MNIST_learn_by_FC.ipynb

{
  "nbformat": 4,
  "nbformat_minor": 0,
  "metadata": {
    "colab": {
      "provenance": [],
      "authorship_tag": "ABX9TyPZMKaQDH9vvVRIcVtBK7Yq",
      "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/stepka123/basic-NL-SA/blob/main/MNIST_learn_by_FC.ipynb\" target=\"_parent\"><img src=\"https://colab.research.google.com/assets/colab-badge.svg\" alt=\"Open In Colab\"/></a>"
      ]
    },
    {
      "cell_type": "markdown",
      "source": [
        "MNIST датасет. Определение рукописного ввода цифр. Полносвязная модель на 754 первого, 254 второго и 10 третьего(выходного) слоев\n"
      ],
      "metadata": {
        "id": "AEtylf-ctGcY"
      }
    },
    {
      "cell_type": "code",
      "execution_count": 1,
      "metadata": {
        "id": "BBSSnFSrtDLQ"
      },
      "outputs": [],
      "source": [
        "import matplotlib.pyplot as plt\n",
        "from tensorflow import keras\n",
        "from tensorflow.keras.layers import Dense\n",
        "from tensorflow.keras.layers import Flatten\n",
        "import numpy as np\n",
        "import scapy as sc\n",
        "from tensorflow.keras.datasets import mnist"
      ]
    },
    {
      "cell_type": "markdown",
      "source": [
        "показывает первые 25 цифр из первичного датасета"
      ],
      "metadata": {
        "id": "ckioIwDr2kwG"
      }
    },
    {
      "cell_type": "code",
      "source": [
        "#Загружаем датасет 60к изобрадений.\n",
        "#1-й вектор: х -изображение цифр обучающей выборки, у - вектор соответствия который проверяется на выходе т.к обучение с учителем\n",
        "#2-й вектор: х -изображение цифр тестовой выборки. у -тоже самое жля тестовой выборки\n",
        "(x_train, y_train), (x_test, y_test) = mnist.load_data()\n",
        "\n",
        "#стандартизация входных данных, для получение значения каждого пикселя от 0 до 1(отображение )\n",
        "x_train = x_train / 255\n",
        "x_test = x_test / 255\n",
        "\n",
        "plt.figure(figsize=(15, 10))\n",
        "for i in range(25):\n",
        "  plt.subplot(5,5, i+1)\n",
        "  plt.imshow(x_train[i], cmap=plt.cm.binary)\n",
        "\n",
        "plt.show()"
      ],
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 833
        },
        "id": "LPVdZWOkuvh7",
        "outputId": "567d9301-ace1-45b6-f306-fbbf09efcd4f"
      },
      "execution_count": 2,
      "outputs": [
        {
          "output_type": "display_data",
          "data": {
            "text/plain": [
              "<Figure size 1500x1000 with 25 Axes>"
            ],
            "image/png": "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\n"
          },
          "metadata": {}
        }
      ]
    },
    {
      "cell_type": "code",
      "source": [
        "#создается модель\n",
        "model = keras.Sequential([\n",
        "  Flatten(input_shape=(28, 28, 1)),#1 - байт = 1 пиксель в градации серого 28 - строк,28-столбцов это изображение\n",
        "  Dense(256, activation='relu'), #создается слой из 254(можно любое кол-во) нейронов и 28*28 входных точек скрытый слой\n",
        "  Dense(154, activation='relu'),\n",
        "  #создается выходной слой из 10 нейронов  т.к цифр всего 0-9 всего 10 каждому выходу свой номер, а активация будет давать й на выход\n",
        "  Dense(10, activation='softmax')\n",
        "])\n",
        "#кол-во параметров слоя(кол-во весов) = кол-во входных связей(предыдущих нейронов(предыдущего слоя)) * кол-во нейронов текущего слоя\n",
        "#т.е первый слой 28*28 - 754 + 1 биас\n",
        "# второй слой это 256 нейронов, т.е 754*256+1 - 200960 параметров(связей между первым и вторым слоем) так для третьего тоже\n",
        "\n",
        "print(model.summary())"
      ],
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "cR2otmVy46WY",
        "outputId": "e547e13b-edbe-489d-9a74-497db1c94c56"
      },
      "execution_count": 16,
      "outputs": [
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "Model: \"sequential_5\"\n",
            "_________________________________________________________________\n",
            " Layer (type)                Output Shape              Param #   \n",
            "=================================================================\n",
            " flatten_5 (Flatten)         (None, 784)               0         \n",
            "                                                                 \n",
            " dense_25 (Dense)            (None, 256)               200960    \n",
            "                                                                 \n",
            " dense_26 (Dense)            (None, 154)               39578     \n",
            "                                                                 \n",
            " dense_27 (Dense)            (None, 10)                1550      \n",
            "                                                                 \n",
            "=================================================================\n",
            "Total params: 242088 (945.66 KB)\n",
            "Trainable params: 242088 (945.66 KB)\n",
            "Non-trainable params: 0 (0.00 Byte)\n",
            "_________________________________________________________________\n",
            "None\n"
          ]
        }
      ]
    },
    {
      "cell_type": "code",
