{
  "nbformat": 4,
  "nbformat_minor": 0,
  "metadata": {
    "colab": {
      "provenance": []
    },
    "kernelspec": {
      "name": "python3",
      "display_name": "Python 3"
    },
    "language_info": {
      "name": "python"
    }
  },
  "cells": [
    {
      "cell_type": "code",
      "execution_count": 209,
      "metadata": {
        "id": "lbvH8mqmcyXu"
      },
      "outputs": [],
      "source": [
        "import numpy as np\n",
        "import matplotlib.pyplot as plt\n",
        "from tensorflow.keras.datasets import fashion_mnist\n",
        "from tensorflow import keras\n",
        "import keras.backend as K\n",
        "from tensorflow.keras.layers import Dense, Flatten, Reshape, Input, Lambda, BatchNormalization, Dropout\n",
        "from tensorflow.keras.utils import plot_model"
      ]
    },
    {
      "cell_type": "code",
      "source": [
        "reflist = [0, 0, 0, 0, 0, 1, 0, 1, 2, 1]\n",
        "\n",
        "def change_classes(l):\n",
        "  res = []\n",
        "  for i in range(len(l)):\n",
        "    res.append(reflist[l[i]])\n",
        "\n",
        "  return np.array(res)"
      ],
      "metadata": {
        "id": "c3t97S1ugxY7"
      },
      "execution_count": 210,
      "outputs": []
    },
    {
      "cell_type": "code",
      "source": [
        "(x_train, y_train), (x_test, y_test) = fashion_mnist.load_data()\n",
        "\n",
        "x_train = x_train / 255\n",
        "x_test = x_test / 255\n",
        "\n",
        "y_test = change_classes(y_test)\n",
        "y_test_categ = keras.utils.to_categorical(y_test, 3)\n",
        "\n",
        "y_train = change_classes(y_train)\n",
        "y_train_categ = keras.utils.to_categorical(y_train, 3)"
      ],
      "metadata": {
        "id": "m1rqqP9ldFV5"
      },
      "execution_count": 211,
      "outputs": []
    },
    {
      "cell_type": "code",
      "source": [
        "pivot = int(len(x_train) * 0.8)\n",
        "\n",
        "x_val = x_train[pivot:]\n",
        "y_val = y_train_categ[pivot:]\n",
        "\n",
        "x_train = x_train[:pivot]\n",
        "y_train_categ = y_train_categ[:pivot]\n",
        "y_train = y_train[:pivot]"
      ],
      "metadata": {
        "id": "eSJvKV6axaeu"
      },
      "execution_count": 212,
      "outputs": []
    },
    {
      "cell_type": "code",
      "source": [
        "input_img = Input((28, 28))\n",
        "x = Flatten()(input_img)\n",
        "x = Dense(256, activation='relu')(x)\n",
        "x = Dense(100, activation='relu')(x)\n",
        "x = Dense(50, activation='relu')(x)\n",
        "ans = Dense(3, activation='softmax')(x)"
      ],
      "metadata": {
        "id": "JapzHASDdRPN"
      },
      "execution_count": 213,
      "outputs": []
    },
    {
      "cell_type": "code",
      "source": [
        "model = keras.Model(input_img, ans)\n",
        "model.compile(optimizer='adam', loss = 'categorical_crossentropy')"
      ],
      "metadata": {
        "id": "KQzpI89rdidC"
      },
      "execution_count": 214,
      "outputs": []
    },
    {
      "cell_type": "code",
      "source": [
        "model.fit(x_train, y_train_categ, epochs = 10, batch_size = 30, validation_data=(x_val, y_val), shuffle=True)"
      ],
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "qqAef6VAdllB",
        "outputId": "f66d0f5b-7607-4a21-ab7b-6cca05c76af1"
      },
      "execution_count": 219,
      "outputs": [
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "Epoch 1/10\n",
            "1600/1600 [==============================] - 6s 3ms/step - loss: 0.0508 - val_loss: 0.0313\n",
            "Epoch 2/10\n",
            "1600/1600 [==============================] - 6s 4ms/step - loss: 0.0292 - val_loss: 0.0249\n",
            "Epoch 3/10\n",
            "1600/1600 [==============================] - 5s 3ms/step - loss: 0.0229 - val_loss: 0.0229\n",
