train.ipynb

{
  "nbformat": 4,
  "nbformat_minor": 0,
  "metadata": {
    "colab": {
      "name": "Copy of Untitled2 (1).ipynb",
      "provenance": [],
      "collapsed_sections": [],
      "include_colab_link": true
    },
    "kernelspec": {
      "name": "python3",
      "display_name": "Python 3"
    },
    "accelerator": "GPU"
  },
  "cells": [
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "view-in-github",
        "colab_type": "text"
      },
      "source": [
        "<a href=\"https://colab.research.google.com/github/mehuldev/Avishkar-Self-Driving-Car/blob/main/train.ipynb\" target=\"_parent\"><img src=\"https://colab.research.google.com/assets/colab-badge.svg\" alt=\"Open In Colab\"/></a>"
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "Xu0XaIu2DyTY"
      },
      "source": [
        "Mounting the google drive"
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "j_40Q-PaClbw",
        "outputId": "37b745a2-4285-401d-89d1-70fd9619dfaa",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 34
        }
      },
      "source": [
        "from google.colab import drive\n",
        "drive.mount('/content/drive')                                  "
      ],
      "execution_count": 42,
      "outputs": [
        {
          "output_type": "stream",
          "text": [
            "Drive already mounted at /content/drive; to attempt to forcibly remount, call drive.mount(\"/content/drive\", force_remount=True).\n"
          ],
          "name": "stdout"
        }
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "Z1mhGIYRC2Hv"
      },
      "source": [
        "Importing Required Modules"
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "f_3YJ0joFzju"
      },
      "source": [
        "import tensorflow as tf    \n",
        "import pandas as pd\n",
        "from __future__ import division, print_function, absolute_import\n",
        "from tqdm import tqdm\n",
        "from keras.preprocessing import image\n",
        "import numpy as np\n",
        "import scipy.io as io\n",
        "import matplotlib.pyplot as plt\n",
        "from sklearn.model_selection import train_test_split\n",
        "\n",
        "from keras.layers.pooling import MaxPooling2D, AveragePooling2D\n",
        "from keras.layers import Cropping2D\n",
        "from keras.layers.core import Dropout\n",
        "from keras.layers.normalization import BatchNormalization\n",
        "import tensorflow as tf\n",
        "from tensorflow import keras\n",
        "from tensorflow.keras import layers\n",
        "import os\n",
        "from PIL import Image\n"
      ],
      "execution_count": 43,
      "outputs": []
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "c5mr_MKxC24H"
      },
      "source": [
        "Importing the train data"
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "LpnUJB8KC7Us",
        "outputId": "243a85d6-2f3b-44b5-dcee-0ef6cce05e18",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 68
        }
      },
      "source": [
        "data = io.loadmat('/content/drive/My Drive/Colab Notebooks/data')\n",
        "X = []\n",
        "Y = []\n",
        "del data['__header__']\n",
        "del data['__version__']\n",
        "del data['__globals__']\n",
        "\n",
        "for k in tqdm(data.keys()):\n",
        "    img = np.asarray(data[k])\n",
        "    X.append(img)\n",
        "    j = k.find('angle')+6\n",
        "    angle = float(k[j:j+4])\n",
        "    j = k.find('throttle')+9\n",
        "    throttle = float(k[j:j+4])\n",
        "    Y.append([angle,throttle])\n",
        "X = np.array(X)\n",
        "Y = np.array(Y)\n",
        "print(X.shape)\n",
        "assert Y.shape[1] == 2\n",
        "print(Y.shape)"
      ],
      "execution_count": 44,
      "outputs": [
        {
          "output_type": "stream",
          "text": [
            "100%|██████████| 2807/2807 [00:00<00:00, 350086.57it/s]\n"
          ],
          "name": "stderr"
        },
        {
          "output_type": "stream",
          "text": [
            "(2807, 240, 320, 3)\n",
            "(2807, 2)\n"
          ],
          "name": "stdout"
        }
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "rV2vEqo4hhrX"
      },
      "source": [
        "Split data into train data and test data"
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "PIvRmoUvGAYs",
        "outputId": "a1449c76-3bd2-4018-8658-81fede60b3e1",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 34
        }
      },
      "source": [
        "X_train, X_test, Y_train, Y_test = train_test_split(X, Y, random_state=1337, test_size=100, shuffle = True)\n",
        "print(X_train.shape,X_test.shape)"
      ],
      "execution_count": 45,
      "outputs": [
        {
          "output_type": "stream",
          "text": [
            "(2707, 240, 320, 3) (100, 240, 320, 3)\n"
          ],
          "name": "stdout"
        }
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "a79Yz9jQd-Ch"
      },
      "source": [
        "Defining the Model"
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "k1CGwjO7Bmia"
      },
      "source": [
        "Convolution Layers"
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "I5jK3Kq4g3Ed",
        "outputId": "5f7a10b2-dc4e-4744-bddb-58bebef1bc5e",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 425
        }
      },
      "source": [
        "model = keras.Sequential()\n",
        "model.add(keras.Input(shape = (240,320,3)))\n",
        "model.add(layers.experimental.preprocessing.Rescaling(1./255))\n",
        "model.add(layers.Conv2D(24, 5, strides=2, activation=\"elu\", use_bias=False))\n",
        "model.add(layers.Conv2D(36,5, strides=2,activation=\"elu\", use_bias=False))\n",
