model-lsr-cont.ipynb

{
 "cells": [
  {
   "cell_type": "code",
   "execution_count": 1,
   "metadata": {},
   "outputs": [],
   "source": [
    "# SiFi-NN trained on the simulation with continuous source then evaluated on \n",
    "# the Bragg containing simulation\n",
    "import numpy as np\n",
    "import uproot\n",
    "import matplotlib.pyplot as plt\n",
    "from tensorflow import keras\n",
    "import tensorflow.keras.backend as K\n",
    "\n",
    "from sificc_lib import AI, utils, Event, Simulation, root_files, DataModel\n",
    "np.set_printoptions(precision=2, linewidth=85, suppress=True)\n",
    "\n",
    "%matplotlib inline"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 2,
   "metadata": {},
   "outputs": [],
   "source": [
    "# The main AI trained on all the data of the new simulation\n",
    "model_name = 'model-lsr-cont'\n",
    "\n",
    "shuffle_clusters = False\n",
    "\n",
    "# load the training data\n",
    "data = DataModel('data-cont-top-8.npz', \n",
    "                 batch_size = 128, validation_percent = .05, test_percent = .05, \n",
    "                 weight_compton=1, weight_non_compton=1)\n",
    "\n",
    "# append an extra dimention to the features since we are using convolutional layers\n",
    "data.append_dim = True\n",
    "\n",
    "# create an AI instance\n",
    "ai = AI(data, model_name)\n",
    "\n",
    "ai.weight_type = 1.4\n",
    "ai.weight_pos_x = 2.5\n",
    "ai.weight_pos_y = .5\n",
    "ai.weight_pos_z = 2\n",
    "ai.weight_energy = .8\n",
    "ai.weight_e_cluster = .6\n",
    "ai.weight_p_cluster = .4\n",
    "\n",
    "# shuffle the clusters within each event\n",
    "if shuffle_clusters:\n",
    "    ai.data.shuffle_training_clusters()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 5,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Model: \"model_1\"\n",
      "__________________________________________________________________________________________________\n",
      "Layer (type)                    Output Shape         Param #     Connected to                     \n",
      "==================================================================================================\n",
      "inputs (InputLayer)             [(None, 72, 1)]      0                                            \n",
      "__________________________________________________________________________________________________\n",
      "conv_1 (Conv1D)                 (None, 8, 128)       1280        inputs[0][0]                     \n",
      "__________________________________________________________________________________________________\n",
      "conv_2 (Conv1D)                 (None, 8, 64)        8256        conv_1[0][0]                     \n",
      "__________________________________________________________________________________________________\n",
      "flatting (Flatten)              (None, 512)          0           conv_2[0][0]                     \n",
      "__________________________________________________________________________________________________\n",
      "dense_cluster_1 (Dense)         (None, 32)           16416       flatting[0][0]                   \n",
      "__________________________________________________________________________________________________\n",
      "e_cluster (Dense)               (None, 8)            264         dense_cluster_1[0][0]            \n",
      "__________________________________________________________________________________________________\n",
      "p_cluster (Dense)               (None, 8)            264         dense_cluster_1[0][0]            \n",
      "__________________________________________________________________________________________________\n",
      "e_hardmax (Lambda)              (None, 8)            0           e_cluster[0][0]                  \n",