            "Epoch 4/10\n",
            "1600/1600 [==============================] - 6s 3ms/step - loss: 0.0185 - val_loss: 0.0267\n",
            "Epoch 5/10\n",
            "1600/1600 [==============================] - 5s 3ms/step - loss: 0.0162 - val_loss: 0.0726\n",
            "Epoch 6/10\n",
            "1600/1600 [==============================] - 6s 4ms/step - loss: 0.0144 - val_loss: 0.0214\n",
            "Epoch 7/10\n",
            "1600/1600 [==============================] - 5s 3ms/step - loss: 0.0122 - val_loss: 0.0200\n",
            "Epoch 8/10\n",
            "1600/1600 [==============================] - 6s 3ms/step - loss: 0.0123 - val_loss: 0.0257\n",
            "Epoch 9/10\n",
            "1600/1600 [==============================] - 5s 3ms/step - loss: 0.0099 - val_loss: 0.0355\n",
            "Epoch 10/10\n",
            "1600/1600 [==============================] - 5s 3ms/step - loss: 0.0111 - val_loss: 0.0241\n"
          ]
        },
        {
          "output_type": "execute_result",
          "data": {
            "text/plain": [
              "<keras.callbacks.History at 0x7fe001f21b70>"
            ]
          },
          "metadata": {},
          "execution_count": 219
        }
      ]
    },
    {
      "cell_type": "code",
      "source": [
        "model.evaluate(x_test, y_test_categ)"
      ],
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "hKJZBPgrd6sV",
        "outputId": "fb2ddd2b-eb68-4d90-a2da-bc94a2a03ad5"
      },
      "execution_count": 224,
      "outputs": [
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "313/313 [==============================] - 0s 1ms/step - loss: 0.0274\n"
          ]
        },
        {
          "output_type": "execute_result",
          "data": {
            "text/plain": [
              "0.02739698439836502"
            ]
          },
          "metadata": {},
          "execution_count": 224
        }
      ]
    },
    {
      "cell_type": "code",
      "source": [
        "res = model.predict(x_test)\n",
        "res = np.array([np.argmax(r) for r in res])"
      ],
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "YBOG-gwXxFTh",
        "outputId": "932b8095-e311-4465-c90d-83c6c28fc1df"
      },
      "execution_count": 225,
      "outputs": [
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "313/313 [==============================] - 0s 1ms/step\n"
          ]
        }
      ]
    },
    {
      "cell_type": "code",
      "source": [
        "right_count = 0\n",
        "\n",
        "for i in range(len(y_test)):\n",
        "  if res[i] == y_test[i]:\n",
        "    right_count += 1\n",
        "\n",
        "print(f\"Процент верно предсказаных: {right_count / len(res)}\")"
      ],
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "MKgK5UCLrkHP",
        "outputId": "70818507-8cdd-4d22-df56-fc47f150dcc0"
      },
      "execution_count": 226,
      "outputs": [
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "Процент верно предсказаных: 0.9939\n"
          ]
        }
      ]
    },
    {
      "cell_type": "code",
      "source": [
        "plot_model(model, expand_nested=True, show_shapes = True, show_layer_names=False, dpi = 70)"
      ],
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 474
        },
        "id": "yIlfzCqs0aks",
        "outputId": "f5153e31-3e11-482d-cdb7-16c46699eefd"
      },
      "execution_count": 227,
      "outputs": [
        {
          "output_type": "execute_result",
          "data": {
            "image/png": "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\n",
            "text/plain": [
              "<IPython.core.display.Image object>"
            ]
          },
          "metadata": {},
          "execution_count": 227
        }
      ]
    }
  ]
}