        "model.add(layers.Conv2D(48,5, strides=2,activation=\"elu\",use_bias=False))\n",
        "model.add(layers.Conv2D(64,5, strides=2,activation=\"elu\",use_bias= False))\n",
        "model.add(layers.Conv2D(64,5, strides=2,activation=\"elu\", use_bias=False))\n",
        "model.add(MaxPooling2D(4, strides=4))\n",
        "model.add(layers.Dropout(.4))\n",
        "model.summary()"
      ],
      "execution_count": 74,
      "outputs": [
        {
          "output_type": "stream",
          "text": [
            "Model: \"sequential_11\"\n",
            "_________________________________________________________________\n",
            "Layer (type)                 Output Shape              Param #   \n",
            "=================================================================\n",
            "rescaling_9 (Rescaling)      (None, 240, 320, 3)       0         \n",
            "_________________________________________________________________\n",
            "conv2d_55 (Conv2D)           (None, 118, 158, 24)      1800      \n",
            "_________________________________________________________________\n",
            "conv2d_56 (Conv2D)           (None, 57, 77, 36)        21600     \n",
            "_________________________________________________________________\n",
            "conv2d_57 (Conv2D)           (None, 27, 37, 48)        43200     \n",
            "_________________________________________________________________\n",
            "conv2d_58 (Conv2D)           (None, 12, 17, 64)        76800     \n",
            "_________________________________________________________________\n",
            "conv2d_59 (Conv2D)           (None, 4, 7, 64)          102400    \n",
            "_________________________________________________________________\n",
            "max_pooling2d (MaxPooling2D) (None, 1, 1, 64)          0         \n",
            "_________________________________________________________________\n",
            "dropout_7 (Dropout)          (None, 1, 1, 64)          0         \n",
            "=================================================================\n",
            "Total params: 245,800\n",
            "Trainable params: 245,800\n",
            "Non-trainable params: 0\n",
            "_________________________________________________________________\n"
          ],
          "name": "stdout"
        }
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "_QDvFK1aBqrT"
      },
      "source": [
        "Linear Layers"
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "GT_dnF_5d8-u",
        "outputId": "fdd176fc-2870-4332-ec23-a927b7754786",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 595
        }
      },
      "source": [
        "model.add(Flatten())\n",
        "model.add(layers.Dense(100,activation='elu',use_bias=False, input_shape = (64*1*18,)))\n",
        "model.add(layers.Dense(50,activation='elu',use_bias=False, input_shape = (100,)))\n",
        "model.add(layers.Dense(10,use_bias=False, input_shape = (50,)))\n",
        "model.add(layers.Dense(2,use_bias=False, input_shape = (10,)))\n",
        "model.summary()"
      ],
      "execution_count": 75,
      "outputs": [
        {
          "output_type": "stream",
          "text": [
            "Model: \"sequential_11\"\n",
            "_________________________________________________________________\n",
            "Layer (type)                 Output Shape              Param #   \n",
            "=================================================================\n",
            "rescaling_9 (Rescaling)      (None, 240, 320, 3)       0         \n",
            "_________________________________________________________________\n",
            "conv2d_55 (Conv2D)           (None, 118, 158, 24)      1800      \n",
            "_________________________________________________________________\n",
            "conv2d_56 (Conv2D)           (None, 57, 77, 36)        21600     \n",
            "_________________________________________________________________\n",
            "conv2d_57 (Conv2D)           (None, 27, 37, 48)        43200     \n",
            "_________________________________________________________________\n",
            "conv2d_58 (Conv2D)           (None, 12, 17, 64)        76800     \n",
            "_________________________________________________________________\n",
            "conv2d_59 (Conv2D)           (None, 4, 7, 64)          102400    \n",
            "_________________________________________________________________\n",
            "max_pooling2d (MaxPooling2D) (None, 1, 1, 64)          0         \n",
            "_________________________________________________________________\n",
            "dropout_7 (Dropout)          (None, 1, 1, 64)          0         \n",
            "_________________________________________________________________\n",
            "flatten_6 (Flatten)          (None, 64)                0         \n",
            "_________________________________________________________________\n",
            "dense_24 (Dense)             (None, 100)               6400      \n",
            "_________________________________________________________________\n",
            "dense_25 (Dense)             (None, 50)                5000      \n",
            "_________________________________________________________________\n",
            "dense_26 (Dense)             (None, 10)                500       \n",
            "_________________________________________________________________\n",
            "dense_27 (Dense)             (None, 2)                 20        \n",
            "=================================================================\n",
            "Total params: 257,720\n",