      "__________________________________________________________________________________________________\n",
      "p_hardmax (Lambda)              (None, 8)            0           p_cluster[0][0]                  \n",
      "__________________________________________________________________________________________________\n",
      "join_layer (Concatenate)        (None, 528)          0           flatting[0][0]                   \n",
      "                                                                 e_hardmax[0][0]                  \n",
      "                                                                 p_hardmax[0][0]                  \n",
      "__________________________________________________________________________________________________\n",
      "dense_type_1 (Dense)            (None, 16)           8464        join_layer[0][0]                 \n",
      "__________________________________________________________________________________________________\n",
      "dense_pos_1 (Dense)             (None, 64)           33856       join_layer[0][0]                 \n",
      "__________________________________________________________________________________________________\n",
      "dense_energy_1 (Dense)          (None, 32)           16928       join_layer[0][0]                 \n",
      "__________________________________________________________________________________________________\n",
      "dense_type_2 (Dense)            (None, 8)            136         dense_type_1[0][0]               \n",
      "__________________________________________________________________________________________________\n",
      "dense_pos_2 (Dense)             (None, 32)           2080        dense_pos_1[0][0]                \n",
      "__________________________________________________________________________________________________\n",
      "dense_energy_2 (Dense)          (None, 16)           528         dense_energy_1[0][0]             \n",
      "__________________________________________________________________________________________________\n",
      "type (Dense)                    (None, 1)            9           dense_type_2[0][0]               \n",
      "__________________________________________________________________________________________________\n",
      "pos_x (Dense)                   (None, 2)            66          dense_pos_2[0][0]                \n",
      "__________________________________________________________________________________________________\n",
      "pos_y (Dense)                   (None, 2)            66          dense_pos_2[0][0]                \n",
      "__________________________________________________________________________________________________\n",
      "pos_z (Dense)                   (None, 2)            66          dense_pos_2[0][0]                \n",
      "__________________________________________________________________________________________________\n",
      "energy (Dense)                  (None, 2)            34          dense_energy_2[0][0]             \n",
      "==================================================================================================\n",
      "Total params: 88,713\n",
      "Trainable params: 88,713\n",
      "Non-trainable params: 0\n",
      "__________________________________________________________________________________________________\n"
     ]
    }
   ],
   "source": [
    "# define and create the neural network architecture\n",
    "ai.create_model(conv_layers=[128, 64], classifier_layers=[32], type_layers=[16, 8], \n",
    "                pos_layers=[64,32], energy_layers=[32, 16], base_l2=.0001, limbs_l2=.0001)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 6,
   "metadata": {},
   "outputs": [],
   "source": [
