tf2_linear_classification.ipynb

{
  "nbformat": 4,
  "nbformat_minor": 0,
  "metadata": {
    "colab": {
      "name": "tf2_linear_classification.ipynb",
      "provenance": [],
      "collapsed_sections": [],
      "mount_file_id": "1bU-_KuSSMT6PZiajmIhHXXbhhj-THsPe",
      "authorship_tag": "ABX9TyPBIBx4l1ukUHwItUa1KuuZ",
      "include_colab_link": true
    },
    "kernelspec": {
      "name": "python3",
      "display_name": "Python 3"
    },
    "language_info": {
      "name": "python"
    }
  },
  "cells": [
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "view-in-github",
        "colab_type": "text"
      },
      "source": [
        "<a href=\"https://colab.research.google.com/github/xmsi/tf2_tasks/blob/main/tf2_linear_classification.ipynb\" target=\"_parent\"><img src=\"https://colab.research.google.com/assets/colab-badge.svg\" alt=\"Open In Colab\"/></a>"
      ]
    },
    {
      "cell_type": "code",
      "execution_count": 1,
      "metadata": {
        "id": "MTghGvCUBsJe"
      },
      "outputs": [],
      "source": [
        "from sklearn.datasets import load_iris"
      ]
    },
    {
      "cell_type": "code",
      "source": [
        "data = load_iris()"
      ],
      "metadata": {
        "id": "EYy4qh6UB84l"
      },
      "execution_count": 2,
      "outputs": []
    },
    {
      "cell_type": "code",
      "source": [
        "type(data)"
      ],
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "d33m9XrvDb6s",
        "outputId": "61a43bb3-4b8d-4f31-eb87-5e7124d2d2f3"
      },
      "execution_count": 3,
      "outputs": [
        {
          "output_type": "execute_result",
          "data": {
            "text/plain": [
              "sklearn.utils.Bunch"
            ]
          },
          "metadata": {},
          "execution_count": 3
        }
      ]
    },
    {
      "cell_type": "code",
      "source": [
        "data.keys()"
      ],
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "OG73Vsx1DiI2",
        "outputId": "d0c11c53-2c7a-4ec4-95da-b5236dee2320"
      },
      "execution_count": 4,
      "outputs": [
        {
          "output_type": "execute_result",
          "data": {
            "text/plain": [
              "dict_keys(['data', 'target', 'frame', 'target_names', 'DESCR', 'feature_names', 'filename', 'data_module'])"
            ]
          },
          "metadata": {},
          "execution_count": 4
        }
      ]
    },
    {
      "cell_type": "code",
      "source": [
        "data.data.shape\n",
        "\n",
        "# 150 items 4 features X"
      ],
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "AVWnjBnJDrdX",
        "outputId": "cc6cff68-b9b1-4004-9135-43938382c838"
      },
      "execution_count": 5,
      "outputs": [
        {
          "output_type": "execute_result",
          "data": {
            "text/plain": [
              "(150, 4)"
            ]
          },
          "metadata": {},
          "execution_count": 5
        }
      ]
    },
    {
      "cell_type": "code",
      "source": [
        "data.target[0:100]\n",
        "\n",
        "# only 0 and 1 in 0:100 Y"
      ],
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "y9KY1rExDyQV",
        "outputId": "5ed102ba-744d-4194-cfb1-f2270f374612"
      },
      "execution_count": 6,
      "outputs": [
        {
          "output_type": "execute_result",
          "data": {
            "text/plain": [
              "array([0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0,\n",
              "       0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0,\n",
              "       0, 0, 0, 0, 0, 0, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1,\n",
              "       1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1,\n",
              "       1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1])"
            ]
          },
          "metadata": {},
          "execution_count": 6
        }
      ]
    },
    {
      "cell_type": "code",
      "source": [
        "data.target_names"
      ],
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "Cb_KvKHIEHci",
        "outputId": "81b7c6d1-0983-41e4-f32a-1f8358d1f110"
      },
      "execution_count": 7,
      "outputs": [
        {
          "output_type": "execute_result",
          "data": {
            "text/plain": [
              "array(['setosa', 'versicolor', 'virginica'], dtype='<U10')"
            ]
          },
          "metadata": {},
          "execution_count": 7
        }
      ]
    },
    {
      "cell_type": "code",
      "source": [
        "data.feature_names"
      ],
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "nzmI5ecpEaur",
        "outputId": "ce17f658-3b15-4734-d65d-2883e7eddabb"
      },
      "execution_count": 8,
      "outputs": [
        {
          "output_type": "execute_result",
          "data": {
            "text/plain": [
              "['sepal length (cm)',\n",
              " 'sepal width (cm)',\n",
              " 'petal length (cm)',\n",
              " 'petal width (cm)']"
            ]
          },
          "metadata": {},
          "execution_count": 8
        }
      ]
    },
    {
      "cell_type": "code",
      "source": [
        "from sklearn.model_selection import train_test_split\n",
        "\n",
        "X_train, X_test, y_train, y_test = train_test_split(data.data[0:100], data.target[0:100], test_size=0.33)\n",
        "N, D = X_train.shape "
      ],
      "metadata": {
        "id": "ng7JBDpLEkaB"
      },
      "execution_count": 9,
      "outputs": []
    },
    {
      "cell_type": "code",
      "source": [
        "D"
      ],