            "Trainable params: 257,720\n",
            "Non-trainable params: 0\n",
            "_________________________________________________________________\n"
          ],
          "name": "stdout"
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "HXcSMcS2eNfo"
      },
      "source": [
        "tf.debugging.set_log_device_placement(True)\n",
        "model.compile(optimizer=tf.keras.optimizers.Adam(learning_rate=0.0001), \n",
        "              loss=tf.keras.losses.MeanSquaredError(), metrics=[tf.keras.metrics.Accuracy()])  \n",
        "num_epochs = 50"
      ],
      "execution_count": 76,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "xuz0wT2uA41D",
        "outputId": "b2ec3eb3-4c43-4255-b4a5-1dbf2603b009",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 1000
        }
      },
      "source": [
        "hist=model.fit(X_train, Y_train, epochs=num_epochs, validation_data=(X_test, Y_test),batch_size=32, shuffle=True)"
      ],
      "execution_count": 77,
      "outputs": [
        {
          "output_type": "stream",
          "text": [
            "Epoch 1/50\n",
            "85/85 [==============================] - 3s 30ms/step - loss: 0.2629 - accuracy: 0.0000e+00 - val_loss: 0.2015 - val_accuracy: 0.0000e+00\n",
            "Epoch 2/50\n",
            "85/85 [==============================] - 2s 28ms/step - loss: 0.2058 - accuracy: 0.0000e+00 - val_loss: 0.1688 - val_accuracy: 0.0000e+00\n",
            "Epoch 3/50\n",
            "85/85 [==============================] - 2s 28ms/step - loss: 0.1699 - accuracy: 0.0000e+00 - val_loss: 0.1413 - val_accuracy: 0.0000e+00\n",
            "Epoch 4/50\n",
            "85/85 [==============================] - 2s 28ms/step - loss: 0.1428 - accuracy: 0.0000e+00 - val_loss: 0.1159 - val_accuracy: 0.0000e+00\n",
            "Epoch 5/50\n",
            "85/85 [==============================] - 2s 28ms/step - loss: 0.1231 - accuracy: 0.0000e+00 - val_loss: 0.1079 - val_accuracy: 0.0000e+00\n",
            "Epoch 6/50\n",
            "85/85 [==============================] - 2s 28ms/step - loss: 0.1116 - accuracy: 0.0000e+00 - val_loss: 0.0981 - val_accuracy: 0.0000e+00\n",
            "Epoch 7/50\n",
            "85/85 [==============================] - 2s 28ms/step - loss: 0.1003 - accuracy: 0.0000e+00 - val_loss: 0.0962 - val_accuracy: 0.0000e+00\n",
            "Epoch 8/50\n",
            "85/85 [==============================] - 2s 28ms/step - loss: 0.0910 - accuracy: 0.0000e+00 - val_loss: 0.0880 - val_accuracy: 0.0000e+00\n",
            "Epoch 9/50\n",
            "85/85 [==============================] - 2s 28ms/step - loss: 0.0835 - accuracy: 0.0000e+00 - val_loss: 0.0811 - val_accuracy: 0.0000e+00\n",
            "Epoch 10/50\n",
            "85/85 [==============================] - 2s 28ms/step - loss: 0.0794 - accuracy: 0.0000e+00 - val_loss: 0.0740 - val_accuracy: 0.0000e+00\n",
            "Epoch 11/50\n",
            "85/85 [==============================] - 2s 28ms/step - loss: 0.0736 - accuracy: 0.0000e+00 - val_loss: 0.0740 - val_accuracy: 0.0000e+00\n",
            "Epoch 12/50\n",
            "85/85 [==============================] - 2s 28ms/step - loss: 0.0667 - accuracy: 0.0000e+00 - val_loss: 0.0709 - val_accuracy: 0.0000e+00\n",
            "Epoch 13/50\n",
            "85/85 [==============================] - 2s 28ms/step - loss: 0.0628 - accuracy: 0.0000e+00 - val_loss: 0.0699 - val_accuracy: 0.0000e+00\n",
            "Epoch 14/50\n",
            "85/85 [==============================] - 2s 28ms/step - loss: 0.0604 - accuracy: 0.0000e+00 - val_loss: 0.0716 - val_accuracy: 0.0000e+00\n",
            "Epoch 15/50\n",
            "85/85 [==============================] - 2s 28ms/step - loss: 0.0577 - accuracy: 0.0000e+00 - val_loss: 0.0711 - val_accuracy: 0.0000e+00\n",
            "Epoch 16/50\n",
            "85/85 [==============================] - 2s 28ms/step - loss: 0.0528 - accuracy: 0.0000e+00 - val_loss: 0.0690 - val_accuracy: 0.0000e+00\n",
            "Epoch 17/50\n",
            "85/85 [==============================] - 2s 28ms/step - loss: 0.0485 - accuracy: 0.0000e+00 - val_loss: 0.0619 - val_accuracy: 0.0000e+00\n",
            "Epoch 18/50\n",
            "85/85 [==============================] - 2s 28ms/step - loss: 0.0499 - accuracy: 0.0000e+00 - val_loss: 0.0584 - val_accuracy: 0.0000e+00\n",
            "Epoch 19/50\n",
            "85/85 [==============================] - 2s 28ms/step - loss: 0.0449 - accuracy: 0.0000e+00 - val_loss: 0.0570 - val_accuracy: 0.0000e+00\n",
            "Epoch 20/50\n",
            "85/85 [==============================] - 2s 28ms/step - loss: 0.0456 - accuracy: 0.0000e+00 - val_loss: 0.0567 - val_accuracy: 0.0000e+00\n",
            "Epoch 21/50\n",
            "85/85 [==============================] - 2s 27ms/step - loss: 0.0438 - accuracy: 0.0000e+00 - val_loss: 0.0564 - val_accuracy: 0.0000e+00\n",
            "Epoch 22/50\n",
            "85/85 [==============================] - 2s 27ms/step - loss: 0.0399 - accuracy: 0.0000e+00 - val_loss: 0.0568 - val_accuracy: 0.0000e+00\n",
            "Epoch 23/50\n",
            "85/85 [==============================] - 2s 27ms/step - loss: 0.0388 - accuracy: 0.0000e+00 - val_loss: 0.0576 - val_accuracy: 0.0000e+00\n",
            "Epoch 24/50\n",
            "85/85 [==============================] - 2s 27ms/step - loss: 0.0384 - accuracy: 0.0000e+00 - val_loss: 0.0557 - val_accuracy: 0.0000e+00\n",
            "Epoch 25/50\n",
            "85/85 [==============================] - 2s 27ms/step - loss: 0.0378 - accuracy: 0.0000e+00 - val_loss: 0.0578 - val_accuracy: 0.0000e+00\n",
            "Epoch 26/50\n",