    "# compile the ai\n",
    "ai.compile_model(learning_rate=0.003)\n",
    "\n",
    "# define the learning rate scheduler for the training phase\n",
    "def lr_scheduler(epoch):\n",
    "    if epoch < 60:\n",
    "        return .003\n",
    "    elif epoch < 110:\n",
    "        return .001\n",
    "    elif epoch < 140:\n",
    "        return .0003\n",
    "    elif epoch < 165:\n",
    "        return .0001\n",
    "    elif epoch < 185:\n",
    "        return .00003\n",
    "    elif epoch < 195:\n",
    "        return .00001\n",
    "    else:\n",
    "        return .000003\n",
    "l_callbacks = [\n",
    "    keras.callbacks.LearningRateScheduler(lr_scheduler),\n",
    "]"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 6,
   "metadata": {},
   "outputs": [],
   "source": [
    "#LOADING\n",
    "ai.data = DataModel('data-mtx-top-8.npz', \n",
    "                 batch_size = 128, validation_percent = .1, test_percent = .2, \n",
    "                 weight_compton=1, weight_non_compton=1)\n",
    "ai.data.append_dim = True\n",
    "ai.load(model_name, optimizer=False)\n",
    "ai.compile_model()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 7,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "AI model\n",
      "  Loss:        2.43298\n",
      "    -Type:         0.42653 *  1.40 = 0.59714\n",
      "    -Pos X:        0.05752 *  2.50 = 0.14380\n",
      "    -Pos Y:        1.17863 *  0.50 = 0.58931\n",
      "    -Pos Z:        0.02579 *  2.00 = 0.05158\n",
      "    -Energy:       1.12075 *  0.80 = 0.89660\n",
      "    -Cls e:        0.01835 *  0.60 = 0.01101\n",
      "    -Cls p:        0.05176 *  0.40 = 0.02070\n",
      "  Accuracy:     0.79162\n",
      "    -Precision:    0.43735\n",
      "    -Recall:       0.86306\n",
      "    -Cls e rate:   0.95838\n",
      "    -Cls p rate:   0.87384\n",
      "  Efficiency:   0.20632\n",
      "  Purity:       0.10455\n",
      "  Euc mean:    10.52290\n",
      "  Euc std:     17.80590\n",
      "  Energy mean:  0.35085\n",
      "  Energy std:   0.81540\n",
      "\n",
      "Reco\n",
      "  Accuracy:     0.66023\n",
      "    -Recall:       0.76769\n",
      "  Efficiency:   0.13425\n",
      "  Purity:       0.05225\n",
      "  Euc mean:    11.82275\n",
      "  Euc std:     19.30429\n",
      "  Energy mean:  0.50111\n",
      "  Energy std:   1.01502\n"
     ]
    }
   ],
   "source": [
    "ai.type_threshold = .5\n",
    "ai.evaluate()\n",
    "ai.export_predictions_root(f'sificc-nn-0mm-shift.root')"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 9,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "AI model\n",
      "  Loss:        2.43298\n",
      "    -Type:         0.42653 *  1.40 = 0.59714\n",
      "    -Pos X:        0.05752 *  2.50 = 0.14380\n",
      "    -Pos Y:        1.17863 *  0.50 = 0.58931\n",
      "    -Pos Z:        0.02579 *  2.00 = 0.05158\n",
      "    -Energy:       1.12075 *  0.80 = 0.89660\n",
      "    -Cls e:        0.01835 *  0.60 = 0.01101\n",
      "    -Cls p:        0.05176 *  0.40 = 0.02070\n",
      "  Accuracy:     0.79162\n",
      "    -Precision:    0.57210\n",
      "    -Recall:       0.55901\n",
      "    -Cls e rate:   0.95838\n",
      "    -Cls p rate:   0.87384\n",
      "  Efficiency:   0.12772\n",
      "  Purity:       0.13071\n",
      "  Euc mean:    11.02753\n",
      "  Euc std:     17.19937\n",
      "  Energy mean:  0.28427\n",
      "  Energy std:   0.70752\n",
      "\n",
      "Reco\n",
      "  Accuracy:     0.66023\n",
      "    -Recall:       0.76769\n",
      "  Efficiency:   0.13425\n",
      "  Purity:       0.05225\n",
      "  Euc mean:    11.82275\n",
      "  Euc std:     19.30429\n",
      "  Energy mean:  0.50111\n",
      "  Energy std:   1.01502\n"
     ]
    }
