TensorFlow_DeepCifar10_NTI.ipynb

{
  "nbformat": 4,
  "nbformat_minor": 0,
  "metadata": {
    "colab": {
      "provenance": []
    },
    "kernelspec": {
      "name": "python3",
      "display_name": "Python 3"
    },
    "language_info": {
      "name": "python"
    },
    "accelerator": "GPU",
    "gpuClass": "standard"
  },
  "cells": [
    {
      "cell_type": "code",
      "execution_count": null,
      "metadata": {
        "id": "-DeP4sHeoeSQ"
      },
      "outputs": [],
      "source": [
        "import numpy as np \n",
        "import matplotlib.pyplot as plt \n",
        "from tensorflow import keras\n",
        "import tensorflow as tf \n",
        "\n"
      ]
    },
    {
      "cell_type": "code",
      "source": [
        "cifar=tf.keras.datasets.cifar10\n",
        "(x_train,y_train),(x_test,y_test)=cifar.load_data()\n",
        "print(x_train.shape,y_train.shape)"
      ],
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "FQmb2UIRqBB4",
        "outputId": "2946a4db-fab0-4a2d-9060-be8555907125"
      },
      "execution_count": null,
      "outputs": [
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "(50000, 32, 32, 3) (50000, 1)\n"
          ]
        }
      ]
    },
    {
      "cell_type": "code",
      "source": [
        "x_train_final=x_train.reshape((-1,32*32*3))\n",
        "x_train_final=x_train_final/255"
      ],
      "metadata": {
        "id": "3dR03BCqrW5w"
      },
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "code",
      "source": [
        "x_test_final=x_test.reshape((-1,32*32*3))\n",
        "x_test_final=x_test_final/255"
      ],
      "metadata": {
        "id": "kuYILyX2rmUB"
      },
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "code",
      "source": [
        "np.unique(y_train)"
      ],
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "1ltOS4M2LZkz",
        "outputId": "6363b305-8532-4d4e-baf3-364fd6df349c"
      },
      "execution_count": null,
      "outputs": [
        {
          "output_type": "execute_result",
          "data": {
            "text/plain": [
              "array([0, 1, 2, 3, 4, 5, 6, 7, 8, 9], dtype=uint8)"
            ]
          },
          "metadata": {},
          "execution_count": 5
        }
      ]
    },
    {
      "cell_type": "code",
      "source": [
        "model=keras.Sequential()\n",
        "model.add(keras.layers.Dense(10000,input_shape=(32*32*3,),activation=\"relu\"))\n",
        "#model.add(keras.layers.Dense(2048,activation=\"relu\"))\n",
        "#model.add(keras.layers.Dense(1024,activation=\"relu\"))\n",
        "#model.add(keras.layers.Dense(512,activation=\"relu\"))\n",
        "#model.add(keras.layers.Dense(256,activation=\"relu\"))\n",
        "#model.add(keras.layers.Dense(128,activation=\"relu\"))\n",
        "#model.add(keras.layers.Dense(64,activation=\"relu\"))\n",
        "#model.add(keras.layers.Dense(32,activation=\"relu\"))\n",
        "model.add(keras.layers.Dense(10,activation=\"softmax\"))\n",
        "model.summary()"
      ],
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "ZUlA9rrAsWVN",
        "outputId": "9b7836ce-8d1e-416b-b13c-95699407ce00"
      },
      "execution_count": null,
      "outputs": [
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "Model: \"sequential\"\n",
            "_________________________________________________________________\n",
            " Layer (type)                Output Shape              Param #   \n",
            "=================================================================\n",
            " dense (Dense)               (None, 10000)             30730000  \n",
            "                                                                 \n",
            " dense_1 (Dense)             (None, 10)                100010    \n",
            "                                                                 \n",
            "=================================================================\n",
            "Total params: 30,830,010\n",
            "Trainable params: 30,830,010\n",
            "Non-trainable params: 0\n",
            "_________________________________________________________________\n"
          ]
        }
      ]
    },
    {
      "cell_type": "code",
      "source": [
        "model.compile(optimizer='sgd',loss='sparse_categorical_crossentropy',metrics=['accuracy'])"
      ],
      "metadata": {
        "id": "QoHHeUNzta01"
      },
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "code",