            "85/85 [==============================] - 2s 27ms/step - loss: 0.0364 - accuracy: 0.0000e+00 - val_loss: 0.0538 - val_accuracy: 0.0000e+00\n",
            "Epoch 27/50\n",
            "85/85 [==============================] - 2s 27ms/step - loss: 0.0359 - accuracy: 0.0000e+00 - val_loss: 0.0533 - val_accuracy: 0.0000e+00\n",
            "Epoch 28/50\n",
            "85/85 [==============================] - 2s 27ms/step - loss: 0.0356 - accuracy: 0.0000e+00 - val_loss: 0.0529 - val_accuracy: 0.0000e+00\n",
            "Epoch 29/50\n",
            "85/85 [==============================] - 2s 27ms/step - loss: 0.0328 - accuracy: 0.0000e+00 - val_loss: 0.0589 - val_accuracy: 0.0000e+00\n",
            "Epoch 30/50\n",
            "85/85 [==============================] - 2s 28ms/step - loss: 0.0321 - accuracy: 0.0000e+00 - val_loss: 0.0518 - val_accuracy: 0.0000e+00\n",
            "Epoch 31/50\n",
            "85/85 [==============================] - 2s 27ms/step - loss: 0.0305 - accuracy: 0.0000e+00 - val_loss: 0.0523 - val_accuracy: 0.0000e+00\n",
            "Epoch 32/50\n",
            "85/85 [==============================] - 2s 27ms/step - loss: 0.0297 - accuracy: 0.0000e+00 - val_loss: 0.0549 - val_accuracy: 0.0000e+00\n",
            "Epoch 33/50\n",
            "85/85 [==============================] - 2s 27ms/step - loss: 0.0288 - accuracy: 0.0000e+00 - val_loss: 0.0480 - val_accuracy: 0.0000e+00\n",
            "Epoch 34/50\n",
            "85/85 [==============================] - 2s 28ms/step - loss: 0.0283 - accuracy: 0.0000e+00 - val_loss: 0.0526 - val_accuracy: 0.0000e+00\n",
            "Epoch 35/50\n",
            "85/85 [==============================] - 2s 27ms/step - loss: 0.0280 - accuracy: 0.0000e+00 - val_loss: 0.0541 - val_accuracy: 0.0000e+00\n",
            "Epoch 36/50\n",
            "85/85 [==============================] - 2s 27ms/step - loss: 0.0281 - accuracy: 0.0000e+00 - val_loss: 0.0529 - val_accuracy: 0.0000e+00\n",
            "Epoch 37/50\n",
            "85/85 [==============================] - 2s 28ms/step - loss: 0.0266 - accuracy: 0.0000e+00 - val_loss: 0.0548 - val_accuracy: 0.0000e+00\n",
            "Epoch 38/50\n",
            "85/85 [==============================] - 2s 28ms/step - loss: 0.0268 - accuracy: 0.0000e+00 - val_loss: 0.0539 - val_accuracy: 0.0000e+00\n",
            "Epoch 39/50\n",
            "85/85 [==============================] - 2s 28ms/step - loss: 0.0262 - accuracy: 0.0000e+00 - val_loss: 0.0515 - val_accuracy: 0.0000e+00\n",
            "Epoch 40/50\n",
            "85/85 [==============================] - 2s 27ms/step - loss: 0.0254 - accuracy: 0.0000e+00 - val_loss: 0.0545 - val_accuracy: 0.0000e+00\n",
            "Epoch 41/50\n",
            "85/85 [==============================] - 2s 28ms/step - loss: 0.0246 - accuracy: 0.0000e+00 - val_loss: 0.0574 - val_accuracy: 0.0000e+00\n",
            "Epoch 42/50\n",
            "85/85 [==============================] - 2s 28ms/step - loss: 0.0255 - accuracy: 0.0000e+00 - val_loss: 0.0507 - val_accuracy: 0.0000e+00\n",
            "Epoch 43/50\n",
            "85/85 [==============================] - 2s 28ms/step - loss: 0.0235 - accuracy: 0.0000e+00 - val_loss: 0.0508 - val_accuracy: 0.0000e+00\n",
            "Epoch 44/50\n",
            "85/85 [==============================] - 2s 28ms/step - loss: 0.0229 - accuracy: 0.0000e+00 - val_loss: 0.0524 - val_accuracy: 0.0000e+00\n",
            "Epoch 45/50\n",
            "85/85 [==============================] - 2s 28ms/step - loss: 0.0233 - accuracy: 0.0000e+00 - val_loss: 0.0490 - val_accuracy: 0.0000e+00\n",
            "Epoch 46/50\n",
            "85/85 [==============================] - 2s 28ms/step - loss: 0.0230 - accuracy: 0.0000e+00 - val_loss: 0.0514 - val_accuracy: 0.0000e+00\n",
            "Epoch 47/50\n",
            "85/85 [==============================] - 2s 27ms/step - loss: 0.0230 - accuracy: 0.0000e+00 - val_loss: 0.0537 - val_accuracy: 0.0000e+00\n",
            "Epoch 48/50\n",
            "85/85 [==============================] - 2s 28ms/step - loss: 0.0224 - accuracy: 0.0000e+00 - val_loss: 0.0508 - val_accuracy: 0.0000e+00\n",
            "Epoch 49/50\n",
            "85/85 [==============================] - 2s 27ms/step - loss: 0.0209 - accuracy: 0.0000e+00 - val_loss: 0.0517 - val_accuracy: 0.0000e+00\n",
            "Epoch 50/50\n",
            "85/85 [==============================] - 2s 27ms/step - loss: 0.0215 - accuracy: 0.0000e+00 - val_loss: 0.0514 - val_accuracy: 0.0000e+00\n"
          ],
          "name": "stdout"
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "dvxgisQJfZ42",
        "outputId": "a5d019db-954f-4bbc-ebd6-f6508cb30d5d",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 332
        }
      },
      "source": [
        "print(hist.history)\n",
        "\n",
        "plt.plot(hist.history['loss'])\n",
        "plt.plot(hist.history['val_loss'])\n",
        "plt.title('Model loss')\n",
        "plt.ylabel('Loss')\n",
        "plt.xlabel('Epoch')\n",
        "plt.legend(['Train', 'Val'], loc='upper right')\n",
        "plt.show()\n",
        "plt.pause(5.)"
      ],
      "execution_count": 78,
      "outputs": [
        {
          "output_type": "stream",
          "text": [