   ],
   "source": [
    "# evaluate the AI on the dataset with Bragg peak with a higher threshold\n",
    "ai.type_threshold = .8\n",
    "ai.evaluate()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 10,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "AI model\n",
      "  Loss:        2.43298\n",
      "    -Type:         0.42653 *  1.40 = 0.59714\n",
      "    -Pos X:        0.05752 *  2.50 = 0.14380\n",
      "    -Pos Y:        1.17863 *  0.50 = 0.58931\n",
      "    -Pos Z:        0.02579 *  2.00 = 0.05158\n",
      "    -Energy:       1.12075 *  0.80 = 0.89660\n",
      "    -Cls e:        0.01835 *  0.60 = 0.01101\n",
      "    -Cls p:        0.05176 *  0.40 = 0.02070\n",
      "  Accuracy:     0.79162\n",
      "    -Precision:    0.61629\n",
      "    -Recall:       0.38172\n",
      "    -Cls e rate:   0.95838\n",
      "    -Cls p rate:   0.87384\n",
      "  Efficiency:   0.08319\n",
      "  Purity:       0.13432\n",
      "  Euc mean:    11.38133\n",
      "  Euc std:     16.39092\n",
      "  Energy mean:  0.25692\n",
      "  Energy std:   0.67644\n",
      "\n",
      "Reco\n",
      "  Accuracy:     0.66023\n",
      "    -Recall:       0.76769\n",
      "  Efficiency:   0.13425\n",
      "  Purity:       0.05225\n",
      "  Euc mean:    11.82275\n",
      "  Euc std:     19.30429\n",
      "  Energy mean:  0.50111\n",
      "  Energy std:   1.01502\n"
     ]
    }
   ],
   "source": [
    "# evaluate the AI on the dataset with Bragg peak with a higher threshold\n",
    "ai.type_threshold = .85\n",
    "ai.evaluate()\n",
    "ai.export_predictions_root(f'sificc-nnq-0mm-shift.root')"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 7,
   "metadata": {},
   "outputs": [],
   "source": [
    "#LOADING\n",
    "ai.data = DataModel('data-simulated-5mm-shift.npz', \n",
    "                 batch_size = 128, validation_percent = .0, test_percent = 1.0, \n",
    "                 weight_compton=1, weight_non_compton=1)\n",
    "ai.data.append_dim = True\n",
    "ai.load(model_name, optimizer=False)\n",
    "ai.compile_model()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 9,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "AI model\n",
      "  Loss:        2.46595\n",
      "    -Type:         0.43234 *  1.40 = 0.60528\n",
      "    -Pos X:        0.06580 *  2.50 = 0.16449\n",
      "    -Pos Y:        1.17977 *  0.50 = 0.58988\n",
      "    -Pos Z:        0.02725 *  2.00 = 0.05450\n",
      "    -Energy:       1.12113 *  0.80 = 0.89690\n",
      "    -Cls e:        0.01805 *  0.60 = 0.01083\n",
      "    -Cls p:        0.05309 *  0.40 = 0.02124\n",
      "  Accuracy:     0.78827\n",
      "    -Precision:    0.43502\n",
      "    -Recall:       0.85893\n",
      "    -Cls e rate:   0.95899\n",
      "    -Cls p rate:   0.87407\n",
      "  Efficiency:   0.20680\n",
      "  Purity:       0.10474\n",
      "  Euc mean:    10.41328\n",
      "  Euc std:     17.46351\n",
      "  Energy mean:  0.35066\n",
      "  Energy std:   0.80565\n",
      "\n",
      "Reco\n",
      "  Accuracy:     0.69363\n",
      "    -Recall:       0.70333\n",
      "  Efficiency:   0.11652\n",
      "  Purity:       0.05237\n",
      "  Euc mean:    11.49812\n",
      "  Euc std:     18.17223\n",
      "  Energy mean:  0.50401\n",
      "  Energy std:   0.99822\n"
     ]
    }
   ],
   "source": [
    "ai.type_threshold = .5\n",
    "ai.evaluate()\n",
    "ai.export_predictions_root(f'sificc-nn-simulated-5mm-shift.root')"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 10,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "AI model\n",
      "  Loss:        2.46595\n",
      "    -Type:         0.43234 *  1.40 = 0.60528\n",
      "    -Pos X:        0.06580 *  2.50 = 0.16449\n",
      "    -Pos Y:        1.17977 *  0.50 = 0.58988\n",