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "YVv-_85yFJRJ",
        "outputId": "f9e208dc-5f48-4e7a-a580-5389dc44aadb"
      },
      "execution_count": 10,
      "outputs": [
        {
          "output_type": "execute_result",
          "data": {
            "text/plain": [
              "4"
            ]
          },
          "metadata": {},
          "execution_count": 10
        }
      ]
    },
    {
      "cell_type": "code",
      "source": [
        "from sklearn.preprocessing import StandardScaler\n",
        "\n",
        "scaler = StandardScaler()\n",
        "X_train = scaler.fit_transform(X_train)\n",
        "X_test = scaler.transform(X_test)"
      ],
      "metadata": {
        "id": "kZ3lNVk1FRsX"
      },
      "execution_count": 11,
      "outputs": []
    },
    {
      "cell_type": "code",
      "source": [
        "#tf2 works here!\n",
        "import tensorflow as tf\n",
        "\n",
        "model = tf.keras.models.Sequential([\n",
        "                         tf.keras.layers.Input(shape=(D,)),\n",
        "                         tf.keras.layers.Dense(1, activation='sigmoid')          \n",
        "])\n",
        "\n",
        "#model = tf.keras.models.Sequential()\n",
        "#model.add(tf.keras.layers.Dense(1, input_shape=(D,), activation='sigmoid'))\n",
        "\n",
        "model.compile(optimizer='adam',\n",
        "              loss='binary_crossentropy',\n",
        "              metrics=['accuracy'])\n",
        "\n",
        "r = model.fit(X_train, y_train, validation_data=(X_test, y_test), epochs=200)"
      ],
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "3NNTyD9UFjKo",
        "outputId": "51260f1a-9af5-4239-c760-caca619232dd"
      },
      "execution_count": 12,
      "outputs": [
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "Epoch 1/200\n",
            "3/3 [==============================] - 1s 88ms/step - loss: 0.9981 - accuracy: 0.2687 - val_loss: 0.9658 - val_accuracy: 0.3030\n",
            "Epoch 2/200\n",
            "3/3 [==============================] - 0s 16ms/step - loss: 0.9921 - accuracy: 0.2687 - val_loss: 0.9600 - val_accuracy: 0.3030\n",
            "Epoch 3/200\n",
            "3/3 [==============================] - 0s 15ms/step - loss: 0.9861 - accuracy: 0.2836 - val_loss: 0.9541 - val_accuracy: 0.3030\n",
            "Epoch 4/200\n",
            "3/3 [==============================] - 0s 15ms/step - loss: 0.9801 - accuracy: 0.2836 - val_loss: 0.9482 - val_accuracy: 0.3333\n",
            "Epoch 5/200\n",
            "3/3 [==============================] - 0s 13ms/step - loss: 0.9742 - accuracy: 0.2985 - val_loss: 0.9424 - val_accuracy: 0.3333\n",
            "Epoch 6/200\n",
            "3/3 [==============================] - 0s 14ms/step - loss: 0.9684 - accuracy: 0.2985 - val_loss: 0.9366 - val_accuracy: 0.3333\n",
            "Epoch 7/200\n",
            "3/3 [==============================] - 0s 19ms/step - loss: 0.9626 - accuracy: 0.2985 - val_loss: 0.9310 - val_accuracy: 0.3333\n",
            "Epoch 8/200\n",
            "3/3 [==============================] - 0s 14ms/step - loss: 0.9570 - accuracy: 0.3134 - val_loss: 0.9254 - val_accuracy: 0.3333\n",
            "Epoch 9/200\n",
            "3/3 [==============================] - 0s 16ms/step - loss: 0.9513 - accuracy: 0.3284 - val_loss: 0.9198 - val_accuracy: 0.3333\n",
            "Epoch 10/200\n",
            "3/3 [==============================] - 0s 20ms/step - loss: 0.9457 - accuracy: 0.3433 - val_loss: 0.9143 - val_accuracy: 0.3333\n",
            "Epoch 11/200\n",
            "3/3 [==============================] - 0s 14ms/step - loss: 0.9402 - accuracy: 0.3582 - val_loss: 0.9088 - val_accuracy: 0.3333\n",
            "Epoch 12/200\n",
            "3/3 [==============================] - 0s 15ms/step - loss: 0.9347 - accuracy: 0.3582 - val_loss: 0.9034 - val_accuracy: 0.3333\n",
            "Epoch 13/200\n",
            "3/3 [==============================] - 0s 15ms/step - loss: 0.9293 - accuracy: 0.3731 - val_loss: 0.8979 - val_accuracy: 0.3333\n",
            "Epoch 14/200\n",
            "3/3 [==============================] - 0s 13ms/step - loss: 0.9238 - accuracy: 0.3731 - val_loss: 0.8925 - val_accuracy: 0.3333\n",
            "Epoch 15/200\n",
            "3/3 [==============================] - 0s 13ms/step - loss: 0.9184 - accuracy: 0.3881 - val_loss: 0.8869 - val_accuracy: 0.3333\n",
            "Epoch 16/200\n",
            "3/3 [==============================] - 0s 15ms/step - loss: 0.9131 - accuracy: 0.3881 - val_loss: 0.8812 - val_accuracy: 0.3333\n",
            "Epoch 17/200\n",
            "3/3 [==============================] - 0s 15ms/step - loss: 0.9076 - accuracy: 0.4030 - val_loss: 0.8755 - val_accuracy: 0.3333\n",
            "Epoch 18/200\n",
            "3/3 [==============================] - 0s 15ms/step - loss: 0.9023 - accuracy: 0.4030 - val_loss: 0.8701 - val_accuracy: 0.3636\n",
            "Epoch 19/200\n",
            "3/3 [==============================] - 0s 15ms/step - loss: 0.8969 - accuracy: 0.4030 - val_loss: 0.8648 - val_accuracy: 0.3636\n",
            "Epoch 20/200\n",
            "3/3 [==============================] - 0s 14ms/step - loss: 0.8918 - accuracy: 0.4030 - val_loss: 0.8593 - val_accuracy: 0.3636\n",
            "Epoch 21/200\n",
            "3/3 [==============================] - 0s 19ms/step - loss: 0.8035 - accuracy: 0.4478 - val_loss: 0.7705 - val_accuracy: 0.4848\n",
            "Epoch 39/200\n",
            "3/3 [==============================] - 0s 14ms/step - loss: 0.7990 - accuracy: 0.4478 - val_loss: 0.7658 - val_accuracy: 0.4848\n",
