DigitClassification.ipynb

{
  "nbformat": 4,
  "nbformat_minor": 0,
  "metadata": {
    "colab": {
      "name": "DigitClassification.ipynb",
      "provenance": [],
      "collapsed_sections": []
    },
    "kernelspec": {
      "name": "python3",
      "display_name": "Python 3"
    },
    "language_info": {
      "name": "python"
    }
  },
  "cells": [
    {
      "cell_type": "code",
      "execution_count": 2,
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "n7a_q6TVQn9q",
        "outputId": "52483747-3cfb-480a-ea0c-68226af3d03c"
      },
      "outputs": [
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "Downloading data from https://storage.googleapis.com/tensorflow/tf-keras-datasets/mnist.npz\n",
            "11493376/11490434 [==============================] - 0s 0us/step\n",
            "11501568/11490434 [==============================] - 0s 0us/step\n"
          ]
        },
        {
          "output_type": "execute_result",
          "data": {
            "text/plain": [
              "(60000, 28, 28)"
            ]
          },
          "metadata": {},
          "execution_count": 2
        }
      ],
      "source": [
        "import tensorflow as tf\n",
        "(x_train, y_train), (x_test, y_test) = tf.keras.datasets.mnist.load_data()\n",
        "x_train.shape"
      ]
    },
    {
      "cell_type": "code",
      "source": [
        "import matplotlib.pyplot as plt\n",
        "%matplotlib inline\n",
        "fig, axs = plt.subplots(4, 4, figsize = (20,20))\n",
        "plt.gray()\n",
        "for i , ax in enumerate(axs.flat):\n",
        "    ax.matshow(x_train[i])\n",
        "    ax.axis('off')\n",
        "    ax.set_title('Number {}'.format(y_train[i]))\n",
        "fig.show()"
      ],
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 1000
        },
        "id": "ly4u8AUyRiZj",
        "outputId": "bb5bb457-e012-46ee-a010-e1b33277bb21"
      },
      "execution_count": 3,
      "outputs": [
        {
          "output_type": "display_data",
          "data": {
            "text/plain": [
              "<Figure size 1440x1440 with 16 Axes>"
            ],
            "image/png": "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\n"
          },
          "metadata": {
            "needs_background": "light"
          }
        }
      ]
    },
    {
      "cell_type": "code",
      "source": [
        "x_train = x_train.reshape(x_train.shape[0], 28, 28, 1)\n",
        "x_test = x_test.reshape(x_test.shape[0], 28, 28, 1)\n",
        "input_shape = (28, 28,1)"
      ],
      "metadata": {
        "id": "CNHX4Z0nR7vP"
      },
      "execution_count": 4,
      "outputs": []
    },
    {
      "cell_type": "code",
      "source": [
        "x_train = x_train.astype('float32')\n",
        "x_test = x_test.astype('float32')\n",
        "x_train /= 255\n",
        "x_test /= 255\n",
        "print('x_train shape:', x_train.shape)\n",
        "print('Number of images in x_train', x_train.shape[0])\n",
        "print('Number of images in x_test', x_test.shape[0])"
      ],
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "6gEDhk2IT6VP",
        "outputId": "9d8042d5-b4c1-4310-9698-cbe8024569c8"
      },
      "execution_count": 5,
      "outputs": [
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "x_train shape: (60000, 28, 28, 1)\n",
            "Number of images in x_train 60000\n",
            "Number of images in x_test 10000\n"
          ]
        }
      ]
    },
    {
      "cell_type": "code",
      "source": [
        "from tensorflow.keras.models import Sequential\n",
        "from tensorflow.keras.layers import Dense, Conv2D, Dropout, Flatten, MaxPooling2D\n",
        "model = Sequential()\n",
        "model.add(Conv2D(28, kernel_size=(3,3), input_shape=input_shape))\n",
        "model.add(MaxPooling2D(pool_size=(2,2)))\n",
        "model.add(Flatten())\n",
        "model.add(Dense(128, activation=tf.nn.relu))\n",
        "model.add(Dropout(0.2))\n",
        "model.add(Dense(10,activation=tf.nn.softmax))"
      ],
      "metadata": {
        "id": "wI0U3-aiUcbv"
      },
      "execution_count": 6,
      "outputs": []
    },
    {
      "cell_type": "code",
      "source": [
        "model.compile(optimizer='adam',\n",
        "              loss='sparse_categorical_crossentropy',\n",