      "source": [
        "#преобразуем выходные данные в вектора(наборы [1000000000] = это 0, т.к стоит в начале, от куда и начинается счет,\n",
        "#да примитивно, но что поделаешь, я бы взял 4 бита из двоичной системы и смотрел по ней) по 10 значений 0-9 в виде 0 и 1, где 1 будет означать цифру\n",
        "y_train_cat = keras.utils.to_categorical(y_train, 10)\n",
        "y_test_cat = keras.utils.to_categorical(y_test, 10)"
      ],
      "metadata": {
        "id": "ku2iZQu2At8I"
      },
      "execution_count": 5,
      "outputs": []
    },
    {
      "cell_type": "markdown",
      "source": [
        "Компиляция нс с оптимизатором и критерием категории кросс-энтропией"
      ],
      "metadata": {
        "id": "CdggyQnyCLml"
      }
    },
    {
      "cell_type": "code",
      "source": [
        "model.compile(optimizer='adam', loss='categorical_crossentropy', metrics=['accuracy'])\n",
        "#запуск процеса обучения ф-я fit\n",
        "model.fit(x_train, y_train_cat, batch_size=32, epochs=5, validation_split=0.2)\n",
        "#x_train - входное множество,\n",
        "#у_train_cat - требуемое выходное значение в виде 0 и 1,\n",
        "#batch_size - размер батча (после каждого 32-ой картинки корректируются веса)\n",
        "#epochs - эпохи обучения(круги),\n",
        "#validation_split=0.2 - делитобучающую выборку (train) на обучающую и проверочную. Т.е 0.2 - 20% картинок из обучающей"
      ],
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "OcXcZzRJCEen",
        "outputId": "a61382aa-daf3-4ae9-abc7-cc07a3845799"
      },
      "execution_count": 18,
      "outputs": [
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "Epoch 1/5\n",
            "1500/1500 [==============================] - 20s 12ms/step - loss: 0.0621 - accuracy: 0.9838 - val_loss: 0.0931 - val_accuracy: 0.9747\n",
            "Epoch 2/5\n",
            "1500/1500 [==============================] - 14s 9ms/step - loss: 0.0406 - accuracy: 0.9881 - val_loss: 0.0915 - val_accuracy: 0.9749\n",
            "Epoch 3/5\n",
            "1500/1500 [==============================] - 13s 9ms/step - loss: 0.0308 - accuracy: 0.9901 - val_loss: 0.0966 - val_accuracy: 0.9767\n",
            "Epoch 4/5\n",
            "1500/1500 [==============================] - 15s 10ms/step - loss: 0.0252 - accuracy: 0.9923 - val_loss: 0.0893 - val_accuracy: 0.9783\n",
            "Epoch 5/5\n",
            "1500/1500 [==============================] - 13s 9ms/step - loss: 0.0193 - accuracy: 0.9939 - val_loss: 0.1014 - val_accuracy: 0.9761\n"
          ]
        },
        {
          "output_type": "execute_result",
          "data": {
            "text/plain": [
              "<keras.src.callbacks.History at 0x7ab3cd464c10>"
            ]
          },
          "metadata": {},
          "execution_count": 18
        }
      ]
    },
    {
      "cell_type": "code",
      "source": [
        "#запускаем обучение по тестовой метрике loss - критерий качества, accuracy - процент правильно распознанных изображений\n",
        "model.evaluate(x_test, y_test_cat)"
      ],
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "nGlJNlPfG2cT",
        "outputId": "d96c1f34-9134-4d0c-a6bc-390f9a850b0c"
      },
      "execution_count": 19,
      "outputs": [
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "313/313 [==============================] - 1s 4ms/step - loss: 0.0895 - accuracy: 0.9809\n"
          ]
        },
        {
          "output_type": "execute_result",
          "data": {
            "text/plain": [
              "[0.0895262360572815, 0.98089998960495]"
            ]
          },
          "metadata": {},
          "execution_count": 19
        }
      ]
    },
    {
      "cell_type": "markdown",
      "source": [
        "Теперь нс обучена и расспознает цифры с вероятностью успеха в 97.8%\n"
      ],
      "metadata": {
        "id": "NjIn1dEEHvsF"
      }
    },
    {
      "cell_type": "code",
      "source": [
        "n = 9\n",
        "\n",
        "x = np.expand_dims(x_test[n], axis=0)#axis - это добавляем в двумерную матрицу третье измерение, x_train[n]-двумерная матрица\n",
        "#т.к передаем несколько изображений нужно подавать в виде трехмерного тензора, каждым элементом тензора будет изображение n\n",
        "res = model.predict(x)\n",
        "print(res)\n",
        "print(f\"Цифра: {np.argmax(res)}\")\n",
        "\n",
        "plt.imshow(x_test[n], cmap=plt.cm.binary)\n",
        "plt.show()"
      ],
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 503
        },
        "id": "XPeghI7qHoTC",
        "outputId": "9d4c874c-2a93-41d7-e04e-5259d7f0d776"
      },
      "execution_count": 20,
      "outputs": [
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "1/1 [==============================] - 0s 84ms/step\n",
            "[[1.5243593e-13 3.0011021e-11 2.6738419e-11 1.4084025e-08 2.0209940e-01\n",
            "  1.7478507e-12 5.2882216e-13 1.3789013e-07 5.8470402e-11 7.9790044e-01]]\n",
            "Цифра: 9\n"
          ]
        },
        {
          "output_type": "display_data",
          "data": {
            "text/plain": [
              "<Figure size 640x480 with 1 Axes>"
            ],
            "image/png": "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\n"
          },
          "metadata": {}
        }
      ]
    }
  ]
}