            "Epoch 40/200\n",
            "3/3 [==============================] - 0s 13ms/step - loss: 0.7944 - accuracy: 0.4776 - val_loss: 0.7610 - val_accuracy: 0.5152\n",
            "Epoch 41/200\n",
            "3/3 [==============================] - 0s 14ms/step - loss: 0.7898 - accuracy: 0.4776 - val_loss: 0.7562 - val_accuracy: 0.5152\n",
            "Epoch 42/200\n",
            "3/3 [==============================] - 0s 16ms/step - loss: 0.7851 - accuracy: 0.4776 - val_loss: 0.7514 - val_accuracy: 0.5152\n",
            "Epoch 43/200\n",
            "3/3 [==============================] - 0s 15ms/step - loss: 0.7805 - accuracy: 0.4776 - val_loss: 0.7466 - val_accuracy: 0.5152\n",
            "Epoch 44/200\n",
            "3/3 [==============================] - 0s 14ms/step - loss: 0.7760 - accuracy: 0.4776 - val_loss: 0.7418 - val_accuracy: 0.5152\n",
            "Epoch 45/200\n",
            "3/3 [==============================] - 0s 16ms/step - loss: 0.7714 - accuracy: 0.4925 - val_loss: 0.7370 - val_accuracy: 0.5152\n",
            "Epoch 46/200\n",
            "3/3 [==============================] - 0s 14ms/step - loss: 0.7669 - accuracy: 0.4925 - val_loss: 0.7323 - val_accuracy: 0.5152\n",
            "Epoch 47/200\n",
            "3/3 [==============================] - 0s 13ms/step - loss: 0.7625 - accuracy: 0.4925 - val_loss: 0.7277 - val_accuracy: 0.5152\n",
            "Epoch 48/200\n",
            "3/3 [==============================] - 0s 14ms/step - loss: 0.7581 - accuracy: 0.5075 - val_loss: 0.7232 - val_accuracy: 0.5455\n",
            "Epoch 49/200\n",
            "3/3 [==============================] - 0s 16ms/step - loss: 0.7537 - accuracy: 0.5075 - val_loss: 0.7188 - val_accuracy: 0.5455\n",
            "Epoch 50/200\n",
            "3/3 [==============================] - 0s 16ms/step - loss: 0.7495 - accuracy: 0.5075 - val_loss: 0.7145 - val_accuracy: 0.5758\n",
            "Epoch 51/200\n",
            "3/3 [==============================] - 0s 15ms/step - loss: 0.7452 - accuracy: 0.5224 - val_loss: 0.7103 - val_accuracy: 0.5758\n",
            "Epoch 52/200\n",
            "3/3 [==============================] - 0s 14ms/step - loss: 0.7409 - accuracy: 0.5224 - val_loss: 0.7062 - val_accuracy: 0.5758\n",
            "Epoch 53/200\n",
            "3/3 [==============================] - 0s 14ms/step - loss: 0.7368 - accuracy: 0.5373 - val_loss: 0.7021 - val_accuracy: 0.5758\n",
            "Epoch 54/200\n",
            "3/3 [==============================] - 0s 16ms/step - loss: 0.7327 - accuracy: 0.5373 - val_loss: 0.6979 - val_accuracy: 0.5758\n",
            "Epoch 55/200\n",
            "3/3 [==============================] - 0s 15ms/step - loss: 0.7286 - accuracy: 0.5373 - val_loss: 0.6937 - val_accuracy: 0.5758\n",
            "Epoch 56/200\n",
            "3/3 [==============================] - 0s 16ms/step - loss: 0.7244 - accuracy: 0.5522 - val_loss: 0.6895 - val_accuracy: 0.6061\n",
            "Epoch 57/200\n",
            "3/3 [==============================] - 0s 14ms/step - loss: 0.7203 - accuracy: 0.5522 - val_loss: 0.6852 - val_accuracy: 0.6061\n",
            "Epoch 58/200\n",
            "3/3 [==============================] - 0s 14ms/step - loss: 0.7163 - accuracy: 0.5522 - val_loss: 0.6810 - val_accuracy: 0.6364\n",
            "Epoch 59/200\n",
            "3/3 [==============================] - 0s 14ms/step - loss: 0.7122 - accuracy: 0.5672 - val_loss: 0.6768 - val_accuracy: 0.6364\n",
            "Epoch 60/200\n",
            "3/3 [==============================] - 0s 16ms/step - loss: 0.7083 - accuracy: 0.5672 - val_loss: 0.6727 - val_accuracy: 0.6364\n",
            "Epoch 61/200\n",
            "3/3 [==============================] - 0s 14ms/step - loss: 0.7043 - accuracy: 0.5672 - val_loss: 0.6687 - val_accuracy: 0.6364\n",
            "Epoch 62/200\n",
            "3/3 [==============================] - 0s 17ms/step - loss: 0.7004 - accuracy: 0.5672 - val_loss: 0.6646 - val_accuracy: 0.6364\n",
            "Epoch 63/200\n",
            "3/3 [==============================] - 0s 15ms/step - loss: 0.6964 - accuracy: 0.5672 - val_loss: 0.6606 - val_accuracy: 0.6364\n",
            "Epoch 64/200\n",
            "3/3 [==============================] - 0s 16ms/step - loss: 0.6925 - accuracy: 0.5672 - val_loss: 0.6567 - val_accuracy: 0.6364\n",
            "Epoch 65/200\n",
            "3/3 [==============================] - 0s 15ms/step - loss: 0.6885 - accuracy: 0.5970 - val_loss: 0.6528 - val_accuracy: 0.6364\n",
            "Epoch 66/200\n",
            "3/3 [==============================] - 0s 17ms/step - loss: 0.6847 - accuracy: 0.5970 - val_loss: 0.6489 - val_accuracy: 0.6364\n",
            "Epoch 67/200\n",
            "3/3 [==============================] - 0s 16ms/step - loss: 0.6807 - accuracy: 0.5970 - val_loss: 0.6451 - val_accuracy: 0.6364\n",
            "Epoch 68/200\n",
            "3/3 [==============================] - 0s 17ms/step - loss: 0.6770 - accuracy: 0.6119 - val_loss: 0.6414 - val_accuracy: 0.6364\n",
            "Epoch 69/200\n",
            "3/3 [==============================] - 0s 17ms/step - loss: 0.6731 - accuracy: 0.6119 - val_loss: 0.6377 - val_accuracy: 0.6364\n",
            "Epoch 70/200\n",
            "3/3 [==============================] - 0s 15ms/step - loss: 0.6694 - accuracy: 0.6119 - val_loss: 0.6340 - val_accuracy: 0.6364\n",
            "Epoch 71/200\n",
            "3/3 [==============================] - 0s 14ms/step - loss: 0.6656 - accuracy: 0.6119 - val_loss: 0.6303 - val_accuracy: 0.6364\n",
            "Epoch 72/200\n",
