MathFundamentalsOfNN_first_look.ipynb

mathfundamentalsofnn_first_look.ipynb

{
  "nbformat": 4,
  "nbformat_minor": 0,
  "metadata": {
    "colab": {
      "name": "MathFundamentalsOfNN_first_look.ipynb",
      "provenance": [],
      "collapsed_sections": [],
      "authorship_tag": "ABX9TyM7nqBzDp8opubbA2LtJyn1",
      "include_colab_link": true
    },
    "kernelspec": {
      "name": "python3",
      "display_name": "Python 3"
    },
    "language_info": {
      "name": "python"
    },
    "gpuClass": "standard"
  },
  "cells": [
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "view-in-github",
        "colab_type": "text"
      },
      "source": [
        "<a href=\"https://colab.research.google.com/gist/rhiskey/6b0575028fb2c33e06cb1f5d996e517c/mathfundamentalsofnn_first_look.ipynb\" target=\"_parent\"><img src=\"https://colab.research.google.com/assets/colab-badge.svg\" alt=\"Open In Colab\"/></a>"
      ]
    },
    {
      "cell_type": "markdown",
      "source": [
        "#Математические основы нейронных сетей\n",
        "##Первое знакомство с нейросетью"
      ],
      "metadata": {
        "id": "fORbbiaBB_r_"
      }
    },
    {
      "cell_type": "markdown",
      "source": [
        "Загрузка набора данных MNIST в Keras"
      ],
      "metadata": {
        "id": "3EqwS3CCCHmu"
      }
    },
    {
      "cell_type": "code",
      "execution_count": 1,
      "metadata": {
        "id": "PT-5-MLLB20C"
      },
      "outputs": [],
      "source": [
        "from tensorflow.keras.datasets import mnist\n",
        "(train_images, train_labels), (test_images, test_labels) = mnist.load_data()"
      ]
    },
    {
      "cell_type": "code",
      "source": [
        "import matplotlib.pyplot as plt\n",
        "%matplotlib inline\n",
        "\n",
        "num = 4\n",
        "images = train_images[:num]\n",
        "labels = train_labels[:num]\n",
        "\n",
        "fig, axes = plt.subplots(1, num, figsize=(1.5*num,2*num))\n",
        "for i in range(num):\n",
        "    ax = axes[i]\n",
        "    ax.imshow(images[i], cmap='gray')\n",
        "    ax.set_title('Метка: {}'.format(labels[i]))\n",
        "plt.tight_layout()\n",
        "plt.show()"
      ],
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 139
        },
        "id": "TaoNNIgBCzDZ",
        "outputId": "8fd011be-da81-4ce6-dabd-cb05df3a6aa7"
      },
      "execution_count": 2,
      "outputs": [
        {
          "output_type": "display_data",
          "data": {
            "text/plain": [
              "<Figure size 432x576 with 4 Axes>"
            ],
            "image/png": "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\n"
          },
          "metadata": {
            "needs_background": "light"
          }
        }
      ]
    },
    {
      "cell_type": "markdown",
      "source": [
        "Первое знакомство с нейросетью и загрузка набора данных MNIST"
      ],
      "metadata": {
        "id": "ePt7458DSFkq"
      }
    },
    {
      "cell_type": "code",
      "source": [
        "from tensorflow.keras.datasets import mnist\n",
        "(train_images, train_labels), (test_images, test_labels) = mnist.load_data()"
      ],
      "metadata": {
        "id": "MqrZ31sBSNCl"
      },
      "execution_count": 3,
      "outputs": []
    },
    {
      "cell_type": "code",
      "source": [
        "train_images.shape"
      ],
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "faOrrzRBVsp0",
        "outputId": "5c488857-2acd-40fc-9857-b1a5ebf2ce45"
      },
      "execution_count": 4,
      "outputs": [
        {
          "output_type": "execute_result",
          "data": {
            "text/plain": [
              "(60000, 28, 28)"
            ]
          },
          "metadata": {},
          "execution_count": 4
        }
      ]
    },
    {
      "cell_type": "code",
      "source": [
        "len(train_images)"
      ],
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "yr-Iqtf1Vv_7",
        "outputId": "f3190a18-d419-4fa8-ac12-c480d8139436"
      },
      "execution_count": 5,
      "outputs": [
        {
          "output_type": "execute_result",
          "data": {
            "text/plain": [
              "60000"
            ]
          },
          "metadata": {},
          "execution_count": 5
