 267 | steps =  4114 | episode_length = 12 | epsilon = 0.15736705212824215\n","Episode = 268 | steps =  4127 | episode_length = 13 | epsilon = 0.15658021686760093\n","Episode = 269 | steps =  4141 | episode_length = 14 | epsilon = 0.15579731578326292\n","Episode = 270 | steps =  4154 | episode_length = 13 | epsilon = 0.1550183292043466\n","Episode = 271 | steps =  4164 | episode_length = 10 | epsilon = 0.15424323755832486\n","Episode = 272 | steps =  4174 | episode_length = 10 | epsilon = 0.15347202137053323\n","Episode = 273 | steps =  4188 | episode_length = 14 | epsilon = 0.15270466126368057\n","Episode = 274 | steps =  4199 | episode_length = 11 | epsilon = 0.15194113795736217\n","Episode = 275 | steps =  4210 | episode_length = 11 | epsilon = 0.15118143226757536\n","Episode = 276 | steps =  4220 | episode_length = 10 | epsilon = 0.15042552510623747\n","Episode = 277 | steps =  4232 | episode_length = 12 | epsilon = 0.14967339748070627\n","Episode = 278 | steps =  4243 | episode_length = 11 | epsilon = 0.14892503049330275\n","Episode = 279 | steps =  4257 | episode_length = 14 | epsilon = 0.14818040534083624\n","Episode = 280 | steps =  4269 | episode_length = 12 | epsilon = 0.14743950331413205\n","Episode = 281 | steps =  4282 | episode_length = 13 | epsilon = 0.14670230579756138\n","Episode = 282 | steps =  4296 | episode_length = 14 | epsilon = 0.14596879426857357\n","Episode = 283 | steps =  4311 | episode_length = 15 | epsilon = 0.1452389502972307\n","Episode = 284 | steps =  4321 | episode_length = 10 | epsilon = 0.14451275554574453\n","Episode = 285 | steps =  4331 | episode_length = 10 | epsilon = 0.1437901917680158\n","Episode = 286 | steps =  4345 | episode_length = 14 | epsilon = 0.14307124080917571\n","Episode = 287 | steps =  4356 | episode_length = 11 | epsilon = 0.14235588460512982\n","Episode = 288 | steps =  4367 | episode_length = 11 | epsilon = 0.14164410518210419\n","Episode = 289 | steps =  4378 | episode_length = 11 | epsilon = 0.14093588465619367\n","Episode = 290 | steps =  4391 | episode_length = 13 | epsilon = 0.1402312052329127\n","Episode = 291 | steps =  4402 | episode_length = 11 | epsilon = 0.13953004920674814\n","Episode = 292 | steps =  4414 | episode_length = 12 | epsilon = 0.1388323989607144\n","Episode = 293 | steps =  4426 | episode_length = 12 | epsilon = 0.13813823696591082\n","Episode = 294 | steps =  4437 | episode_length = 11 | epsilon = 0.13744754578108126\n","Episode = 295 | steps =  4452 | episode_length = 15 | epsilon = 0.13676030805217584\n","Episode = 296 | steps =  4462 | episode_length = 10 | epsilon = 0.13607650651191497\n","Episode = 297 | steps =  4474 | episode_length = 12 | epsilon = 0.13539612397935538\n","Episode = 298 | steps =  4486 | episode_length = 12 | epsilon = 0.13471914335945862\n","Episode = 299 | steps =  4496 | episode_length = 10 | epsilon = 0.13404554764266133\n","Episode = 300 | steps =  4506 | episode_length = 10 | epsilon = 0.133375319904448\n","Episode = 301 | steps =  4516 | episode_length = 10 | epsilon = 0.13270844330492576\n","Episode = 302 | steps =  4526 | episode_length = 10 | epsilon = 0.13204490108840114\n","Episode = 303 | steps =  4539 | episode_length = 13 | epsilon = 0.13138467658295913\n","Episode = 304 | steps =  4551 | episode_length = 12 | epsilon = 0.13072775320004434\n","Episode = 305 | steps =  4561 | episode_length = 10 | epsilon = 0.13007411443404412\n","Episode = 306 | steps =  4572 | episode_length = 11 | epsilon = 0.1294237438618739\n","Episode = 307 | steps =  4584 | episode_length = 12 | epsilon = 0.12877662514256455\n","Episode = 308 | steps =  4596 | episode_length = 12 | epsilon = 0.12813274201685174\n","Episode = 309 | steps =  4610 | episode_length = 14 | epsilon = 0.12749207830676748\n","Episode = 310 | steps =  4622 | episode_length = 12 | epsilon = 0.12685461791523364\n","Episode = 311 | steps =  4635 | episode_length = 13 | epsilon = 0.12622034482565747\n","Episode = 312 | steps =  4645 | episode_length = 10 | epsilon = 0.1255892431015292\n","Episode = 313 | steps =  4657 | episode_length = 12 | epsilon = 0.12496129688602155\n","Episode = 314 | steps =  4670 | episode_length = 13 | epsilon = 0.12433649040159143\n","Episode = 315 | steps =  4680 | episode_length = 10 | epsilon = 0.12371480794958348\n","Episode = 316 | steps =  4696 | episode_length = 16 | epsilon = 0.12309623390983555\n","Episode = 317 | steps =  4706 | episode_length = 10 | epsilon = 0.12248075274028637\n","Episode = 318 | steps =  4720 | episode_length = 14 | epsilon = 0.12186834897658494\n","Episode = 319 | steps =  4730 | episode_length = 10 | epsilon = 0.12125900723170202\n","Episode = 320 | steps =  4741 | episode_length = 11 | epsilon = 0.12065271219554351\n","Episode = 321 | steps =  4754 | episode_length = 13 | epsilon = 0.12004944863456579\n","Episode = 322 | steps =  4770 | episode_length = 16 | epsilon = 0.11944920139139296\n","Episode = 323 | steps =  4784 | episode_length = 14 | epsilon = 0.118851955384436\n","Episode = 324 | steps =  4800 | episode_length = 16 | epsilon = 0.11825769560751383\n","Episode = 325 | steps =  4812 | episode_length = 12 | epsilon = 0.11766640712947626\n","Episode = 326 | steps =  4823 | episode_length = 11 | epsilon = 0.11707807509382888\n","Episode = 327 | steps =  4835 | episode_length = 12 | epsilon = 0.11649268471835973\n","Episode = 328 | steps =  4849 | episode_length = 14 | epsilon = 0.11591022129476793\n","Episode = 329 | steps =  4863 | episode_length = 14 | epsilon = 0.11533067018829408\n","Episode = 330 | steps =  4872 | episode_length = 9 | epsilon = 0.11475401683735262\n","Episode = 331 | steps =  4883 | episode_length = 11 | epsilon = 0.11418024675316585\n","Episode = 332 | steps =  4894 | episode_length = 11 | epsilon = 0.11360934551940002\n","Episode = 333 | steps =  4906 | episode_length = 12 | epsilon = 0.11304129879180302\n","Episode = 334 | steps =  4916 | episode_length = 10 | epsilon = 0.11247609229784401\n","Episode = 335 | steps =  4927 | episode_length = 11 | epsilon = 0.11191371183635479\n","Episode = 336 | steps =  4937 | episode_length = 10 | epsilon = 0.11135414327717301\n","Episode = 337 | steps =  4949 | episode_length = 12 | epsilon = 0.11079737256078714\n","Episode = 338 | steps =  4961 | episode_length = 12 | epsilon = 0.11024338569798321\n","Episode = 339 | steps =  4972 | episode_length = 11 | epsilon = 0.1096921687694933\n","Episode = 340 | steps =  4985 | episode_length = 13 | epsilon = 0.10914370792564583\n","Episode = 341 | steps =  4996 | episode_length = 11 | epsilon = 0.1085979893860176\n","Episode = 342 | steps =  5007 | episode_length = 11 | epsilon = 0.10805499943908752\n","Episode = 343 | steps =  5020 | episode_length = 13 | epsilon = 0.10751472444189207\n","Episode = 344 | steps =  5032 | episode_length = 12 | epsilon = 0.1069771508196826\n","Episode = 345 | steps =  5042 | episode_length = 10 | epsilon = 0.1064422650655842\n","Episode = 346 | steps =  5055 | episode_length = 13 | epsilon = 0.10591005374025628\n","Episode = 347 | steps =  5065 | episode_length = 10 | epsilon = 0.10538050347155499\n","Episode = 348 | steps =  5076 | episode_length = 11 | epsilon = 0.10485360095419721\n","Episode = 349 | steps =  5090 | episode_length = 14 | epsilon = 0.10432933294942623\n","Episode = 350 | steps =  5103 | episode_length = 13 | epsilon = 0.1038076862846791\n","Episode = 351 | steps =  5113 | episode_length = 10 | epsilon = 0.1032886478532557\n","Episode = 352 | steps =  5123 | episode_length = 10 | epsilon = 0.10277220461398942\n","Episode = 353 | steps =  5134 | episode_length = 11 | epsilon = 0.10225834359091947\n","Episode = 354 | steps =  5147 | episode_length = 13 | epsilon = 0.10174705187296487\n","Episode = 355 | steps =  5158 | episode_length = 11 | epsilon = 0.10123831661360004\n","Episode = 356 | steps =  5171 | episode_length = 13 | epsilon = 0.10073212503053204\n","Episode = 357 | steps =  5184 | episode_length = 13 | epsilon = 0.10022846440537939\n","Episode = 358 | steps =  5197 | episode_length = 13 | epsilon = 0.09972732208335248\n","Episode = 359 | steps =  5209 | episode_length = 12 | epsilon = 0.09972732208335248\n","Episode = 360 | steps =  5220 | episode_length = 11 | epsilon = 0.09972732208335248\n","Episode = 361 | steps =  5230 | episode_length = 10 | epsilon = 0.09972732208335248\n","Episode = 362 | steps =  5242 | episode_length = 12 | epsilon = 0.09972732208335248\n","Episode = 363 | steps =  5255 | episode_length = 13 | epsilon = 0.09972732208335248\n","Episode = 364 | steps =  5265 | episode_length = 10 | epsilon = 0.09972732208335248\n","Episode = 365 | steps =  5276 | episode_length = 11 | epsilon = 0.09972732208335248\n","Episode = 366 | steps =  5289 | episode_length = 13 | epsilon = 0.09972732208335248\n","Episode = 367 | steps =  5300 | episode_length = 11 | epsilon = 0.09972732208335248\n","Episode = 368 | steps =  5311 | episode_length = 11 | epsilon = 0.09972732208335248\n","Episode = 369 | steps =  5322 | episode_length = 11 | epsilon = 0.09972732208335248\n","Episode = 370 | steps =  5336 | episode_length = 14 | epsilon = 0.09972732208335248\n","Episode = 371 | steps =  5345 | episode_length = 9 | epsilon = 0.09972732208335248\n","Episode = 372 | steps =  5354 | episode_length = 9 | epsilon = 0.09972732208335248\n","Episode = 373 | steps =  5364 | episode_length = 10 | epsilon = 0.09972732208335248\n","Episode = 374 | steps =  5375 | episode_length = 11 | epsilon = 0.09972732208335248\n","Episode = 375 | steps =  5387 | episode_length = 12 | epsilon = 0.09972732208335248\n","Episode = 376 | steps =  5402 | episode_length = 15 | epsilon = 0.09972732208335248\n","Episode = 377 | steps =  5414 | episode_length = 12 | epsilon = 0.09972732208335248\n","Episode = 378 | steps =  5427 | episode_length = 13 | epsilon = 0.09972732208335248\n","Episode = 379 | steps =  5441 | episode_length = 14 | epsilon = 0.09972732208335248\n","Episode = 380 | steps =  5453 | episode_length = 12 | epsilon = 0.09972732208335248\n","Episode = 381 | steps =  5465 | episode_length = 12 | epsilon = 0.09972732208335248\n","Episode = 382 | steps =  5477 | episode_length = 12 | epsilon = 0.09972732208335248\n","Episode = 383 | steps =  5489 | episode_length = 12 | epsilon = 0.09972732208335248\n","Episode = 384 | steps =  5506 | episode_length = 17 | epsilon = 0.09972732208335248\n","Episode = 385 | steps =  5516 | episode_length = 10 | epsilon = 0.09972732208335248\n","Episode = 386 | steps =  5526 | episode_length = 10 | epsilon = 0.09972732208335248\n","Episode = 387 | steps =  5541 | episode_length = 15 | epsilon = 0.09972732208335248\n","Episode = 388 | steps =  5550 | episode_length = 9 | epsilon = 0.09972732208335248\n","Episode = 389 | steps =  5561 | episode_length = 11 | epsilon = 0.09972732208335248\n","Episode = 390 | steps =  5573 | episode_length = 12 | epsilon = 0.09972732208335248\n","Episode = 391 | steps =  5584 | episode_length = 11 | epsilon = 0.09972732208335248\n","Episode = 392 | steps =  5595 | episode_length = 11 | epsilon = 0.09972732208335248\n","Episode = 393 | steps =  5612 | episode_length = 17 | epsilon = 0.09972732208335248\n","Episode = 394 | steps =  5623 | episode_length = 11 | epsilon = 0.09972732208335248\n","Episode = 395 | steps =  5635 | episode_length = 12 | epsilon = 0.09972732208335248\n","Episode = 396 | steps =  5644 | episode_length = 9 | epsilon = 0.09972732208335248\n","Episode = 397 | steps =  5654 | episode_length = 10 | epsilon = 0.09972732208335248\n","Episode = 398 | steps =  5663 | episode_length = 9 | epsilon = 0.09972732208335248\n","Episode = 399 | steps =  5679 | episode_length = 16 | epsilon = 0.09972732208335248\n","Episode = 400 | steps =  5695 | episode_length = 16 | epsilon = 0.09972732208335248\n","Episode = 401 | steps =  5708 | episode_length = 13 | epsilon = 0.09972732208335248\n","Episode = 402 | steps =  5718 | episode_length = 10 | epsilon = 0.09972732208335248\n","Episode = 403 | steps =  5728 | episode_length = 10 | epsilon = 0.09972732208335248\n","Episode = 404 | steps =  5740 | episode_length = 12 | epsilon = 0.09972732208335248\n","Episode = 405 | steps =  5752 | episode_length = 12 | epsilon = 0.09972732208335248\n","Episode = 406 | steps =  5765 | episode_length = 13 | epsilon = 0.09972732208335248\n","Episode = 407 | steps =  5775 | episode_length = 10 | epsilon = 0.09972732208335248\n","Episode = 408 | steps =  5787 | episode_length = 12 | epsilon = 0.09972732208335248\n","Episode = 409 | steps =  5796 | episode_length = 9 | epsilon = 0.09972732208335248\n","Episode = 410 | steps =  5807 | episode_length = 11 | epsilon = 0.09972732208335248\n","Episode = 411 | steps =  5823 | episode_length = 16 | epsilon = 0.09972732208335248\n","Episode = 412 | steps =  5835 | episode_length = 12 | epsilon = 0.09972732208335248\n","Episode = 413 | steps =  5847 | episode_length = 12 | epsilon = 0.09972732208335248\n","Episode = 414 | steps =  5859 | episode_length = 12 | epsilon = 0.09972732208335248\n","Episode = 415 | steps =  5873 | episode_length = 14 | epsilon = 0.09972732208335248\n","Episode = 416 | steps =  5887 | episode_length = 14 | epsilon = 0.09972732208335248\n","Episode = 417 | steps =  5900 | episode_length = 13 | epsilon = 0.09972732208335248\n","Episode = 418 | steps =  5910 | episode_length = 10 | epsilon = 0.09972732208335248\n","Episode = 419 | steps =  5923 | episode_length = 13 | epsilon = 0.09972732208335248\n","Episode = 420 | steps =  5936 | episode_length = 13 | epsilon = 0.09972732208335248\n","Episode = 421 | steps =  5952 | episode_length = 16 | epsilon = 0.09972732208335248\n","Episode = 422 | steps =  5966 | episode_length = 14 | epsilon = 0.09972732208335248\n","Episode = 423 | steps =  5981 | episode_length = 15 | epsilon = 0.09972732208335248\n","Episode = 424 | steps =  5995 | episode_length = 14 | epsilon = 0.09972732208335248\n","Episode = 425 | steps =  6008 | episode_length = 13 | epsilon = 0.09972732208335248\n","Episode = 426 | steps =  6022 | episode_length = 14 | epsilon = 0.09972732208335248\n","Episode = 427 | steps =  6032 | episode_length = 10 | epsilon = 0.09972732208335248\n","Episode = 428 | steps =  6045 | episode_length = 13 | epsilon = 0.09972732208335248\n","Episode = 429 | steps =  6057 | episode_length = 12 | epsilon = 0.09972732208335248\n","Episode = 430 | steps =  6070 | episode_length = 13 | epsilon = 0.09972732208335248\n","Episode = 431 | steps =  6084 | episode_length = 14 | epsilon = 0.09972732208335248\n","Episode = 432 | steps =  6099 | episode_length = 15 | epsilon = 0.09972732208335248\n","Episode = 433 | steps =  6112 | episode_length = 13 | epsilon = 0.09972732208335248\n","Episode = 434 | steps =  6124 | episode_length = 12 | epsilon = 0.09972732208335248\n","Episode = 435 | steps =  6140 | episode_length = 16 | epsilon = 0.09972732208335248\n","Episode = 436 | steps =  6155 | episode_length = 15 | epsilon = 0.09972732208335248\n","Episode = 437 | steps =  6168 | episode_length = 13 | epsilon = 0.09972732208335248\n","Episode = 438 | steps =  6185 | episode_length = 17 | epsilon = 0.09972732208335248\n","Episode = 439 | steps =  6200 | episode_length = 15 | epsilon = 0.09972732208335248\n","Episode = 440 | steps =  6215 | episode_length = 15 | epsilon = 0.09972732208335248\n","Episode = 441 | steps =  6232 | episode_length = 17 | epsilon = 0.09972732208335248\n","Episode = 442 | steps =  6245 | episode_length = 13 | epsilon = 0.09972732208335248\n","Episode = 443 | steps =  6261 | episode_length = 16 | epsilon = 0.09972732208335248\n","Episode = 444 | steps =  6273 | episode_length = 12 | epsilon = 0.09972732208335248\n","Episode = 445 | steps =  6289 | episode_length = 16 | epsilon = 0.09972732208335248\n","Episode = 446 | steps =  6304 | episode_length = 15 | epsilon = 0.09972732208335248\n","Episode = 447 | steps =  6318 | episode_length = 14 | epsilon = 0.09972732208335248\n","Episode = 448 | steps =  6330 | episode_length = 12 | epsilon = 0.09972732208335248\n","Episode = 449 | steps =  6347 | episode_length = 17 | epsilon = 0.09972732208335248\n","Episode = 450 | steps =  6362 | episode_length = 15 | epsilon = 0.09972732208335248\n","Episode = 451 | steps =  6377 | episode_length = 15 | epsilon = 0.09972732208335248\n","Episode = 452 | steps =  6393 | episode_length = 16 | epsilon = 0.09972732208335248\n","Episode = 453 | steps =  6408 | episode_length = 15 | epsilon = 0.09972732208335248\n","Episode = 454 | steps =  6426 | episode_length = 18 | epsilon = 0.09972732208335248\n","Episode = 455 | steps =  6441 | episode_length = 15 | epsilon = 0.09972732208335248\n","Episode = 456 | steps =  6455 | episode_length = 14 | epsilon = 0.09972732208335248\n","Episode = 457 | steps =  6470 | episode_length = 15 | epsilon = 0.09972732208335248\n","Episode = 458 | steps =  6484 | episode_length = 14 | epsilon = 0.09972732208335248\n","Episode = 459 | steps =  6499 | episode_length = 15 | epsilon = 0.09972732208335248\n","Episode = 460 | steps =  6513 | episode_length = 14 | epsilon = 0.09972732208335248\n","Episode = 461 | steps =  6530 | episode_length = 17 | epsilon = 0.09972732208335248\n","Episode = 462 | steps =  6548 | episode_length = 18 | epsilon = 0.09972732208335248\n","Episode = 463 | steps =  6564 | episode_length = 16 | epsilon = 0.09972732208335248\n","Episode = 464 | steps =  6621 | episode_length = 57 | epsilon = 0.09972732208335248\n","Episode = 465 | steps =  6642 | episode_length = 21 | epsilon = 0.09972732208335248\n","Episode = 466 | steps =  6659 | episode_length = 17 | epsilon = 0.09972732208335248\n","Episode = 467 | steps =  6680 | episode_length = 21 | epsilon = 0.09972732208335248\n","Episode = 468 | steps =  6743 | episode_length = 63 | epsilon = 0.09972732208335248\n","Episode = 469 | steps =  6812 | episode_length = 69 | epsilon = 0.09972732208335248\n","Episode = 470 | steps =  6899 | episode_length = 87 | epsilon = 0.09972732208335248\n","Episode = 471 | steps =  6919 | episode_length = 20 | epsilon = 0.09972732208335248\n","Episode = 472 | steps =  6998 | episode_length = 79 | epsilon = 0.09972732208335248\n","Episode = 473 | steps =  7018 | episode_length = 20 | epsilon = 0.09972732208335248\n","Episode = 474 | steps =  7121 | episode_length = 103 | epsilon = 0.09972732208335248\n","Episode = 475 | steps =  7201 | episode_length = 80 | epsilon = 0.09972732208335248\n","Episode = 476 | steps =  7298 | episode_length = 97 | epsilon = 0.09972732208335248\n","Episode = 477 | steps =  7314 | episode_length = 16 | epsilon = 0.09972732208335248\n","Episode = 478 | steps =  7379 | episode_length = 65 | epsilon = 0.09972732208335248\n","Episode = 479 | steps =  7414 | episode_length = 35 | epsilon = 0.09972732208335248\n","Episode = 480 | steps =  7455 | episode_length = 41 | epsilon = 0.09972732208335248\n","Episode = 481 | steps =  7488 | episode_length = 33 | epsilon = 0.09972732208335248\n","Episode = 482 | steps =  7522 | episode_length = 34 | epsilon = 0.09972732208335248\n","Episode = 483 | steps =  7556 | episode_length = 34 | epsilon = 0.09972732208335248\n","Episode = 484 | steps =  7602 | episode_length = 46 | epsilon = 0.09972732208335248\n","Episode = 485 | steps =  7651 | episode_length = 49 | epsilon = 0.09972732208335248\n","Episode = 486 | steps =  7659 | episode_length = 8 | epsilon = 0.09972732208335248\n","Episode = 487 | steps =  7668 | episode_length = 9 | epsilon = 0.09972732208335248\n","Episode = 488 | steps =  7693 | episode_length = 25 | epsilon = 0.09972732208335248\n","Episode = 489 | steps =  7713 | episode_length = 20 | epsilon = 0.09972732208335248\n","Episode = 490 | steps =  7723 | episode_length = 10 | epsilon = 0.09972732208335248\n","Episode = 491 | steps =  7733 | episode_length = 10 | epsilon = 0.09972732208335248\n","Episode = 492 | steps =  7744 | episode_length = 11 | epsilon = 0.09972732208335248\n","Episode = 493 | steps =  7753 | episode_length = 9 | epsilon = 0.09972732208335248\n","Episode = 494 | steps =  7763 | episode_length = 10 | epsilon = 0.09972732208335248\n","Episode = 495 | steps =  7773 | episode_length = 10 | epsilon = 0.09972732208335248\n","Episode = 496 | steps =  7783 | episode_length = 10 | epsilon = 0.09972732208335248\n","Episode = 497 | steps =  7793 | episode_length = 10 | epsilon = 0.09972732208335248\n","Episode = 498 | steps =  7802 | episode_length = 9 | epsilon = 0.09972732208335248\n","Episode = 499 | steps =  7811 | episode_length = 9 | epsilon = 0.09972732208335248\n","Episode = 500 | steps =  7821 | episode_length = 10 | epsilon = 0.09972732208335248\n","Episode = 501 | steps =  7833 | episode_length = 12 | epsilon = 0.09972732208335248\n","Episode = 502 | steps =  7842 | episode_length = 9 | epsilon = 0.09972732208335248\n","Episode = 503 | steps =  7851 | episode_length = 9 | epsilon = 0.09972732208335248\n","Episode = 504 | steps =  7861 | episode_length = 10 | epsilon = 0.09972732208335248\n","Episode = 505 | steps =  7871 | episode_length = 10 | epsilon = 0.09972732208335248\n","Episode = 506 | steps =  7882 | episode_length = 11 | epsilon = 0.09972732208335248\n","Episode = 507 | steps =  7891 | episode_length = 9 | epsilon = 0.09972732208335248\n","Episode = 508 | steps =  7903 | episode_length = 12 | epsilon = 0.09972732208335248\n","Episode = 509 | steps =  7916 | episode_length = 13 | epsilon = 0.09972732208335248\n","Episode = 510 | steps =  7924 | episode_length = 8 | epsilon = 0.09972732208335248\n","Episode = 511 | steps =  7933 | episode_length = 9 | epsilon = 0.09972732208335248\n","Episode = 512 | steps =  7953 | episode_length = 20 | epsilon = 0.09972732208335248\n","Episode = 513 | steps =  7970 | episode_length = 17 | epsilon = 0.09972732208335248\n","Episode = 514 | steps =  7988 | episode_length = 18 | epsilon = 0.09972732208335248\n","Episode = 515 | steps =  8013 | episode_length = 25 | epsilon = 0.09972732208335248\n"]}]},{"cell_type":"code","source":["def visualize_progress_multiple(experiments_progresses: dict):\n","  plt.clf()\n","  for experiment_name, experiment_progress in experiments_progresses.items():\n","    # plt.plot(experiment_progress, label=f\"{experiment_name}: progress\")\n","    smoothed_progress = [0]\n","    for x in experiment_progress:\n","      smoothed_progress.append(0.8*smoothed_progress[-1] + 0.2*x)\n","    plt.plot(smoothed_progress, label=f\"{experiment_name}: learning (smoothed)\")\n","  plt.legend(loc=\"upper left\")\n","  plt.show()\n","\n","experiments_progresses ={\n","    'DQN standard, unlimited buffer': progress_standard,\n","    'DQN no target, unlimited buffer': progress_no_target,\n","    'DQN standard, buffer_size=500': progress_limit_buffer,\n","    'DQN no target, buffer_size=500': progress_no_target_limit_buffer,\n","}\n","\n","visualize_progress_multiple(experiments_progresses)"],"metadata":{"colab":{"base_uri":"https://localhost:8080/","height":432},"id":"g4h6iI_sJJh-","executionInfo":{"status":"ok","timestamp":1680284317122,"user_tz":-120,"elapsed":680,"user":{"displayName":"Michał Janik","userId":"10483929333441523386"}},"outputId":"62487cbe-168a-48bb-f399-6c1ce558e528"},"execution_count":23,"outputs":[{"output_type":"display_data","data":{"text/plain":["<Figure size 640x480 with 1 Axes>"],"image/png":"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\n"},"metadata":{}}]},{"cell_type":"markdown","source":["Using target network seems to stabilise the training process and make model learn faster.\n","\n","Limiting the size of replay buffer results in poorer performance and sometimes forgetting useful transitions."],"metadata":{"id":"ej6L4TNBrfmK"}},{"cell_type":"code","source":[],"metadata":{"id":"BMTFXj66r3BK"},"execution_count":null,"outputs":[]}]}