            "3/3 [==============================] - 0s 16ms/step - loss: 0.6619 - accuracy: 0.6119 - val_loss: 0.6265 - val_accuracy: 0.6364\n",
            "Epoch 73/200\n",
            "3/3 [==============================] - 0s 18ms/step - loss: 0.6582 - accuracy: 0.6269 - val_loss: 0.6227 - val_accuracy: 0.6364\n",
            "Epoch 74/200\n",
            "3/3 [==============================] - 0s 15ms/step - loss: 0.6545 - accuracy: 0.6418 - val_loss: 0.6190 - val_accuracy: 0.6667\n",
            "Epoch 75/200\n",
            "3/3 [==============================] - 0s 17ms/step - loss: 0.6508 - accuracy: 0.6418 - val_loss: 0.6153 - val_accuracy: 0.6667\n",
            "Epoch 76/200\n",
            "3/3 [==============================] - 0s 20ms/step - loss: 0.6472 - accuracy: 0.6418 - val_loss: 0.6117 - val_accuracy: 0.6667\n",
            "Epoch 77/200\n",
            "3/3 [==============================] - 0s 16ms/step - loss: 0.6437 - accuracy: 0.6567 - val_loss: 0.6080 - val_accuracy: 0.6667\n",
            "Epoch 78/200\n",
            "3/3 [==============================] - 0s 15ms/step - loss: 0.6401 - accuracy: 0.6567 - val_loss: 0.6044 - val_accuracy: 0.6970\n",
            "Epoch 79/200\n",
            "3/3 [==============================] - 0s 16ms/step - loss: 0.6366 - accuracy: 0.6567 - val_loss: 0.6008 - val_accuracy: 0.6970\n",
            "Epoch 80/200\n",
            "3/3 [==============================] - 0s 17ms/step - loss: 0.6331 - accuracy: 0.6716 - val_loss: 0.5972 - val_accuracy: 0.6970\n",
            "Epoch 81/200\n",
            "3/3 [==============================] - 0s 15ms/step - loss: 0.6296 - accuracy: 0.6866 - val_loss: 0.5935 - val_accuracy: 0.6970\n",
            "Epoch 82/200\n",
            "3/3 [==============================] - 0s 14ms/step - loss: 0.6260 - accuracy: 0.6866 - val_loss: 0.5900 - val_accuracy: 0.6970\n",
            "Epoch 83/200\n",
            "3/3 [==============================] - 0s 15ms/step - loss: 0.6226 - accuracy: 0.6866 - val_loss: 0.5864 - val_accuracy: 0.6970\n",
            "Epoch 84/200\n",
            "3/3 [==============================] - 0s 15ms/step - loss: 0.6192 - accuracy: 0.6866 - val_loss: 0.5829 - val_accuracy: 0.6970\n",
            "Epoch 85/200\n",
            "3/3 [==============================] - 0s 16ms/step - loss: 0.6158 - accuracy: 0.7164 - val_loss: 0.5796 - val_accuracy: 0.6970\n",
            "Epoch 86/200\n",
            "3/3 [==============================] - 0s 14ms/step - loss: 0.6125 - accuracy: 0.7164 - val_loss: 0.5763 - val_accuracy: 0.6970\n",
            "Epoch 87/200\n",
            "3/3 [==============================] - 0s 16ms/step - loss: 0.6093 - accuracy: 0.7164 - val_loss: 0.5731 - val_accuracy: 0.6970\n",
            "Epoch 88/200\n",
            "3/3 [==============================] - 0s 16ms/step - loss: 0.6061 - accuracy: 0.7164 - val_loss: 0.5698 - val_accuracy: 0.6970\n",
            "Epoch 89/200\n",
            "3/3 [==============================] - 0s 15ms/step - loss: 0.6029 - accuracy: 0.7164 - val_loss: 0.5667 - val_accuracy: 0.6970\n",
            "Epoch 90/200\n",
            "3/3 [==============================] - 0s 14ms/step - loss: 0.5997 - accuracy: 0.7164 - val_loss: 0.5635 - val_accuracy: 0.6970\n",
            "Epoch 91/200\n",
            "3/3 [==============================] - 0s 14ms/step - loss: 0.5964 - accuracy: 0.7164 - val_loss: 0.5603 - val_accuracy: 0.6970\n",
            "Epoch 92/200\n",
            "3/3 [==============================] - 0s 19ms/step - loss: 0.5933 - accuracy: 0.7164 - val_loss: 0.5571 - val_accuracy: 0.6970\n",
            "Epoch 93/200\n",
            "3/3 [==============================] - 0s 14ms/step - loss: 0.5900 - accuracy: 0.7313 - val_loss: 0.5539 - val_accuracy: 0.6970\n",
            "Epoch 94/200\n",
            "3/3 [==============================] - 0s 17ms/step - loss: 0.5868 - accuracy: 0.7313 - val_loss: 0.5505 - val_accuracy: 0.7273\n",
            "Epoch 95/200\n",
            "3/3 [==============================] - 0s 15ms/step - loss: 0.5836 - accuracy: 0.7463 - val_loss: 0.5472 - val_accuracy: 0.7273\n",
            "Epoch 96/200\n",
            "3/3 [==============================] - 0s 16ms/step - loss: 0.5803 - accuracy: 0.7463 - val_loss: 0.5439 - val_accuracy: 0.7273\n",
            "Epoch 97/200\n",
            "3/3 [==============================] - 0s 16ms/step - loss: 0.5771 - accuracy: 0.7463 - val_loss: 0.5408 - val_accuracy: 0.7273\n",
            "Epoch 98/200\n",
            "3/3 [==============================] - 0s 15ms/step - loss: 0.5740 - accuracy: 0.7463 - val_loss: 0.5375 - val_accuracy: 0.7273\n",
            "Epoch 99/200\n",
            "3/3 [==============================] - 0s 15ms/step - loss: 0.5708 - accuracy: 0.7463 - val_loss: 0.5343 - val_accuracy: 0.7273\n",
            "Epoch 100/200\n",
            "3/3 [==============================] - 0s 14ms/step - loss: 0.5676 - accuracy: 0.7463 - val_loss: 0.5311 - val_accuracy: 0.7273\n",
            "Epoch 101/200\n",
            "3/3 [==============================] - 0s 14ms/step - loss: 0.5645 - accuracy: 0.7463 - val_loss: 0.5280 - val_accuracy: 0.7273\n",
            "Epoch 102/200\n",
            "3/3 [==============================] - 0s 15ms/step - loss: 0.5613 - accuracy: 0.7463 - val_loss: 0.5249 - val_accuracy: 0.7273\n",
            "Epoch 103/200\n",
            "3/3 [==============================] - 0s 16ms/step - loss: 0.5583 - accuracy: 0.7612 - val_loss: 0.5219 - val_accuracy: 0.7273\n",
            "Epoch 104/200\n",
            "3/3 [==============================] - 0s 16ms/step - loss: 0.5552 - accuracy: 0.7612 - val_loss: 0.5189 - val_accuracy: 0.7273\n",