        }
      ]
    },
    {
      "cell_type": "code",
      "source": [
        "test_images.shape"
      ],
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "hgMmpnTeWBm1",
        "outputId": "867242a0-1912-46c6-ece6-880dbc6e2d48"
      },
      "execution_count": 6,
      "outputs": [
        {
          "output_type": "execute_result",
          "data": {
            "text/plain": [
              "(10000, 28, 28)"
            ]
          },
          "metadata": {},
          "execution_count": 6
        }
      ]
    },
    {
      "cell_type": "code",
      "source": [
        "train_labels"
      ],
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "K19J_-IvWGsp",
        "outputId": "07dd180c-62b3-410b-ed63-ce2af23ed617"
      },
      "execution_count": 7,
      "outputs": [
        {
          "output_type": "execute_result",
          "data": {
            "text/plain": [
              "array([5, 0, 4, ..., 5, 6, 8], dtype=uint8)"
            ]
          },
          "metadata": {},
          "execution_count": 7
        }
      ]
    },
    {
      "cell_type": "code",
      "source": [
        "len(test_images)"
      ],
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "jHX1k0q8WJNk",
        "outputId": "036a3a00-e024-4dfe-b1c7-bf7dd860a2cc"
      },
      "execution_count": 8,
      "outputs": [
        {
          "output_type": "execute_result",
          "data": {
            "text/plain": [
              "10000"
            ]
          },
          "metadata": {},
          "execution_count": 8
        }
      ]
    },
    {
      "cell_type": "code",
      "source": [
        "test_labels"
      ],
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "MCifMYrbWNEn",
        "outputId": "0da58b48-8a73-4261-e210-dae50de4bd77"
      },
      "execution_count": 9,
      "outputs": [
        {
          "output_type": "execute_result",
          "data": {
            "text/plain": [
              "array([7, 2, 1, ..., 4, 5, 6], dtype=uint8)"
            ]
          },
          "metadata": {},
          "execution_count": 9
        }
      ]
    },
    {
      "cell_type": "markdown",
      "source": [
        "### Архитектура нейросети"
      ],
      "metadata": {
        "id": "M_D_x8LAKhc8"
      }
    },
    {
      "cell_type": "code",
      "source": [
        "from tensorflow import keras\n",
        "from tensorflow.keras import layers\n",
        "\n",
        "model = keras.Sequential([\n",
        "  layers.Dense(512, activation=\"relu\"),\n",
        "  layers.Dense(10, activation=\"softmax\")\n",
        "])"
      ],
      "metadata": {
        "id": "_EpVhdQBKjis"
      },
      "execution_count": 10,
      "outputs": []
    },
    {
      "cell_type": "markdown",
      "source": [
        "### Компиляция"
      ],
      "metadata": {
        "id": "vJ2dQtbKKorF"
      }
    },
    {
      "cell_type": "code",
      "source": [
        "model.compile(optimizer=\"rmsprop\",\n",
        "              loss=\"sparse_categorical_crossentropy\",\n",
        "              metrics=[\"accuracy\"])"
      ],
      "metadata": {
        "id": "jqmrnSG5Ktrr"
      },
      "execution_count": 11,
      "outputs": []
    },
    {
      "cell_type": "markdown",
      "source": [
        "### Подготовка графических данных"
      ],
      "metadata": {
        "id": "y7ZSRLgEKtDb"
      }
    },
    {
      "cell_type": "code",
      "source": [
        "train_images = train_images.reshape((60000, 28*28))\n",
        "train_images = train_images.astype(\"float32\")/255\n",
        "\n",
        "test_images = test_images.reshape((10000, 28*28))\n",
        "test_images = test_images.astype(\"float32\")/255"
      ],
      "metadata": {
        "id": "WFV6sW0_K5NT"
      },
      "execution_count": 12,
      "outputs": []
    },
    {
      "cell_type": "markdown",
      "source": [
        "### \"Обучение\" модели"
      ],
      "metadata": {
        "id": "7JmmNtVeK5gu"
      }
    },
    {
      "cell_type": "code",
      "source": [
        "model.fit(train_images, train_labels, epochs=5, batch_size=128)"
      ],
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "PP6Zi7t8SU5D",
        "outputId": "3e19d22e-81e3-42f2-c7ca-8d9f65599bf9"
      },
      "execution_count": 13,
      "outputs": [