            "{'loss': [0.26289284229278564, 0.20583052933216095, 0.16993831098079681, 0.1427505612373352, 0.1230955645442009, 0.11159228533506393, 0.10031221807003021, 0.09100194275379181, 0.08353665471076965, 0.07940086722373962, 0.07362554222345352, 0.06665235757827759, 0.0627659484744072, 0.06037740409374237, 0.057698141783475876, 0.0527561716735363, 0.048519477248191833, 0.049908969551324844, 0.04493633657693863, 0.04557623714208603, 0.04379918798804283, 0.03990194946527481, 0.038826797157526016, 0.0383857823908329, 0.03781187906861305, 0.036439381539821625, 0.03590894117951393, 0.035552021116018295, 0.03284377232193947, 0.03205110505223274, 0.030464448034763336, 0.029689276590943336, 0.028750646859407425, 0.028305772691965103, 0.027964036911725998, 0.028084471821784973, 0.026565594598650932, 0.02679612673819065, 0.026203975081443787, 0.025436213240027428, 0.024639295414090157, 0.025498056784272194, 0.02349049039185047, 0.022948389872908592, 0.023323001340031624, 0.023005113005638123, 0.023009780794382095, 0.022447455674409866, 0.0208725705742836, 0.021450461819767952], 'accuracy': [0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0], 'val_loss': [0.20153270661830902, 0.16881494224071503, 0.14133933186531067, 0.11587663739919662, 0.10785558819770813, 0.0981455072760582, 0.09618759155273438, 0.08802375942468643, 0.08105769753456116, 0.07398543506860733, 0.0740441158413887, 0.0709025114774704, 0.06990296393632889, 0.07159950584173203, 0.07111496478319168, 0.06904042512178421, 0.061913322657346725, 0.05835916846990585, 0.05697745457291603, 0.0566856674849987, 0.05635274574160576, 0.05680384114384651, 0.05764758959412575, 0.055695708841085434, 0.057805780321359634, 0.053808312863111496, 0.053299274295568466, 0.05293015018105507, 0.05887984856963158, 0.05181821808218956, 0.05226382240653038, 0.05487225949764252, 0.04804747551679611, 0.052572473883628845, 0.054085418581962585, 0.05288757383823395, 0.05480743944644928, 0.053905587643384933, 0.05146576464176178, 0.05448545143008232, 0.057397861033678055, 0.05070675536990166, 0.05084472522139549, 0.05244624987244606, 0.04900689050555229, 0.05135035887360573, 0.05367494747042656, 0.05077317729592323, 0.05167418345808983, 0.05138850584626198], 'val_accuracy': [0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0]}\n"
          ],
          "name": "stdout"
        },
        {
          "output_type": "display_data",
          "data": {
            "image/png": "iVBORw0KGgoAAAANSUhEUgAAAYgAAAEWCAYAAAB8LwAVAAAABHNCSVQICAgIfAhkiAAAAAlwSFlzAAALEgAACxIB0t1+/AAAADh0RVh0U29mdHdhcmUAbWF0cGxvdGxpYiB2ZXJzaW9uMy4yLjIsIGh0dHA6Ly9tYXRwbG90bGliLm9yZy+WH4yJAAAgAElEQVR4nO3dd3hc5Zn38e89o1Hvstwk2XIT7gWE6cZgwKYaiCkGElggWdhkU1kC7CYkpEJIKLvZd5MQWgrGoZNgjHHoprhgG/cqW5KbZFm9a+73j3Nkj6WRLdsajaS5P9c118ycMuc5RsxvnnKeI6qKMcYY05Yn3AUwxhjTM1lAGGOMCcoCwhhjTFAWEMYYY4KygDDGGBOUBYQxxpigLCCMOQEikisiKiJRndj2FhH58EQ/x5juYgFhIoaIFIhIo4j0a7P8c/fLOTc8JTOmZ7KAMJFmOzC39Y2ITADiw1ccY3ouCwgTaf4EfCXg/c3As4EbiEiKiDwrIiUiskNE/ktEPO46r4g8LCKlIrINuDTIvn8Ukd0iUiwiPxUR77EWUkQGi8hrIlImIltE5KsB66aKyDIRqRSRvSLyG3d5rIj8WUT2i0i5iCwVkQHHemxjWllAmEjzCZAsImPcL+7rgT+32ea/gRRgOHAuTqD8i7vuq8BlwBQgH5jTZt+ngWZgpLvNRcDtx1HOeUARMNg9xs9F5Hx33WPAY6qaDIwA5rvLb3bLnQNkAHcAdcdxbGMACwgTmVprERcC64Hi1hUBoXGvqlapagHwa+DL7ibXAo+qaqGqlgG/CNh3AHAJ8G1VrVHVfcAj7ud1mojkAGcB31fVelVdCTzBoZpPEzBSRPqparWqfhKwPAMYqaotqrpcVSuP5djGBLKAMJHoT8ANwC20aV4C+gE+YEfAsh1Alvt6MFDYZl2roe6+u90mnnLgd0D/YyzfYKBMVas6KMNtQB6wwW1GuizgvBYC80Rkl4g8JCK+Yzy2MQdZQJiIo6o7cDqrLwFearO6FOeX+NCAZUM4VMvYjdOEE7iuVSHQAPRT1VT3kayq446xiLuAdBFJClYGVd2sqnNxgudB4AURSVDVJlX9saqOBc7EaQr7CsYcJwsIE6luA85X1ZrAharagtOm/zMRSRKRocB3OdRPMR/4pohki0gacE/AvruBt4Bfi0iyiHhEZISInHssBVPVQmAJ8Au343miW94/A4jITSKSqap+oNzdzS8i54nIBLeZrBIn6PzHcmxjAllAmIikqltVdVkHq/8dqAG2AR8CfwWedNf9AacZZxWwgvY1kK8A0cA64ADwAjDoOIo4F8jFqU28DNyvqm+762YBa0WkGqfD+npVrQMGuserxOlbeQ+n2cmY4yJ2wyBjjDHBWA3CGGNMUBYQxhhjgrKAMMYYE5QFhDHGmKD6zNTC/fr109zc3HAXwxhjepXly5eXqmpmsHV9JiByc3NZtqyjUYvGGGOCEZEdHa2zJiZjjDFBWUAYY4wJygLCGGNMUH2mD8IYY45VU1MTRUVF1NfXh7soIRcbG0t2djY+X+cn+LWAMMZErKKiIpKSksjNzUVEwl2ckFFV9u/fT1FREcOGDev0ftbEZIyJWPX19WRkZPTpcAAQETIyMo65pmQBYYyJaH09HFodz3lGfEBU1Dbx2NubWVVYfvSNjTEmgkR8QIgHHnl7E59s2x/uohhjIsz+/fuZPHkykydPZuDAgWRlZR1839jYeMR9ly1bxje/+c2Qli/iO6mTY32kxvvYWVYb7qIYYyJMRkYGK1euBOBHP/oRiYmJ3HXXXQfXNzc3ExUV/Gs6Pz+f/Pz8kJYv4msQAEPS4y0gjDE9wi233MIdd9zBaaedxt13381nn33GGWecwZQpUzjzzDPZuHEjAO+++y6XXXYZ4ITLrbfeyvTp0xk+fDiPP/54l5Ql4msQ4ATEmuKKcBfDGBNGP359Let2VXbpZ44dnMz9l4875v2KiopYsmQJXq+XyspKPvjgA6Kionj77be57777ePHFF9vts2HDBt555x2qqqo46aSTuPPOO4/pmodgLCBwAuLNNXtobvET5bVKlTEmvK655hq8Xi8AFRUV3HzzzWzevBkRoampKeg+l156KTExMcTExNC/f3/27t1Ldnb2CZXDAgInIJr9yu6KenLS48NdHGNMGBzPL/1QSUhIOPj6Bz/4Aeeddx4vv/wyBQUFTJ8+Peg+MTExB197vV6am5tPuBz2cxknIAAKrR/CGNPDVFRUkJWVBcDTTz/drce2gICDtQbrqDbG9DR333039957L1OmTOmSWsGxEFUN3YeLzAIeA7zAE6r6yzbrvwvcDjQDJcCtqrrDXdcCfOFuulNVrzjSsfLz8/V4bxjU3OJn9A/e5GvThnP3rNHH9RnGmN5n/fr1jBkzJtzF6DbBzldElqtq0PGyIeuDEBEv8FvgQqAIWCoir6nquoDNPgfyVbVWRO4EHgKuc9fVqerkUJUvUJTXQ1ZanNUgjDEmQCibmKYCW1R1m6o2AvOA2YEbqOo7qtr6rfwJcGJd7idgSHq89UEYY0yAUAZEFlAY8L7IXdaR24AFAe9jRWSZiHwiIlcG20FEvuZus6ykpOSECmsXyxljzOF6xDBXEbkJyAfODVg8VFWLRWQ48E8R+UJVtwbup6q/B34PTh/EiZRhSHo8B2qbqKxvIjn2xC4uMcaYviCUNYhiICfgfba77DAicgHwn8AVqtrQulxVi93nbcC7wJQQlvXgUNed+60WYYwxENqAWAqMEpFhIhINXA+8FriBiEwBfocTDvsClqeJSIz7uh9wFhDYud3lcuxaCGOMOUzIAkJVm4FvAAuB9cB8VV0rIg+ISOuQ1V8BicDfRGSliLQGyBhgmYisAt4Bftlm9FOXG5Jh10IYY7rXeeedx8KFCw9b9uijj3LnnXcG3X769Okc73D+4xHSPghVfQN4o82yHwa8vqCD/ZYAE0JZtrZs2m9jTHebO3cu8+bNY+bMmQeXzZs3j4ceeiiMpTrErqQOYCOZjDHdac6cOfzjH/84eHOggoICdu3axXPPPUd+fj7jxo3j/vvvD1v5esQopp7Cpv02JoItuAf2fHH07Y7FwAlw8S87XJ2ens7UqVNZsGABs2fPZt68eVx77bXcd999pKen09LSwowZM1i9ejUTJ07s2rJ1gtUgAgxJj6foQB0t/tBNP2KMMYFam5nAaV6aO3cu8+fP5+STT2bKlCmsXbuWdetC2gXbIatBBGid9ntXeZ1N+21MpDnCL/1Qmj17Nt/5zndYsWIFtbW1pKen8/DDD7N06VLS0tK45ZZbqK+vD0vZrAYRwKb9NsZ0t8TERM477zxuvfVW5s6dS2VlJQkJCaSkpLB3714WLFhw9A8JEatBBAic9vvMMJfFGBM55s6dy1VXXcW8efMYPXo0U6ZMYfTo0eTk5HDWWWeFrVwWEAEGpcQS5REbyWSM6VZXXnklgbde6OjGQO+++273FMhlTUwBbNpvY4w5xAKiDZv22xhjHBYQbdjFcsZEllDeVbMnOZ7ztIBoI3Dab2NM3xYbG8v+/fv7fEioKvv37yc2NvaY9rNO6jYCp/0en5US5tIYY0IpOzuboqIiTvSGY71BbGws2dnHdtNOC4g2Aqf9toAwpm/z+XwMGzYs3MXosayJqQ2b9tsYYxwWEG3YtN/GGOOwgAjCRjIZY4wFRFA5di2EMcZYQAQz1Kb9NsYYC4