            "Epoch 105/200\n",
            "3/3 [==============================] - 0s 16ms/step - loss: 0.5523 - accuracy: 0.7612 - val_loss: 0.5159 - val_accuracy: 0.7273\n",
            "Epoch 106/200\n",
            "3/3 [==============================] - 0s 22ms/step - loss: 0.5494 - accuracy: 0.7612 - val_loss: 0.5129 - val_accuracy: 0.7273\n",
            "Epoch 107/200\n",
            "3/3 [==============================] - 0s 13ms/step - loss: 0.5464 - accuracy: 0.7612 - val_loss: 0.5098 - val_accuracy: 0.7273\n",
            "Epoch 108/200\n",
            "3/3 [==============================] - 0s 15ms/step - loss: 0.5435 - accuracy: 0.7612 - val_loss: 0.5069 - val_accuracy: 0.7273\n",
            "Epoch 109/200\n",
            "3/3 [==============================] - 0s 15ms/step - loss: 0.5406 - accuracy: 0.7612 - val_loss: 0.5040 - val_accuracy: 0.7576\n",
            "Epoch 110/200\n",
            "3/3 [==============================] - 0s 15ms/step - loss: 0.5378 - accuracy: 0.7612 - val_loss: 0.5012 - val_accuracy: 0.7879\n",
            "Epoch 111/200\n",
            "3/3 [==============================] - 0s 17ms/step - loss: 0.5350 - accuracy: 0.7612 - val_loss: 0.4985 - val_accuracy: 0.7879\n",
            "Epoch 112/200\n",
            "3/3 [==============================] - 0s 15ms/step - loss: 0.5323 - accuracy: 0.7910 - val_loss: 0.4958 - val_accuracy: 0.7879\n",
            "Epoch 113/200\n",
            "3/3 [==============================] - 0s 20ms/step - loss: 0.5296 - accuracy: 0.7910 - val_loss: 0.4932 - val_accuracy: 0.7879\n",
            "Epoch 114/200\n",
            "3/3 [==============================] - 0s 18ms/step - loss: 0.5268 - accuracy: 0.7910 - val_loss: 0.4906 - val_accuracy: 0.7879\n",
            "Epoch 115/200\n",
            "3/3 [==============================] - 0s 19ms/step - loss: 0.5241 - accuracy: 0.7910 - val_loss: 0.4880 - val_accuracy: 0.7879\n",
            "Epoch 116/200\n",
            "3/3 [==============================] - 0s 14ms/step - loss: 0.5214 - accuracy: 0.7910 - val_loss: 0.4853 - val_accuracy: 0.7879\n",
            "Epoch 117/200\n",
            "3/3 [==============================] - 0s 15ms/step - loss: 0.5187 - accuracy: 0.7910 - val_loss: 0.4827 - val_accuracy: 0.7879\n",
            "Epoch 118/200\n",
            "3/3 [==============================] - 0s 14ms/step - loss: 0.5160 - accuracy: 0.7910 - val_loss: 0.4801 - val_accuracy: 0.7879\n",
            "Epoch 119/200\n",
            "3/3 [==============================] - 0s 14ms/step - loss: 0.5133 - accuracy: 0.7910 - val_loss: 0.4775 - val_accuracy: 0.7879\n",
            "Epoch 120/200\n",
            "3/3 [==============================] - 0s 15ms/step - loss: 0.5106 - accuracy: 0.7910 - val_loss: 0.4749 - val_accuracy: 0.7879\n",
            "Epoch 121/200\n",
            "3/3 [==============================] - 0s 16ms/step - loss: 0.5079 - accuracy: 0.7910 - val_loss: 0.4720 - val_accuracy: 0.8182\n",
            "Epoch 122/200\n",
            "3/3 [==============================] - 0s 14ms/step - loss: 0.5051 - accuracy: 0.7910 - val_loss: 0.4692 - val_accuracy: 0.8182\n",
            "Epoch 123/200\n",
            "3/3 [==============================] - 0s 14ms/step - loss: 0.5024 - accuracy: 0.7910 - val_loss: 0.4664 - val_accuracy: 0.8182\n",
            "Epoch 124/200\n",
            "3/3 [==============================] - 0s 15ms/step - loss: 0.4996 - accuracy: 0.7910 - val_loss: 0.4636 - val_accuracy: 0.8182\n",
            "Epoch 125/200\n",
            "3/3 [==============================] - 0s 14ms/step - loss: 0.4969 - accuracy: 0.7910 - val_loss: 0.4609 - val_accuracy: 0.8182\n",
            "Epoch 126/200\n",
            "3/3 [==============================] - 0s 15ms/step - loss: 0.4942 - accuracy: 0.7910 - val_loss: 0.4583 - val_accuracy: 0.8182\n",
            "Epoch 127/200\n",
            "3/3 [==============================] - 0s 16ms/step - loss: 0.4916 - accuracy: 0.7910 - val_loss: 0.4556 - val_accuracy: 0.8182\n",
            "Epoch 128/200\n",
            "3/3 [==============================] - 0s 17ms/step - loss: 0.4890 - accuracy: 0.7910 - val_loss: 0.4530 - val_accuracy: 0.8182\n",
            "Epoch 129/200\n",
            "3/3 [==============================] - 0s 16ms/step - loss: 0.4864 - accuracy: 0.7910 - val_loss: 0.4504 - val_accuracy: 0.8182\n",
            "Epoch 130/200\n",
            "3/3 [==============================] - 0s 15ms/step - loss: 0.4838 - accuracy: 0.8060 - val_loss: 0.4477 - val_accuracy: 0.8182\n",
            "Epoch 131/200\n",
            "3/3 [==============================] - 0s 16ms/step - loss: 0.4812 - accuracy: 0.8060 - val_loss: 0.4451 - val_accuracy: 0.8182\n",
            "Epoch 132/200\n",
            "3/3 [==============================] - 0s 16ms/step - loss: 0.4787 - accuracy: 0.8060 - val_loss: 0.4426 - val_accuracy: 0.8182\n",
            "Epoch 133/200\n",
            "3/3 [==============================] - 0s 13ms/step - loss: 0.4762 - accuracy: 0.8060 - val_loss: 0.4400 - val_accuracy: 0.8182\n",
            "Epoch 134/200\n",
            "3/3 [==============================] - 0s 15ms/step - loss: 0.4737 - accuracy: 0.8060 - val_loss: 0.4375 - val_accuracy: 0.8182\n",
            "Epoch 135/200\n",
            "3/3 [==============================] - 0s 16ms/step - loss: 0.4713 - accuracy: 0.8209 - val_loss: 0.4350 - val_accuracy: 0.8182\n",
            "Epoch 136/200\n",
            "3/3 [==============================] - 0s 16ms/step - loss: 0.4688 - accuracy: 0.8209 - val_loss: 0.4326 - val_accuracy: 0.8182\n",