      "    -Pos Z:        0.02725 *  2.00 = 0.05450\n",
      "    -Energy:       1.12113 *  0.80 = 0.89690\n",
      "    -Cls e:        0.01805 *  0.60 = 0.01083\n",
      "    -Cls p:        0.05309 *  0.40 = 0.02124\n",
      "  Accuracy:     0.78827\n",
      "    -Precision:    0.61247\n",
      "    -Recall:       0.38063\n",
      "    -Cls e rate:   0.95899\n",
      "    -Cls p rate:   0.87407\n",
      "  Efficiency:   0.08628\n",
      "  Purity:       0.13883\n",
      "  Euc mean:    11.27999\n",
      "  Euc std:     16.56245\n",
      "  Energy mean:  0.25061\n",
      "  Energy std:   0.62420\n",
      "\n",
      "Reco\n",
      "  Accuracy:     0.69363\n",
      "    -Recall:       0.70333\n",
      "  Efficiency:   0.11652\n",
      "  Purity:       0.05237\n",
      "  Euc mean:    11.49812\n",
      "  Euc std:     18.17223\n",
      "  Energy mean:  0.50401\n",
      "  Energy std:   0.99822\n"
     ]
    }
   ],
   "source": [
    "ai.type_threshold = .85\n",
    "ai.evaluate()\n",
    "ai.export_predictions_root(f'sificc-nnq-simulated-5mm-shift.root')"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 6,
   "metadata": {},
   "outputs": [],
   "source": [
    "#TRAINING"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 7,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "6983/6983 [==============================] - 12s 2ms/step - loss: 5.5790 - e_cluster_loss: 0.0459 - p_cluster_loss: 0.1641 - type_loss: 0.4328 - pos_x_loss: 0.1740 - pos_y_loss: 3.7442 - pos_z_loss: 0.0681 - energy_loss: 2.8930 - e_cluster__cluster_accuracy: 0.9645 - p_cluster__cluster_accuracy: 0.8662 - type__type_accuracy: 0.8019 - type__type_tp_rate: 0.8415\n",
      "\n",
      "CPU times: user 6h 38min 18s, sys: 2min 7s, total: 6h 40min 26s\n",
      "Wall time: 6h 44min 29s\n"
     ]
    }
   ],
   "source": [
    "%%time\n",
    "# balance the training data since there are too many background events\n",
    "ai.data.balance_training = True\n",
    "\n",
    "# start the training\n",
    "ai.train(epochs=200, shuffle_clusters=shuffle_clusters, verbose=0, callbacks = l_callbacks)\n",
    "\n",
    "# evaluate the AI on the training set\n",
    "ai.model.evaluate(ai.data.train_x, ai.data.train_y, verbose=1)\n",
    "print()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 8,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 864x288 with 2 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "# plot the training loss\n",
    "ai.plot_training_loss(smooth=False)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 9,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "AI model\n",
      "  Loss:        2.17190\n",
      "    -Type:         0.44635 *  1.40 = 0.62489\n",
      "    -Pos X:        0.04579 *  2.50 = 0.11447\n",
      "    -Pos Y:        1.11014 *  0.50 = 0.55507\n",
      "    -Pos Z:        0.02200 *  2.00 = 0.04400\n",
      "    -Energy:       0.85343 *  0.80 = 0.68275\n",
      "    -Cls e:        0.01525 *  0.60 = 0.00915\n",
      "    -Cls p:        0.04859 *  0.40 = 0.01944\n",
      "  Accuracy:     0.77495\n",
      "    -Precision:    0.37512\n",
      "    -Recall:       0.83537\n",
      "    -Cls e rate:   0.95973\n",
      "    -Cls p rate:   0.86501\n",
      "  Efficiency:   0.19362\n",
      "  Purity:       0.08694\n",
      "  Euc mean:    11.27563\n",
      "  Euc std:     18.17045\n",
      "  Energy mean:  0.31195\n",
      "  Energy std:   0.74586\n",
      "\n",
      "Reco\n",
      "  Accuracy:     0.68932\n",
      "    -Recall:       0.67155\n",
      "  Efficiency:   0.12436\n",
      "  Purity:       0.04991\n",
      "  Euc mean:    12.34355\n",
      "  Euc std:     18.71795\n",
      "  Energy mean:  0.44726\n",
      "  Energy std:   0.93658\n"
     ]
    }
   ],