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "Epoch 1/5\n",
            "469/469 [==============================] - 7s 14ms/step - loss: 0.2567 - accuracy: 0.9261\n",
            "Epoch 2/5\n",
            "469/469 [==============================] - 10s 21ms/step - loss: 0.1044 - accuracy: 0.9693\n",
            "Epoch 3/5\n",
            "469/469 [==============================] - 7s 16ms/step - loss: 0.0687 - accuracy: 0.9796\n",
            "Epoch 4/5\n",
            "469/469 [==============================] - 8s 17ms/step - loss: 0.0501 - accuracy: 0.9849\n",
            "Epoch 5/5\n",
            "469/469 [==============================] - 6s 13ms/step - loss: 0.0370 - accuracy: 0.9889\n"
          ]
        },
        {
          "output_type": "execute_result",
          "data": {
            "text/plain": [
              "<keras.callbacks.History at 0x7fe1030043d0>"
            ]
          },
          "metadata": {},
          "execution_count": 13
        }
      ]
    },
    {
      "cell_type": "markdown",
      "source": [
        "### Использование модели для осуществления предсказаний"
      ],
      "metadata": {
        "id": "LhhNEqXqSV4X"
      }
    },
    {
      "cell_type": "code",
      "source": [
        "test_digits = test_images[0:10]\n",
        "predictions = model.predict(test_digits)\n",
        "predictions[0]"
      ],
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "ydOSOZdaSZPx",
        "outputId": "ecb01668-7ec3-4e09-c803-f75b6b2abf79"
      },
      "execution_count": 14,
      "outputs": [
        {
          "output_type": "execute_result",
          "data": {
            "text/plain": [
              "array([1.0593984e-08, 5.7342181e-10, 4.9859304e-06, 2.9029541e-05,\n",
              "       1.9924337e-10, 5.1484651e-08, 5.4766508e-14, 9.9995840e-01,\n",
              "       2.6315294e-07, 7.2423331e-06], dtype=float32)"
            ]
          },
          "metadata": {},
          "execution_count": 14
        }
      ]
    },
    {
      "cell_type": "code",
      "source": [
        "predictions[0].argmax()"
      ],
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "7LP1P_CEYa9m",
        "outputId": "91dc5392-ffeb-4d5e-e4ea-8cdf5af6a1ad"
      },
      "execution_count": 15,
      "outputs": [
        {
          "output_type": "execute_result",
          "data": {
            "text/plain": [
              "7"
            ]
          },
          "metadata": {},
          "execution_count": 15
        }
      ]
    },
    {
      "cell_type": "code",
      "source": [
        "predictions[0][7]"
      ],
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "aUZP_taZYkQ6",
        "outputId": "cb2b5584-bbbd-441c-a035-bdfb5de04e95"
      },
      "execution_count": 16,
      "outputs": [
        {
          "output_type": "execute_result",
          "data": {
            "text/plain": [
              "0.9999584"
            ]
          },
          "metadata": {},
          "execution_count": 16
        }
      ]
    },
    {
      "cell_type": "code",
      "source": [
        "test_labels[0]"
      ],
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "ehuyAhhZYpEj",
        "outputId": "5cafd075-fa50-4e72-98d7-8252b1806f6d"
      },
      "execution_count": 17,
      "outputs": [
        {
          "output_type": "execute_result",
          "data": {
            "text/plain": [
              "7"
            ]
          },
          "metadata": {},
          "execution_count": 17
        }
      ]
    },
    {
      "cell_type": "markdown",
      "source": [
        "### Применение модели на новых данных"
      ],
      "metadata": {
        "id": "jyJ0vKAPSbkq"
      }
    },
    {
      "cell_type": "code",
      "source": [
        "test_loss, test_acc = model.evaluate(test_images, test_labels)\n",
        "print(f\"test_acc: {test_acc}\")"
      ],
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "oapvYX94SkTs",
        "outputId": "9362a649-3f3b-48ea-d8d1-e6516346066d"
      },
      "execution_count": 18,
      "outputs": [
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "313/313 [==============================] - 1s 2ms/step - loss: 0.0725 - accuracy: 0.9776\n",
            "test_acc: 0.9775999784469604\n"
          ]
        }
      ]
    }
  ]
}