            "Epoch 137/200\n",
            "3/3 [==============================] - 0s 14ms/step - loss: 0.4665 - accuracy: 0.8209 - val_loss: 0.4302 - val_accuracy: 0.8182\n",
            "Epoch 138/200\n",
            "3/3 [==============================] - 0s 15ms/step - loss: 0.4640 - accuracy: 0.8209 - val_loss: 0.4278 - val_accuracy: 0.8182\n",
            "Epoch 139/200\n",
            "3/3 [==============================] - 0s 20ms/step - loss: 0.4616 - accuracy: 0.8358 - val_loss: 0.4255 - val_accuracy: 0.8182\n",
            "Epoch 140/200\n",
            "3/3 [==============================] - 0s 15ms/step - loss: 0.4593 - accuracy: 0.8358 - val_loss: 0.4231 - val_accuracy: 0.8485\n",
            "Epoch 141/200\n",
            "3/3 [==============================] - 0s 15ms/step - loss: 0.4569 - accuracy: 0.8507 - val_loss: 0.4208 - val_accuracy: 0.8485\n",
            "Epoch 142/200\n",
            "3/3 [==============================] - 0s 21ms/step - loss: 0.4545 - accuracy: 0.8507 - val_loss: 0.4185 - val_accuracy: 0.9091\n",
            "Epoch 143/200\n",
            "3/3 [==============================] - 0s 13ms/step - loss: 0.4522 - accuracy: 0.8507 - val_loss: 0.4163 - val_accuracy: 0.9091\n",
            "Epoch 144/200\n",
            "3/3 [==============================] - 0s 15ms/step - loss: 0.4499 - accuracy: 0.8507 - val_loss: 0.4141 - val_accuracy: 0.9091\n",
            "Epoch 145/200\n",
            "3/3 [==============================] - 0s 14ms/step - loss: 0.4477 - accuracy: 0.8507 - val_loss: 0.4118 - val_accuracy: 0.9091\n",
            "Epoch 146/200\n",
            "3/3 [==============================] - 0s 14ms/step - loss: 0.4454 - accuracy: 0.8507 - val_loss: 0.4096 - val_accuracy: 0.9091\n",
            "Epoch 147/200\n",
            "3/3 [==============================] - 0s 13ms/step - loss: 0.4432 - accuracy: 0.8507 - val_loss: 0.4073 - val_accuracy: 0.9091\n",
            "Epoch 148/200\n",
            "3/3 [==============================] - 0s 15ms/step - loss: 0.4410 - accuracy: 0.8507 - val_loss: 0.4052 - val_accuracy: 0.9091\n",
            "Epoch 149/200\n",
            "3/3 [==============================] - 0s 15ms/step - loss: 0.4388 - accuracy: 0.8507 - val_loss: 0.4031 - val_accuracy: 0.9091\n",
            "Epoch 150/200\n",
            "3/3 [==============================] - 0s 14ms/step - loss: 0.4366 - accuracy: 0.8507 - val_loss: 0.4010 - val_accuracy: 0.9091\n",
            "Epoch 151/200\n",
            "3/3 [==============================] - 0s 16ms/step - loss: 0.4345 - accuracy: 0.8507 - val_loss: 0.3990 - val_accuracy: 0.9394\n",
            "Epoch 152/200\n",
            "3/3 [==============================] - 0s 16ms/step - loss: 0.4323 - accuracy: 0.8507 - val_loss: 0.3969 - val_accuracy: 0.9394\n",
            "Epoch 153/200\n",
            "3/3 [==============================] - 0s 15ms/step - loss: 0.4301 - accuracy: 0.8507 - val_loss: 0.3948 - val_accuracy: 0.9394\n",
            "Epoch 154/200\n",
            "3/3 [==============================] - 0s 16ms/step - loss: 0.4280 - accuracy: 0.8507 - val_loss: 0.3926 - val_accuracy: 0.9394\n",
            "Epoch 155/200\n",
            "3/3 [==============================] - 0s 16ms/step - loss: 0.4258 - accuracy: 0.8507 - val_loss: 0.3905 - val_accuracy: 0.9394\n",
            "Epoch 156/200\n",
            "3/3 [==============================] - 0s 14ms/step - loss: 0.4236 - accuracy: 0.8507 - val_loss: 0.3883 - val_accuracy: 0.9394\n",
            "Epoch 157/200\n",
            "3/3 [==============================] - 0s 17ms/step - loss: 0.4214 - accuracy: 0.8507 - val_loss: 0.3862 - val_accuracy: 0.9394\n",
            "Epoch 158/200\n",
            "3/3 [==============================] - 0s 15ms/step - loss: 0.4192 - accuracy: 0.8507 - val_loss: 0.3841 - val_accuracy: 0.9394\n",
            "Epoch 159/200\n",
            "3/3 [==============================] - 0s 15ms/step - loss: 0.4172 - accuracy: 0.8507 - val_loss: 0.3821 - val_accuracy: 0.9394\n",
            "Epoch 160/200\n",
            "3/3 [==============================] - 0s 15ms/step - loss: 0.4151 - accuracy: 0.8507 - val_loss: 0.3802 - val_accuracy: 0.9394\n",
            "Epoch 161/200\n",
            "3/3 [==============================] - 0s 16ms/step - loss: 0.4131 - accuracy: 0.8507 - val_loss: 0.3783 - val_accuracy: 0.9394\n",
            "Epoch 162/200\n",
            "3/3 [==============================] - 0s 18ms/step - loss: 0.4112 - accuracy: 0.8657 - val_loss: 0.3764 - val_accuracy: 0.9394\n",
            "Epoch 163/200\n",
            "3/3 [==============================] - 0s 16ms/step - loss: 0.4093 - accuracy: 0.8657 - val_loss: 0.3744 - val_accuracy: 0.9394\n",
            "Epoch 164/200\n",
            "3/3 [==============================] - 0s 16ms/step - loss: 0.4073 - accuracy: 0.8657 - val_loss: 0.3725 - val_accuracy: 0.9394\n",
            "Epoch 165/200\n",
            "3/3 [==============================] - 0s 14ms/step - loss: 0.4053 - accuracy: 0.8657 - val_loss: 0.3705 - val_accuracy: 0.9394\n",
            "Epoch 166/200\n",
            "3/3 [==============================] - 0s 16ms/step - loss: 0.4034 - accuracy: 0.8657 - val_loss: 0.3686 - val_accuracy: 0.9394\n",
            "Epoch 167/200\n",
            "3/3 [==============================] - 0s 16ms/step - loss: 0.4014 - accuracy: 0.8657 - val_loss: 0.3666 - val_accuracy: 0.9394\n",
            "Epoch 168/200\n",
            "3/3 [==============================] - 0s 15ms/step - loss: 0.3995 - accuracy: 0.8657 - val_loss: 0.3647 - val_accuracy: 0.9394\n",