        "              metrics=['accuracy'])\n",
        "model.fit(x=x_train,y=y_train, epochs=2)"
      ],
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "4RJQb95tYjnM",
        "outputId": "46fbbda1-68d1-4c1d-be84-dd2fdb719705"
      },
      "execution_count": 7,
      "outputs": [
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "Epoch 1/2\n",
            "1875/1875 [==============================] - 32s 17ms/step - loss: 0.2047 - accuracy: 0.9387\n",
            "Epoch 2/2\n",
            "1875/1875 [==============================] - 32s 17ms/step - loss: 0.0857 - accuracy: 0.9735\n"
          ]
        },
        {
          "output_type": "execute_result",
          "data": {
            "text/plain": [
              "<keras.callbacks.History at 0x7f29415af610>"
            ]
          },
          "metadata": {},
          "execution_count": 7
        }
      ]
    },
    {
      "cell_type": "code",
      "source": [
        "model.evaluate(x_test, y_test)"
      ],
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "yD3s4lIO4e1f",
        "outputId": "7cc20563-e67b-4b6a-ac77-9e24d729a65e"
      },
      "execution_count": 8,
      "outputs": [
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "313/313 [==============================] - 2s 7ms/step - loss: 0.0643 - accuracy: 0.9786\n"
          ]
        },
        {
          "output_type": "execute_result",
          "data": {
            "text/plain": [
              "[0.06427575647830963, 0.978600025177002]"
            ]
          },
          "metadata": {},
          "execution_count": 8
        }
      ]
    },
    {
      "cell_type": "code",
      "source": [
        "pred = model.predict(x_test)"
      ],
      "metadata": {
        "id": "w62pUC-m44R6"
      },
      "execution_count": 9,
      "outputs": []
    },
    {
      "cell_type": "code",
      "source": [
        "pred"
      ],
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "xEcvS_7X4-t3",
        "outputId": "74dc5a08-9a77-442e-e470-f81024928103"
      },
      "execution_count": 10,
      "outputs": [
        {
          "output_type": "execute_result",
          "data": {
            "text/plain": [
              "array([[1.4866539e-06, 4.4138989e-07, 2.5047470e-04, ..., 9.9971658e-01,\n",
              "        2.7104772e-07, 4.7934245e-06],\n",
              "       [1.7107850e-04, 2.9236596e-04, 9.9952173e-01, ..., 1.8580701e-10,\n",
              "        3.9757597e-06, 8.3450530e-10],\n",
              "       [4.5024942e-07, 9.9992108e-01, 6.3052771e-06, ..., 3.1571304e-05,\n",
              "        1.0080369e-05, 8.3473327e-07],\n",
              "       ...,\n",
              "       [2.8084533e-11, 6.1918826e-08, 4.4963461e-10, ..., 5.0086351e-07,\n",
              "        1.5927895e-05, 1.2693480e-04],\n",
              "       [3.9020708e-07, 1.7551215e-09, 8.9068600e-11, ..., 1.4160776e-09,\n",
              "        9.2826725e-05, 1.4391756e-10],\n",
              "       [1.8068135e-07, 1.9753529e-10, 1.7003383e-07, ..., 1.2728695e-12,\n",
              "        3.1910363e-07, 5.1220561e-09]], dtype=float32)"
            ]
          },
          "metadata": {},
          "execution_count": 10
        }
      ]
    },
    {
      "cell_type": "code",
      "source": [
        "import numpy as np\n",
        "print(np.argmax(pred[2]))"
      ],
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "_dlLjCtb5B5B",
        "outputId": "da47eea5-5086-4d6c-aae2-3980618504d1"
      },
      "execution_count": 19,
      "outputs": [
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "1\n"
          ]
        }
      ]
    },
    {
      "cell_type": "code",
      "source": [
        "plt.imshow(np.squeeze(x_test[2]))\n",
        "plt.show()"
      ],
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 265
        },
        "id": "cJ4MEq2Z5S_3",
        "outputId": "a543f9d1-c895-4666-b092-5dff28a231cb"
      },
      "execution_count": 20,
      "outputs": [
        {
          "output_type": "display_data",
          "data": {
            "text/plain": [
              "<Figure size 432x288 with 1 Axes>"
            ],
            "image/png": "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\n"
          },
          "metadata": {
            "needs_background": "light"
          }
        }
      ]
    }
  ]
}