   "source": [
    "# evaluate the AI on the test dataset\n",
    "ai.evaluate()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 9,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "AI model\n",
      "  Loss:        2.43709\n",
      "    -Type:         0.42801 *  1.40 = 0.59922\n",
      "    -Pos X:        0.06306 *  2.50 = 0.15764\n",
      "    -Pos Y:        1.17162 *  0.50 = 0.58581\n",
      "    -Pos Z:        0.02580 *  2.00 = 0.05161\n",
      "    -Energy:       1.11039 *  0.80 = 0.88831\n",
      "    -Cls e:        0.01803 *  0.60 = 0.01082\n",
      "    -Cls p:        0.05214 *  0.40 = 0.02086\n",
      "  Accuracy:     0.79098\n",
      "    -Precision:    0.43641\n",
      "    -Recall:       0.86416\n",
      "    -Cls e rate:   0.95865\n",
      "    -Cls p rate:   0.87297\n",
      "  Efficiency:   0.20564\n",
      "  Purity:       0.10385\n",
      "  Euc mean:    10.42945\n",
      "  Euc std:     17.45239\n",
      "  Energy mean:  0.34909\n",
      "  Energy std:   0.81039\n",
      "\n",
      "Reco\n",
      "  Accuracy:     0.65981\n",
      "    -Recall:       0.76909\n",
      "  Efficiency:   0.13441\n",
      "  Purity:       0.05219\n",
      "  Euc mean:    11.79185\n",
      "  Euc std:     19.09710\n",
      "  Energy mean:  0.49799\n",
      "  Energy std:   1.00702\n"
     ]
    }
   ],
   "source": [
    "# evaluate the AI on the dataset with Bragg peak\n",
    "ai.data = DataModel('data-mtx-top-8.npz', \n",
    "                 batch_size = 128, validation_percent = .1, test_percent = .4, \n",
    "                 weight_compton=1, weight_non_compton=1)\n",
    "ai.data.append_dim = True\n",
    "ai.type_threshold = .5\n",
    "ai.evaluate()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 10,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "AI model\n",
      "  Loss:        2.43709\n",
      "    -Type:         0.42801 *  1.40 = 0.59922\n",
      "    -Pos X:        0.06306 *  2.50 = 0.15764\n",
      "    -Pos Y:        1.17162 *  0.50 = 0.58581\n",
      "    -Pos Z:        0.02580 *  2.00 = 0.05161\n",
      "    -Energy:       1.11039 *  0.80 = 0.88831\n",
      "    -Cls e:        0.01803 *  0.60 = 0.01082\n",
      "    -Cls p:        0.05214 *  0.40 = 0.02086\n",
      "  Accuracy:     0.79098\n",
      "    -Precision:    0.57154\n",
      "    -Recall:       0.55913\n",
      "    -Cls e rate:   0.95865\n",
      "    -Cls p rate:   0.87297\n",
      "  Efficiency:   0.12952\n",
      "  Purity:       0.13239\n",
      "  Euc mean:    10.92532\n",
      "  Euc std:     16.90697\n",
      "  Energy mean:  0.27513\n",
      "  Energy std:   0.67306\n",
      "\n",
      "Reco\n",
      "  Accuracy:     0.65981\n",
      "    -Recall:       0.76909\n",
      "  Efficiency:   0.13441\n",
      "  Purity:       0.05219\n",
      "  Euc mean:    11.79185\n",
      "  Euc std:     19.09710\n",
      "  Energy mean:  0.49799\n",
      "  Energy std:   1.00702\n"
     ]
    }
   ],
   "source": [
    "# evaluate the AI on the dataset with Bragg peak with a higher threshold\n",
    "ai.type_threshold = .8\n",
    "ai.evaluate()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 8,
   "metadata": {},
   "outputs": [],
   "source": [
    "# save the trained model\n",
    "ai.save(file_name=model_name)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": []
  }
 ],
 "metadata": {
  "kernelspec": {
   "display_name": "Python 3 (ipykernel)",
   "language": "python",
   "name": "python3"
  },
  "language_info": {
   "codemirror_mode": {
    "name": "ipython",
    "version": 3
   },
   "file_extension": ".py",
   "mimetype": "text/x-python",
   "name": "python",
   "nbconvert_exporter": "python",
   "pygments_lexer": "ipython3",
   "version": "3.7.11"
  }
 },
 "nbformat": 4,
 "nbformat_minor": 2
}