            "Epoch 169/200\n",
            "3/3 [==============================] - 0s 15ms/step - loss: 0.3976 - accuracy: 0.8806 - val_loss: 0.3629 - val_accuracy: 0.9394\n",
            "Epoch 170/200\n",
            "3/3 [==============================] - 0s 16ms/step - loss: 0.3958 - accuracy: 0.8806 - val_loss: 0.3610 - val_accuracy: 0.9394\n",
            "Epoch 171/200\n",
            "3/3 [==============================] - 0s 19ms/step - loss: 0.3939 - accuracy: 0.8806 - val_loss: 0.3591 - val_accuracy: 0.9394\n",
            "Epoch 172/200\n",
            "3/3 [==============================] - 0s 15ms/step - loss: 0.3920 - accuracy: 0.8806 - val_loss: 0.3573 - val_accuracy: 0.9394\n",
            "Epoch 173/200\n",
            "3/3 [==============================] - 0s 14ms/step - loss: 0.3901 - accuracy: 0.8806 - val_loss: 0.3555 - val_accuracy: 0.9394\n",
            "Epoch 174/200\n",
            "3/3 [==============================] - 0s 14ms/step - loss: 0.3883 - accuracy: 0.8806 - val_loss: 0.3537 - val_accuracy: 0.9394\n",
            "Epoch 175/200\n",
            "3/3 [==============================] - 0s 17ms/step - loss: 0.3864 - accuracy: 0.8806 - val_loss: 0.3519 - val_accuracy: 0.9394\n",
            "Epoch 176/200\n",
            "3/3 [==============================] - 0s 14ms/step - loss: 0.3845 - accuracy: 0.8806 - val_loss: 0.3500 - val_accuracy: 0.9394\n",
            "Epoch 177/200\n",
            "3/3 [==============================] - 0s 16ms/step - loss: 0.3826 - accuracy: 0.8806 - val_loss: 0.3482 - val_accuracy: 0.9394\n",
            "Epoch 178/200\n",
            "3/3 [==============================] - 0s 16ms/step - loss: 0.3808 - accuracy: 0.8806 - val_loss: 0.3463 - val_accuracy: 0.9394\n",
            "Epoch 179/200\n",
            "3/3 [==============================] - 0s 15ms/step - loss: 0.3788 - accuracy: 0.8955 - val_loss: 0.3445 - val_accuracy: 0.9394\n",
            "Epoch 180/200\n",
            "3/3 [==============================] - 0s 16ms/step - loss: 0.3769 - accuracy: 0.8955 - val_loss: 0.3427 - val_accuracy: 0.9394\n",
            "Epoch 181/200\n",
            "3/3 [==============================] - 0s 17ms/step - loss: 0.3751 - accuracy: 0.8955 - val_loss: 0.3409 - val_accuracy: 0.9697\n",
            "Epoch 182/200\n",
            "3/3 [==============================] - 0s 16ms/step - loss: 0.3733 - accuracy: 0.8955 - val_loss: 0.3391 - val_accuracy: 0.9697\n",
            "Epoch 183/200\n",
            "3/3 [==============================] - 0s 16ms/step - loss: 0.3715 - accuracy: 0.8955 - val_loss: 0.3374 - val_accuracy: 0.9697\n",
            "Epoch 184/200\n",
            "3/3 [==============================] - 0s 16ms/step - loss: 0.3698 - accuracy: 0.8955 - val_loss: 0.3357 - val_accuracy: 0.9697\n",
            "Epoch 185/200\n",
            "3/3 [==============================] - 0s 15ms/step - loss: 0.3681 - accuracy: 0.8955 - val_loss: 0.3341 - val_accuracy: 0.9697\n",
            "Epoch 186/200\n",
            "3/3 [==============================] - 0s 19ms/step - loss: 0.3664 - accuracy: 0.8955 - val_loss: 0.3325 - val_accuracy: 0.9697\n",
            "Epoch 187/200\n",
            "3/3 [==============================] - 0s 16ms/step - loss: 0.3647 - accuracy: 0.8955 - val_loss: 0.3310 - val_accuracy: 0.9697\n",
            "Epoch 188/200\n",
            "3/3 [==============================] - 0s 14ms/step - loss: 0.3630 - accuracy: 0.8955 - val_loss: 0.3294 - val_accuracy: 0.9697\n",
            "Epoch 189/200\n",
            "3/3 [==============================] - 0s 15ms/step - loss: 0.3613 - accuracy: 0.9254 - val_loss: 0.3279 - val_accuracy: 0.9697\n",
            "Epoch 190/200\n",
            "3/3 [==============================] - 0s 15ms/step - loss: 0.3598 - accuracy: 0.9254 - val_loss: 0.3263 - val_accuracy: 0.9697\n",
            "Epoch 191/200\n",
            "3/3 [==============================] - 0s 15ms/step - loss: 0.3581 - accuracy: 0.9254 - val_loss: 0.3248 - val_accuracy: 0.9697\n",
            "Epoch 192/200\n",
            "3/3 [==============================] - 0s 24ms/step - loss: 0.3565 - accuracy: 0.9254 - val_loss: 0.3232 - val_accuracy: 0.9697\n",
            "Epoch 193/200\n",
            "3/3 [==============================] - 0s 14ms/step - loss: 0.3549 - accuracy: 0.9254 - val_loss: 0.3217 - val_accuracy: 0.9697\n",
            "Epoch 194/200\n",
            "3/3 [==============================] - 0s 18ms/step - loss: 0.3533 - accuracy: 0.9254 - val_loss: 0.3202 - val_accuracy: 0.9697\n",
            "Epoch 195/200\n",
            "3/3 [==============================] - 0s 23ms/step - loss: 0.3518 - accuracy: 0.9254 - val_loss: 0.3187 - val_accuracy: 0.9697\n",
            "Epoch 196/200\n",
            "3/3 [==============================] - 0s 14ms/step - loss: 0.3502 - accuracy: 0.9254 - val_loss: 0.3171 - val_accuracy: 0.9697\n",
            "Epoch 197/200\n",
            "3/3 [==============================] - 0s 15ms/step - loss: 0.3486 - accuracy: 0.9254 - val_loss: 0.3154 - val_accuracy: 0.9697\n",
            "Epoch 198/200\n",
            "3/3 [==============================] - 0s 16ms/step - loss: 0.3470 - accuracy: 0.9403 - val_loss: 0.3138 - val_accuracy: 0.9697\n",
            "Epoch 199/200\n",
            "3/3 [==============================] - 0s 15ms/step - loss: 0.3453 - accuracy: 0.9403 - val_loss: 0.3122 - val_accuracy: 0.9697\n",
            "Epoch 200/200\n",
            "3/3 [==============================] - 0s 15ms/step - loss: 0.3437 - accuracy: 0.9403 - val_loss: 0.3106 - val_accuracy: 0.9697\n"