hgAqf9NsaYSGUBEYRN+22MMRYQQQVO+22MMZHKAiIIm/bbGGMsIIKyab+NMcYCokM27bcxJtJZQHTALpYzxkQ6C4gO2LTfxphIZwHRgcBpv40xJhJZQHQgx66FMMZEOAuIDti038aYSGcB0YHWab8LrInJGBOhLCCO4KQBSazbXRnuYhhjTFiENCBEZJaIbBSRLSJyT5D13xWRdSKyWkQWi8jQgHU3i8hm93FzKMvZkck5qazfVUlDc0s4Dm+MMWEVsoAQES/wW+BiYCwwV0TGttnscyBfVScCLwAPufumA/cDpwFTgftFJC0kBW2ogpV/hdLN7VZNykmlscXPht1VITm0Mcb0ZKGsQUwFtqjqNlVtBOYBswM3UNV3VLW1kf8TINt9PRNYpKplqnoAWATMCkkpmxvhlTth/WvtVk3OSQVgVVF5SA5tjDE9WSgDIgsoDHhf5C7ryG3AguPc9/glZEC/PNj5abtVg1JiyUyKYeVOCwhjTOTpEZ3UInITkA/86hj3+5qILBORZSUlJcdfgCGnQ+En4Pe3/XwmZaey0moQxpgIFMqAKAZyAt5nu8sOIyIXAP8JXKGqDceyr6r+XlXzVTU/MzPz+EuaczrUV0DpxnarJueksK2khoo6m3LDGBNZQhkQS4FRIjJMRKKB64HDGvpFZArwO5xw2BewaiFwkYikuZ3TF7nLQmPI6c7zzk/arZrk9kN8UVQRssMbY0xPFLKAUNVm4Bs4X+zrgfmqulZEHhCRK9zNfgUkAn8TkZUi8pq7bxnwE5yQWQo84C4LjfThkJAZNCAmZjsBsbLwQMgOb4wxPVFUKD9cVd8A3miz7IcBry84wr5PAk+GrnQBRA71Q7SREudjeGYCKwutBmGMiSw9opO6R8g5HQ4UQNWedqsmZ6eysrAcVe3+chljTJhYQLQacobz3EE/RGl1A7sr6ru5UMYYEz4WEK0GTYSouKABcfCCuUIb7mqMiRwWEK28PsjOD9oPMXpQEtFeDystIIwxEcQCIlDOabB7NTRUH7Y4JsrLmMHJFhDGmIhiARFoyBmgLVC8vN2qydkpfFFcQYvfOqqNMZHBAiJQzqmABO+HGJJKbWMLW/ZVt9/PGGP6IAuIQLEpMGBc0H6ISXbBnDEmwlhAtJVzGhQuBf/hNwnKzUggOTbKLpgzxkQMC4i2hpwBjVWwd+1hiz0eYVJOqg11NcZEDAuItoac5jx3cD3Exr1V1DXaLUiNMX2fBURbKTmQnNVhP0SLX1mzy5qZjDF9nwVEWyJOP0SwmV1zUgC7otoYExksIIIZcgZUFkN54WGL+yfFkpUaZxfMGWMiggVEMEfph1hltyA1xkQAC4hg+o+D6KTg/RA5KRSW1bG/uiHIjsYY03dYQATjjXIm7gs29bd7wZzVIowxfZ0FREeGnO5cC1F/+Iil8VkpeAS7YM4Y0+dZQHRkyOmAOldVB0iIiSJvQJJ1VBtj+jwLiI5k5YN4YefH7VadPDSNFTsO0NTiD0PBjDGme1hAdCQmEQZPhu3vt1s1bVQm1Q3NrNhhE/cZY/ouC4gjGXE+FC+DusObk84cmUGUR3hvU0mYCmaMMaFnAXEkI2aA+mH7e4ctTo71cfKQNN7fbAFhjOm7LCCOJDsfYpJhy+J2q849KZM1xZWUVNn1EMaYvskC4ki8Phg2Dbb+E/TwW42em5cJwIdbrBZhjOmbLCCOZuQMqCiE0s2HLR47KJmMhGje22gBYYzpmywgjmbEDOd56+HNTB6PMC0vk/c3l+L3a5AdjTGmd7OAOJq0oZAxMmg/xLS8fpTVNLJ2V2UYCmaMMaFlAdEZI86Hgg+hqf6wxeeMcvoh3tu0LxylMsaYkLKA6IwRM6C5rt1V1f0SY5iQlWLXQxhj+iQLiM7IPRs8Pmc0UxvT8vqxYmc5lfVNYSiYMcaEjgVEZ8QkOpP3BQmIc/P60+JXlmwpDUPBjDEmdEIaECIyS0Q2isgWEbknyPppIrJCRJpFZE6bdS0istJ9vBbKcnbKyBmwdw1U7Tls8ZQhqSTGRPHeJgsIY0zfErKAEBEv8FvgYmAsMFdExrbZbCdwC/DXIB9Rp6qT3ccVoSpnpx0c7np4LcLn9XDWyAze31SCqg13Ncb0HZ0KCBFJEBGP+zpPRK4QEd9RdpsKbFHVbaraCMwDZgduoKoFqroa6PnzZg8YDwn9g0+7kdef4vI6tpZUh6FgxhgTGp2tQbwPxIpIFvAW8GXg6aPskwUUBrwvcpd1VqyILBORT0TkymAbiMjX3G2WlZSEeCSRx+MMd932DvgPz7Npef0ArJnJGNOndDYgRFVrgauB/1XVa4BxoSsWAENVNR+4AXhUREa03UBVf6+q+aqan5mZGeLi4ARE7X7YvfKwxdlp8YzITLDhrsaYPqXTASEiZwA3Av9wl3mPsk8xkBPwPttd1imqWuw+bwPeBaZ0dt+QGXG+87w1eDPTp9v2U9/U0s2FMsaY0OhsQHwbuBd4WVXXishw4J2j7LMUGCUiw0QkGrge6NRoJBFJE5EY93U/4CxgXSfLGjqJmTBwImxtf+rT8vrR0Ozn0+1lYSiYMcZ0vU4FhKq+p6pXqOqDbmd1qap+8yj7NAPfABYC64H5brg8ICJXAIjIqSJSBFwD/E5E1rq7jwGWicgqnCD6paqGPyDAGe5a+CnUHz7/0unDM4iJ8tjsrsaYPiOqMxuJyF+BO4AWnJpBsog8pqq/OtJ+qvoG8EabZT8MeL0Up+mp7X5LgAmdKVu3GzEDPnwECj6A0ZceXBzr8zJ1WLrdZc4Y02d0tolprKpWAlcCC4BhOCOZIk/OaRCdGHS463kn9WfLvmq27KsKQ8GMMaZrdTYgfO51D1cCr6lqExCZV4VFRUPuObBlUbu7zF0+aTBRHmH+sqIwFc4YY7pOZwPid0ABkAC8LyJDgci9CULeRVC+E0o2HrY4MymG80f356UVRTS19Pxr/4wx5kg620n9uKpmqeol6tgBnBfisvVcebOc500L2q267tQcSqsbWbze7hFhjOndOjvVRoqI/Kb1qmUR+TVObSIyJQ92hrtuWthu1bl5mQxIjmH+ssIgOxpjTO/R2SamJ4Eq4Fr3UQk8FapC9QonXewMd609/LqHKK+HOadk8+7GfeypqO9gZ2OM6fk6GxAjVPV+d+K9bar6Y2B4KAvW4+XNBPXD5kXtVl2bn4Nf4YXlVoswxvRenQ2IOhE5u/WNiJwF1IWmSL3EoCmQOAA2vdlu1dCMBE4fns78ZUX4/ZE52MsY0/t1NiDuAH4rIgUiUgD8D/CvIStVb+DxwKiLnOshWtrfbvS6U3PYWVbLJ9v3h6Fwxhhz4jo7immVqk4CJgITVXUKcH5IS9Yb5M2ChgrY+XG7VRePH0RSbBTzl1ozkzGmdzqmO8qpaqV7RTXAd0NQnt5l+HTwxsDG9s1MsT4vsycPZsGaPVTUta9hGGNMT3citxyVLitFbxWTCMPOCdoPAXBd/hAamv28trLTs5wbY0yPcSIBYb2v4DQzlW2F0s3tVo3PSmbMoGSet2sijDG90BEDQkSqRKQyyKMKGNxNZezZ8mY6z0FqESLCdfnZrCmuZO2uim4umDHGnJgjBoSqJqlqcpBHkqp2aqrwPi91CAwYH/SqaoArp2QRHeWxzmpjTK9zIk1MplXeTNixBOoOtFuVGh/NrHEDeWXlLrsdqTGmV7GA6Ap5F4O2BL1HBMD1U3OoqGviqY8KurdcxhhzAiwgukLWyRDfr8PRTGcMz2DmuAE88vYmtuyr7ubCGWPM8bGA6Aoer9PMtHkRtDS3Wy0i/OTK8cRHe/mPF1bRYtNvGGN6AQuIrpI3E+rLnRleg+ifFMv9l4/l853lPPXR9m4unDHGHDsLiK4y4nzw+DpsZgK4cnIWF4zpz68WbmR7aU03Fs4YY46dBURXiUmC3LOdgNDgTUgiws+umkBMlIe7X1hlM70aY3o0C4iuNO4qKN0E7/6iw00GJMfyg8vGsrTgAM98XNBtRTPGmGNlAdGVpnwZptwE7z0IHz3e4WZzTslm+kmZPPTmRnbst6YmY0zPZAHRlTweuPxxGHc1LPoBLP1j0M1EhF9cPYEoj/D9F1dbU5MxpkeygOhqHi9c/XtnEr9/fA9WPR90s0EpcfznpWP4ZFsZf/l0RzcX0hhjjs4CIhS8PrjmaafT+pU7Yf3rQTe77tQczhnVj5+/sYGd+2u7t4zGGHMUFhCh4ouDufOcq6xfuDXoNBwiwoNfmkiUR7jLRjUZY3oYC4hQikmEG/8G/U6CeTdC8fJ2mwxOjeOHl4/ls+1lPL2koPvLaIwxHbCACLW4NPjyy5CQCc9/BWr2t9tkzinZzBjdn4cWbmBbic3VZIzpGSwgukNiJlz3LNSUwIu3gv/wab9bRzXFRHm56282V5MxpmewgOgug6fApb+Gbe/CP3/abnX/5FgemD2OFTvLeeKDbd1fPmOMaSOkASEis0Rko4hsEZF7gqyfJiIrRKRZROa0WXeziGx2HzeHspzd5uQvw8k3w4e/gfV/b7f6ikmDmTluAL9etInNe6vCUEBjjDkkZAEhIl7gt8DFwFhgroiMbbPZTuAW4K9t9k0H7gdOA6YC94tIWqjK2q0u+RUMPhlevgNKtxy2qnWupsSYKL73t1U0t/jDVEhjjAltDWIqsEVVt6lqIzAPmB24gaoWqOpqoO034UxgkaqWqeoBYBEwK4Rl7T5RMXDtsxAVDc/fCA2Hd0r3S4zhJ7PHs7qogt++szVMhTTGmNAGRBZQGPC+yF3WZfuKyNdEZJmILCspKTnugna71ByY86Qzsd9r/95u9tdLJw7iqilZPLp4E4vX7w1TIY0xka5Xd1Kr6u9VNV9V8zMzM8NdnGMzfDrM+CGsfQmemAGLfwLb3oOmOgB+ftUExg1O5lvzVrJln/VHGGO6XygDohjICXif7S4L9b69x1nfhgsfAPHCh4/As1fAL4fC05cR9/Gv+eMVmcT6PNz+zDIqapvCXVpjTIQJZUAsBUaJyDARiQauB17r5L4LgYtEJM3tnL7IXda3iMBZ34LbF8E9O+CGv8HUr0J9BbzzcwbMv4ynrkinuLyObzy3wjqtjTHdKmQBoarNwDdwvtjXA/NVda2IPCAiVwCIyKkiUgRcA/xORNa6+5YBP8EJmaXAA+6yvismCfIugpk/gzs+gK9/CupnwuKb+fXMfnywuZRfLtgQ7lIaYyKIaAe3x+xt8vPzddmyZeEuRtfatRKeuRySBvLQ4Ef538/KefiaScw5JTvcJTPG9BEislxV84Ot69Wd1H3e4MnOjLDlO7mr5D5mDIvlvpe+YMXOA+EumTEmAlhA9HS5Z8G1z+LZu4b/i3qYIcnCV59ZxsY9NrLJGBNaFhC9Qd5MuOp3+Ao/5tUBTxDjaWHuHz5h3a7KcJfMGNOHWUD0FhPmwKUPk1CwiLdynyPeq9zwxCesKa4Id8mMMX2UBURvcurtcMGPSNz8Cm9l/YE0Xws3/OETVhWWh7tkxpg+yAKitzn7O3Dpr4nfvog3M35DVlwjNz3xqXVcG2O6nAVEb3Tq7TDnSWL2fM5rCT9nZHw1X/njZywr6NuXihhjupcFRG81/mq4cT6+ih38zfcjJieUcdMfP+U3izZR3dAc7tIZY/oAC4jebMT5cMvrRDVV8ww/4Cu5FTy+eDPTf/UOf/q4gCabmsMYcwIsIHq7rFPg1oV4o2K4b99/8NbVXoZnJvKDV9dy0SPv88YXu+krV8sbY7qXBURfkJkHty2EpIHkvXUzz0+v4Imv5BPlEf7tLyu4+v8tYZPdwtQYc4wsIPqKlGz4lzchczTy/I1c0Pw+C751Dg9+aQKFZbXM/p+PeGlFUbhLaYzpRSwg+pKEDLj5dRhyBrz0VaKW/5HrTh3CG988h4nZKXx3/irueXE19U0t4S6pMaYXsIDoa2KT4cYX4KSL4Y274L1f0T8phr/cfhpfP28E85YWctX/LmF7aU24S2qM6eFsuu++qqUZXv06rJ4Hp93pDItVPysK9vPbdzbh9/u5Y9pwTjvjXIhPD3dpjTFhcqTpvi0g+jK/HxbeC5/+X4ebNEkMUZPmIKfeDlknd2PhjDE9gQVEJFOFomXQUAHiOfho8gvPfriZ2M1/Z070EmL8dTD4ZDj1Nhj/JfDFdf4YNaVQsgF88RCXCrGpEJMM3qjQnZcxpktYQJigVJVH3t7MU4tXcV/WKq6Tt/CUbnS+4EddCGnDIC3XfQyFpEFO4OxbC4WfQdFS5/nA9uAHiE6CuDSnZjLifOeRmtOdp2iMOYojBYT9xItgIsJ3L8wjPd7Hva/H80ruJTx5QyMJq56BnZ/CmhdBA67G9kaDJwqaap33iQMg+1TI/xcYMB6aG6C+wn2UO8/Ve2HHElj3irNPxqhDYTHiPIiK6f4TN8Z0igWE4ZazhpGWEM335q/imgVJPHPr/5GZFAPNjVBZBAcK3McOJwSy851gSB0CIkc/gKrTBLX1n85jxbPw2e+cJq2bXrROcmN6KGtiMge9u3Efd/55Bf2TY3j21qkMzUgIzYGa6mH9a/DqNyB9OHz5ZUgeFJpjGWOO6EhNTHYdhDlo+kn9+ctXT6OiromLHnmfn/59HaXVDV1/IF8sTLwWbnoBKgrhqVlODcUY06NYQJjDnDwkjde/cTaXTRzMkx9tZ9pD7/Dgmxs4UNPY9QcbNg2+8prTV/HkLNi3oeuPYYw5btbEZDq0taSax97ezOurd5EQHcWtZ+Vy2znDSYnzde2B9q6DP10FLY1On4Rdj2FMt7FhruaEbNxTxaNvb2LBmj3E+jxMG5XJzHEDmTGmP6nx0V1zkLJt8OyVUFsG1z4NIy/oms81xhyRBYTpEmt3VfD80kLeWruXPZX1eD3C1Nx0Zo4bwIXjBpKVegwX1wVTucupSZRsgCFnOvffHnVh50ZKGWOOiwWE6VKqyuqiCt5at4eFa/eyZV81APlD05g9eTCXTBhERuJxXt/QWAMr/gRL/tsZYtt/HJz9bRh3tV2ZbUwIWECYkNpaUs2ba/bw6spiNu2tJsojnDOqH7MnZ3Hh2AEkxBzHF3tLE3zxAnz0GJSsd665OPV2yMqHAeOcKT0i2fYP4O0fOeE55vJwl8b0YhYQpluoKhv2VPHqyl28vmoXxeV1xPm83HHuCO6cPoLoqOMYNOf3w+aF8OEjUPjpoeUpOc7V2wPGQf8xEJvizB/li4Mo99kXD9HxzrPH2/ExmuqhodIZTVVTCpXFTnNXZbHzqCiG5nrIOc0ZeZV7DiRmtv+csu2wdTFs+Sfs/Ni5mPD8/4JBE4/9vDuiCh//Dyy632l6Uz9c/jic/OWuO0ZfUlEMq/4K/hY45y6rhQZhAfEkArEAABQESURBVGG6nd+vLN95gKc/KuAfX+xm9MAkHvzSRCblnMAv/8rdsHeN89izBvauhdJNoJ24AVJUrBsYiU5otDQdCoWWDobwRidCchakZDmTHO78FBrdW7f2H+uExaBJsOtz2PK209EOkDIEhpwGm99yPn/8l+C8/4SMEcd/7gANVc7FhetegTFXwCW/glfudK5Ov+incOa/n9jn9xXNDbDxDfj8z7BlMeB+x+XNgjlPQnSILgDtpSwgTFi9vW4v//XKGvZV1XP7OcP5zgV5xEUf4Rf9sWiqdyYLbKxx5ohqqgt41DjPjTWH1jfWQmM1eH3OjLOxye5zivOIS3cCIXmw8z5QSzPsXgXb34OCD2DHx9Bc5wRP7tkwYgaMnAEZI51f93UH4KPHnenWmxtgyk1w7vedzz9WJZvg+Ztg/2a44Edw5jedYzQ3wEtfc0LjnO/B+T84tk79mv3OVe3N9XDqV7v3F/aBAij40AnWpEFOrTA1x7l9btt/+87YswY+/xOsft75t0/Ohsk3OI8tb8OCu2HQZLjheUjs3+Wn01tZQJiwq6xv4hdvrOe5zwrJzYjnF1dP5IwRGeEu1olpboD9W5xAONKkg1V74YOHYdlTTk1k+LnOLLexqU5fSuvr2BTn121MolvTSXAe29+HV77uHGPOk87+gfwt8PfvwIpnIP9WuOThIzep1VfAhn84kzFufedQDWzUTLjmqc7/wlZ1muB2rXSCc/dK2L3amdQxMw/6BTwyT3LCuuADJxS2fwAVO90PEg7+ym8Vk+zMIDxihlNbyjo5ePA1VMPal2D501C83Dn26MucMB4+/fB/hw1vwAu3OuFw00vQb2TnzrOrVe+DTW/CvvVOrSb3HPCE75rlsAWEiMwCHgO8wBOq+ss262OAZ4FTgP3AdapaICK5wHpgo7vpJ6p6x5GOZQHROyzZUso9L33BzrJa5pySzV0XncTAlNhwF6t7HNjhBMWulc5st3UVzn06OiPrFLj2WefXdTCqsPjHTl/NuKth+r1Ok1RDBdRXHmpO2/mJ0/TV0ug0hY2/2mkCK1rq3KJ28Mlww3zn/uYd2fMFvPtL57NqS51l4oHM0U6TW0uT0/RXutmpYbUVlw65Z0HuNBh2jhMgNSVQUQTlO53niiJncMKOJeBvdpr6xlzuPIac4ZRh+dPOQIbGKsgcA6fcDBOvO/Lkj0XL4K/XOX03c+c5TYFH429xzqV4uXPcwZOdf+OoTl4D1DpZ5cY3YOMCpwyoMzOyv9kZgDH5Rqemkzqkc5/ZhcISECLiBTYBFwJFwFJgrqquC9jm34CJqnqHiFwPXKWq17kB8XdVHd/Z41lA9B51jS08+vYmnvqoAK9H+Oq04fzrtOHHN9qpt/O3HD49emtzWGO18+u4scapOUy5qXNTo3/0GCz6YcfrEwc6oTDuamdW3sBf5etfhxduc5p5bnrJ+QUfqO4AvPNzWPqEU+M56WKnyWbwZGfAQHR8m3PzO0OVSzc5TWQeLww9y+m/6ewv5toy2LTQKdvWxU5TmC/BaT6MinPO5eSbIWdq55vWyrbBn7/kDES49NfOFPT+JufLuqXZeW6sdsKgeLkT6K19Tx6fs23SYDj9DjjlluDNYS3NUPiJU2vZ+Mahe6YMngInXeL822WMdGpyn/8Jtr3nrB82zQk5ETcoCw8FZkWR8/fic/vTWvvVfHEwcAJc/mjnzr+NcAXEGcCPVHWm+/5eAFX9RcA2C91tPhaRKGAPkAkMxQKiz9u5v5YHF27gH6t3k5kUw/cuzOOa/By8nkP/o1fUNbF+dyUbdlfSonDjaUOI9XVR/0VfVfCR8+V3sH+lzfORvkh3fAzPXed8+dz4gjMCy++HlX9xhtXWlUH+bXDefd0/TXtDtdOXsHUxDJwIE645/uHONaXw3PVOzakjHh8MHO/U3lofGSOdQQFLHnea/qITnYA6/Q6nZrR1sRMKmxc6geqNdr70T7rEaU7qqP+pfCesfA5W/tl53Sqhv1NrTM1x+lS8UU6/W2t/W7P7On24E3bHIVwBMQeYpaq3u++/DJymqt8I2GaNu02R+34rcBqQCKzFqYFUAv+lqh8EOcbXgK8BDBky5JQdO3aE5FxMaC3fcYCf/WMdK3aWc9KAJGaM6c+mvVWs311FcfnhTRS5GfE8+KWJnDa8l/df9GT71ju/sOsr4aIHnNFAxcsh53Rn5FRXDtsNp6Z62PERB5t7PD5n8ILH6wRk+gjn13pHdq10hhyvecl574mClgandpU30wmFkTMgJqnzZfL7Ye8Xh0bQHen4XaQ3BkQVkKiq+0XkFOAVYJyqVnZ0PKtB9G6qyoI1e/jlgg0Ul9cxvF8CowclM2ZQEmMGJTN2UDJbS6q550Wn/+LLpw/l+xePJjESm6W6Q0WxExIl651fsRf95FDThzlceSEs+6PT95I3y+kj6UXXW4TrlqPFQOANiLPdZcG2KXKbmFKA/eqkVgOAqi53gyMPsAToo0SESyYMYta4gTT5/cREtW9GGpAcy5vfPoeHF27iqSXb+eeGffzi6glMywty0RpO6Ih9oR2flCy4dQGsew3GXXl8w04jRWqOM/S4DwplDSIKp4loBk4QLAVuUNW1Adt8HZgQ0El9tapeKyKZQJmqtojIcOADd7uyjo5nNYjIsnxHGXe/sJqtJTVcOnEQ6fHRlNU0UlrdQFlNI/trGimvbSS3XwKnDk0nPzeNU3PTGZoRb6FhTIBwDnO9BHgUZ5jrk6r6MxF5AFimqq+JSCzwJ2AKUAZcr6rbRORLwANAE+AH7lfV1490LAuIyFPf1MJ//3MzT3ywnbhoL+kJ0fRLiCE9IZqMxGiS43xs3lvF0oIDVNQ1AdAvMZr8oelcMnEQl00YhMdjYWEim10oZ/q0ozUl+f3K1pJqlhYcYFlBGZ9uL6O4vI68AYl898I8Zo4baLUKE7EsIIwJ4Pcrf/9iN4++vYltJTWMG5zMdy/M4/zR/S0oTMSxgDAmiOYWP6+u3MVjizezs6yWyTmp3HJmLiP7J5LbL8FGSJmIYAFhzBE0tfh5cXkRjy/ezK6K+oPL+yXGkJsRz9CMBLLT4kiJ85EYG0VybBRJsT6SYqNIjvWRGu8jOdZn/RmmVwrXMFdjegWf18P1U4dw9cnZbN5XxY79tRTsr2FHqfP84ZYS9lY2HPEzPAIpcT7SEqJJi3cew/rFM3pgMmMGJTOif0LQobvG9GQWEMa4oqM8jBucwrjB7cf8N7X4qWlopqq+mcr6Jqrqm6l2X5fXNnGgttF9NHGgppGiA7W8v7mExmY/AFEeYURmIqMHJTEkPZ6UOB8pcT5S46NJjfeRGucjOc5HfLSX+Oiow6YbMSZcLCCM6QSf1+N+mXdyBk+cPo6C/TWs313Fhj2VbNhdxdLtZby+ahf+o7Tsxvo8JERHER/jJT0hhovGDmD25MFkp8UfeUdjupD1QRjTzfx+pbqxmYpap/ZRXtdIRV0TFXVN1DW2UNPQQm1jM9UNzdQ2tlCwv4bPd5YDcGpuGrMnZ3HphEGkJXQ+rIzpiHVSG9PLFZbV8urKYl5ZuYst+6rxeYVpozI5aWASA1Ni6Z8Uy8CUWAYmx9IvMZoob/huQGN6FwsIY/oIVWXtrkpeXVnMW+v2UnygjuY27VUegWH9EpiUk8qk7FQm5aQyZlCSdZKboCwgjOmj/H5lf00jeyvr2VtZz57KevZU1LN+dyUrCysorXZGX/m8wphByYzqn8SA5BgGJMcyIDmG/smxDEiOpX9SDD6rdUQkG+ZqTB/l8QiZSTFkJsUwPuvw0Veqyu6KelYVlrOqqIJVheV8vLWUfVUN7WodIjAwOZbstDiyUuPITosnOy2OQanO9R9JsVEHr/uIifLYFecRwgLCmD5KRBicGsfg1DgunjDo4PLAWse+qnr2Vjawu7yOovI6ig/UsbTgAK+v3k1LB0OtfF4hKdZHQoyXeJ8z0io+2kucL4qEGC8eEZr9SovfT3OL0uJXmv1KcpyP8YOTmZCVwrisFFLifN31T2GOkwWEMREmsNbh3IKlveYWP3sq69ldUU9lnXPdR1V9E5X1zuiqqvomahtaqGl0RlrVNrZQVlNHbWMzqs51H173EeUVvB4PW/ZV8/qqXQePkZsRz/isFIb3S6DZrzQ2+2lo9tPQ3EJDsxMuyXE+0hN8pMVHk54QTVpCNOnx0QfLb81ioWUBYYxpJ8rrcZuZuva6i7KaRr4ormBNcQVfFFXw+c5y/r56Nz6vEO31EOPzEhPlISbKg8cjVNY1c6C2MWhtRgQyEqLJTHL6Uwa0PrujuQa4j4yEaESgvLaJXRV17KmoZ1dFPbvL66hraiF/aDpnjMgg3YYNt2Od1MaYsPL79YjzWKkqlfXNHKhppKy2kbLqRkqqG9wmsgb2VTrNZHsq69lf3dDuIkSf16nJ1Df5D1se5RF8Xg91TS2IwNhByZw9sh9njuzH1Nx04qIjY9SXdVIbY3qso01yKCIHpybJJeGI2za3+CmpbmBPhRMarSO7mpr9DEyJZXBqHIPc536JMagqq4sr+GhzKR9tLeXJj7bzu/e34fUIafHuVChxzoSMKXHRpMT58Ku6FzE2U93QQk1DMzUNzQDERR/qj3GevcT4gjeDeURIjvORFu80oaXGH2pKG5wa1yOmW7EahDHGuGobmw/eWKq0upGKukbnavfaJsprGymvayLKIyTGRBEfE0VCTBSJMc78WeDc5bC1T6bO7Z9pbPET7Ku+2a9U1TcHLUdCtJfxWSlMykllYnYKk7JTyU6LQ0RQVeqaWg7O+3WgthGf18PpwzOO65ytBmGMMZ0QHx3FuXmZnJuX2S3Ha/ErFXXOZI/ltY0cqGmitLqBtbsqWV1UztMfFdDY4jSNpSdEE+31cKC2kYbmw5vLJuWk8urXz+ry8llAGGNMmHg9QnpCdIcd5A3NLWzcU8WqogrWFFXgVyU9wZk0Mi3+0PTy/ZNiQlI+CwhjjOmhYqK8TMxOZWJ2aliOb4OIjTHGBGUBYYwxJigLCGOMMUFZQBhjjAnKAsIYY0xQFhDGGGOCsoAwxhgTlAWEMcaYoPrMXEwiUgLsOIGP6AeUdlFxehM778hi5x1ZOnPeQ1U16NwifSYgTpSILOtowqq+zM47sth5R5YTPW9rYjLGGBOUBYQxxpigLCAO+X24CxAmdt6Rxc47spzQeVsfhDHGmKCsBmGMMSYoCwhjjDFBRXxAiMgsEdkoIltE5J5wlyeURORJEdknImsClqWLyCIR2ew+p4WzjF1NRHJE5B0RWScia0XkW+7yvn7esSLymYiscs/7x+7yYSLyqfv3/ryIBL+VWS8nIl4R+VxE/u6+j5TzLhCRL0RkpYgsc5cd9996RAeEiHiB3wIXA2OBuSIyNrylCqmngVltlt0DLFbVUcBi931f0gx8T1XHAqcDX3f/G/f1824AzlfVScBkYJaInA48CDyiqiOBA8BtYSxjKH0LWB/wPlLOG+A8VZ0ccP3Dcf+tR3RAAFOBLaq6TVUbgXnA7DCXKWRU9X2grM3i2cAz7utngCu7tVAhpqq7VXWF+7oK50sji75/3qqq1e5bn/tQ4HzgBXd5nztvABHJBi4FnnDfCxFw3kdw3H/rkR4QWUBhwPsid1kkGaCqu93Xe4AB4SxMKIlILjAF+JQIOG+3mWUlsA9YBGwFylW12d2kr/69PwrcDfjd9xlExnmD8yPgLRFZLiJfc5cd9996VFeXzvReqqoi0ifHPYtIIvAi8G1VrXR+VDr66nmragswWURSgZeB0WEuUsiJyGXAPlVdLiLTw12eMDhbVYtFpD+wSEQ2BK481r/1SK9BFAM5Ae+z3WWRZK+IDAJwn/eFuTxdTkR8OOHwF1V9yV3c58+7laqWA+8AZwCpItL6w7Av/r2fBVwhIgU4TcbnA4/R988bAFUtdp/34fwomMoJ/K1HekAsBUa5IxyigeuB18Jcpu72GnCz+/pm4NUwlqXLue3PfwTWq+pvAlb19fPOdGsOiEgccCFO/8s7wBx3sz533qp6r6pmq2ouzv/P/1TVG+nj5w0gIgkiktT6GrgIWMMJ/K1H/JXUInIJTpulF3hSVX8W5iKFjIg8B0zHmQJ4L3A/8AowHxiCM136taratiO71xKRs4EPgC841CZ9H04/RF8+74k4HZJenB+C81X1AREZjvPLOh34HLhJVRvCV9LQcZuY7lLVyyLhvN1zfNl9GwX8VVV/JiIZHOffesQHhDHGmOAivYnJGGNMBywgjDHGBGUBYYwxJigLCGOMMUFZQBhjjAnKAsKYYyAiLe5Mma2PLpvkT0RyA2faNSbcbKoNY45NnapODnchjOkOVoMwpgu48/A/5M7F/5mIjHSX54rIP0VktYgsFpEh7vIBIvKye7+GVSJypvtRXhH5g3sPh7fcq6CNCQsLCGOOTVybJqbrAtZVqOoE4H9wrs4H+G/gGVWdCPwFeNxd/jjwnnu/hpOBte7yUcBvVXUcUA58KcTnY0yH7EpqY46BiFSramKQ5QU4N+jZ5k4OuEdVM0SkFBikqk3u8t2q2k9ESoDswOke3OnIF7k3dkFEvg/4VPWnoT8zY9qzGoQxXUc7eH0sAucHasH6CU0YWUAY03WuC3j+2H29BGdWUYAbcSYOBOfWj3fCwRv7pHRXIY3pLPt1YsyxiXPv0tbqTVVtHeqaJiKrcWoBc91l/w48JSL/AZQA/+Iu/xbwexG5DaemcCewG2N6EOuDMKYLuH0Q+apaGu6yGNNVrInJGGNMUFaDMMYYE5TVIIwxxgRlAWGMMSYoCwhjjDFBWUAYY4wJygLCGGNMUP8f8F1kOwqL1z4AAAAASUVORK5CYII=\n",
            "text/plain": [
              "<Figure size 432x288 with 1 Axes>"
            ]
          },
          "metadata": {
            "tags": [],
            "needs_background": "light"
          }
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "HZXOyGRs7USe",
        "outputId": "efc77489-bcfb-4f67-c49f-93cbf8361b48",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 282
        }
      },
      "source": [
        "y = model.predict(X_test)\n",
        "plt.plot(y[:,1],'ro')\n",
        "plt.plot(Y_test[:,1],'bo')"
      ],
      "execution_count": 87,
      "outputs": [
        {
          "output_type": "execute_result",
          "data": {
            "text/plain": [
              "[<matplotlib.lines.Line2D at 0x7fa4c3261668>]"
            ]
          },
          "metadata": {
            "tags": []
          },
          "execution_count": 87
        },
        {
          "output_type": "display_data",
          "data": {
            "image/png": "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\n",
            "text/plain": [
              "<Figure size 432x288 with 1 Axes>"
            ]
          },
          "metadata": {
            "tags": [],
            "needs_background": "light"
          }
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "cZkMOT29zng-",
        "outputId": "87057754-1b18-45ca-abff-236f0e51d10b",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 204
        }
      },
      "source": [
        "os.chdir('/content/drive/My Drive/Colab Notebooks')\n",
        "i = 2\n",
        "img = []\n",
        "while i <= 22:\n",
        "  img.append(np.asarray(Image.open('image'+str(i)+'.jpg')))\n",
        "  \n",
        "  i += 2\n",
        "\n",
        "img = np.array(img)\n",
        "model.predict(img)"
      ],
      "execution_count": 88,
      "outputs": [
        {
          "output_type": "execute_result",
          "data": {
            "text/plain": [
              "array([[ 1.19842544e-01,  2.24056020e-01],\n",
              "       [ 3.44115913e-01,  3.44994724e-01],\n",
              "       [ 2.40915835e-01,  3.27651232e-01],\n",
              "       [ 2.92024314e-01,  3.14950168e-01],\n",
              "       [ 2.90021509e-01,  3.43325853e-01],\n",
              "       [ 5.27060404e-02,  3.46850336e-01],\n",
              "       [-4.26858664e-04,  2.67756164e-01],\n",
              "       [ 3.16483974e-01,  3.35296094e-01],\n",
              "       [ 5.58743514e-02,  3.67838919e-01],\n",
              "       [ 3.18777472e-01,  4.43879426e-01],\n",
              "       [ 2.08305404e-01,  1.80959981e-02]], dtype=float32)"
            ]
          },
          "metadata": {
            "tags": []
          },
          "execution_count": 88
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "c3kOb2CjwL3Q"
      },
      "source": [
        "model.save(\"model1.h5\")   "
      ],
      "execution_count": 89,
      "outputs": []
    }
  ]
}