          ]
        }
      ]
    },
    {
      "cell_type": "code",
      "source": [
        "print('Train score', model.evaluate(X_train,y_train))\n",
        "print('Test score', model.evaluate(X_test,y_test))"
      ],
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "LxufoIwHMz7m",
        "outputId": "47182de4-be95-4628-a69d-3d8975dc6419"
      },
      "execution_count": 13,
      "outputs": [
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "3/3 [==============================] - 0s 9ms/step - loss: 0.3424 - accuracy: 0.9403\n",
            "Train score [0.342352032661438, 0.9402984976768494]\n",
            "2/2 [==============================] - 0s 7ms/step - loss: 0.3106 - accuracy: 0.9697\n",
            "Test score [0.3105735182762146, 0.9696969985961914]\n"
          ]
        }
      ]
    },
    {
      "cell_type": "code",
      "source": [
        "import matplotlib.pyplot as plt\n",
        "plt.plot(r.history['loss'],label ='loss')\n",
        "plt.plot(r.history['val_loss'],label ='val_loss')\n",
        "plt.legend()"
      ],
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 282
        },
        "id": "2yk1ftePOWGs",
        "outputId": "d4c469b7-f169-4e77-bc7b-20e4f00a0e58"
      },
      "execution_count": 14,
      "outputs": [
        {
          "output_type": "execute_result",
          "data": {
            "text/plain": [
              "<matplotlib.legend.Legend at 0x7f88ffc94ed0>"
            ]
          },
          "metadata": {},
          "execution_count": 14
        },
        {
          "output_type": "display_data",
          "data": {
            "text/plain": [
              "<Figure size 432x288 with 1 Axes>"
            ],
            "image/png": "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\n"
          },
          "metadata": {
            "needs_background": "light"
          }
        }
      ]
    },
    {
      "cell_type": "code",
      "source": [
        "plt.plot(r.history['accuracy'],label ='accuracy')\n",
        "plt.plot(r.history['val_accuracy'],label ='val_accuracy')\n",
        "plt.legend()"
      ],
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 286
        },
        "id": "_EkLmfedOe43",
        "outputId": "da0f042f-fd7f-4ec1-aa51-40dc2d4f5c0c"
      },
      "execution_count": 15,
      "outputs": [
        {
          "output_type": "execute_result",
          "data": {
            "text/plain": [
              "<matplotlib.legend.Legend at 0x7f88fba59c50>"
            ]
          },
          "metadata": {},
          "execution_count": 15
        },
        {
          "output_type": "display_data",
          "data": {
            "text/plain": [
              "<Figure size 432x288 with 1 Axes>"
            ],
            "image/png": "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\n"
          },
          "metadata": {
            "needs_background": "light"
          }
        }
      ]
    },
    {
      "cell_type": "code",
      "source": [
        "P = model.predict(X_test)\n",
        "print(P)"
      ],
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "MehtQGULwzxu",
        "outputId": "46a342c3-cd97-40f3-ca53-b40022c2a6ea"
      },
      "execution_count": 16,
      "outputs": [
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "[[0.90195   ]\n",
            " [0.38037896]\n",
            " [0.6149642 ]\n",
            " [0.32924587]\n",
            " [0.31949356]\n",
            " [0.58183837]\n",
            " [0.2543989 ]\n",
            " [0.21764669]\n",
            " [0.16980287]\n",
            " [0.73259175]\n",
            " [0.2289263 ]\n",
            " [0.28729182]\n",
            " [0.9049752 ]\n",
            " [0.6375998 ]\n",
            " [0.8744104 ]\n",
            " [0.20970899]\n",
            " [0.17934638]\n",
            " [0.12293184]\n",
            " [0.71177244]\n",
            " [0.24923897]\n",
            " [0.4590645 ]\n",
            " [0.65620065]\n",
            " [0.1665904 ]\n",
            " [0.08096433]\n",
            " [0.19530743]\n",
            " [0.77109456]\n",
            " [0.4103527 ]\n",
            " [0.27140516]\n",
            " [0.89201015]\n",
            " [0.3600503 ]\n",
            " [0.8790517 ]\n",
            " [0.08704317]\n",
            " [0.36372024]]\n"
          ]
        }
      ]
    },
    {
      "cell_type": "code",
      "source": [
        "import numpy as np\n",
        "\n",
        "P = np.round(P).flatten()\n",
        "print(P)"
      ],
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "mkHlaOzYxPZY",
        "outputId": "3ab12980-bc6b-49fe-83cd-b1192d144672"
      },
      "execution_count": 20,
      "outputs": [
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "[1. 0. 1. 0. 0. 1. 0. 0. 0. 1. 0. 0. 1. 1. 1. 0. 0. 0. 1. 0. 0. 1. 0. 0.\n",
            " 0. 1. 0. 0. 1. 0. 1. 0. 0.]\n"
          ]
        }
      ]
    },
    {
      "cell_type": "code",
      "source": [
        "print(\"Mannualy calculated accuracy:\", np.mean(P == y_test))\n",
        "print(\"Evaluate output:\", model.evaluate(X_test, y_test))"
      ],
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "MPqfWuzlzcPh",
        "outputId": "d10ca2fe-4104-4c24-8087-94fef2454a83"
      },
      "execution_count": 21,
      "outputs": [
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "Mannualy calculated accuracy: 0.9696969696969697\n",
            "2/2 [==============================] - 0s 7ms/step - loss: 0.3106 - accuracy: 0.9697\n",
            "Evaluate output: [0.3105735182762146, 0.9696969985961914]\n"
          ]
        }
      ]
    }
  ]
}