      "source": [
        "results=model.fit(x=x_train_final,y=y_train,shuffle=True,epochs=50,batch_size=32,validation_data=(x_test_final,y_test),verbose=1)"
      ],
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "MDK6tDQ_tu55",
        "outputId": "50599a28-3b53-4fe9-84d9-a78a1045cded"
      },
      "execution_count": null,
      "outputs": [
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "Epoch 1/50\n",
            "1563/1563 [==============================] - 14s 8ms/step - loss: 1.8202 - accuracy: 0.3569 - val_loss: 1.6942 - val_accuracy: 0.3980\n",
            "Epoch 2/50\n",
            "1563/1563 [==============================] - 8s 5ms/step - loss: 1.6358 - accuracy: 0.4272 - val_loss: 1.6030 - val_accuracy: 0.4376\n",
            "Epoch 3/50\n",
            "1563/1563 [==============================] - 8s 5ms/step - loss: 1.5591 - accuracy: 0.4541 - val_loss: 1.5581 - val_accuracy: 0.4498\n",
            "Epoch 4/50\n",
            "1563/1563 [==============================] - 8s 5ms/step - loss: 1.5034 - accuracy: 0.4750 - val_loss: 1.5194 - val_accuracy: 0.4542\n",
            "Epoch 5/50\n",
            "1563/1563 [==============================] - 9s 6ms/step - loss: 1.4590 - accuracy: 0.4919 - val_loss: 1.5110 - val_accuracy: 0.4631\n",
            "Epoch 6/50\n",
            "1563/1563 [==============================] - 9s 6ms/step - loss: 1.4213 - accuracy: 0.5046 - val_loss: 1.4599 - val_accuracy: 0.4884\n",
            "Epoch 7/50\n",
            "1563/1563 [==============================] - 9s 6ms/step - loss: 1.3876 - accuracy: 0.5175 - val_loss: 1.5035 - val_accuracy: 0.4648\n",
            "Epoch 8/50\n",
            "1563/1563 [==============================] - 8s 5ms/step - loss: 1.3565 - accuracy: 0.5302 - val_loss: 1.4186 - val_accuracy: 0.5051\n",
            "Epoch 9/50\n",
            "1563/1563 [==============================] - 9s 6ms/step - loss: 1.3309 - accuracy: 0.5375 - val_loss: 1.4348 - val_accuracy: 0.4826\n",
            "Epoch 10/50\n",
            "1563/1563 [==============================] - 9s 6ms/step - loss: 1.3053 - accuracy: 0.5470 - val_loss: 1.4240 - val_accuracy: 0.4993\n",
            "Epoch 11/50\n",
            "1563/1563 [==============================] - 9s 6ms/step - loss: 1.2799 - accuracy: 0.5570 - val_loss: 1.3936 - val_accuracy: 0.5008\n",
            "Epoch 12/50\n",
            "1563/1563 [==============================] - 9s 6ms/step - loss: 1.2563 - accuracy: 0.5650 - val_loss: 1.3882 - val_accuracy: 0.5141\n",
            "Epoch 13/50\n",
            "1563/1563 [==============================] - 9s 6ms/step - loss: 1.2363 - accuracy: 0.5721 - val_loss: 1.4152 - val_accuracy: 0.5044\n",
            "Epoch 14/50\n",
            "1563/1563 [==============================] - 8s 5ms/step - loss: 1.2147 - accuracy: 0.5808 - val_loss: 1.4597 - val_accuracy: 0.4841\n",
            "Epoch 15/50\n",
            "1563/1563 [==============================] - 9s 6ms/step - loss: 1.1946 - accuracy: 0.5871 - val_loss: 1.4026 - val_accuracy: 0.5096\n",
            "Epoch 16/50\n",
            "1563/1563 [==============================] - 9s 6ms/step - loss: 1.1743 - accuracy: 0.5938 - val_loss: 1.4249 - val_accuracy: 0.4968\n",
            "Epoch 17/50\n",
            "1563/1563 [==============================] - 8s 5ms/step - loss: 1.1557 - accuracy: 0.6019 - val_loss: 1.4005 - val_accuracy: 0.5080\n",
            "Epoch 18/50\n",
            "1563/1563 [==============================] - 8s 5ms/step - loss: 1.1359 - accuracy: 0.6093 - val_loss: 1.3513 - val_accuracy: 0.5193\n",
            "Epoch 19/50\n",
            "1563/1563 [==============================] - 8s 5ms/step - loss: 1.1207 - accuracy: 0.6159 - val_loss: 1.3317 - val_accuracy: 0.5306\n",
            "Epoch 20/50\n",
            "1563/1563 [==============================] - 9s 6ms/step - loss: 1.1018 - accuracy: 0.6229 - val_loss: 1.4265 - val_accuracy: 0.5023\n",
            "Epoch 21/50\n",
            "1563/1563 [==============================] - 8s 5ms/step - loss: 1.0839 - accuracy: 0.6288 - val_loss: 1.3668 - val_accuracy: 0.5172\n",
            "Epoch 22/50\n",
            "1563/1563 [==============================] - 9s 6ms/step - loss: 1.0647 - accuracy: 0.6321 - val_loss: 1.3804 - val_accuracy: 0.5140\n",
            "Epoch 23/50\n",
            "1563/1563 [==============================] - 9s 6ms/step - loss: 1.0472 - accuracy: 0.6413 - val_loss: 1.6443 - val_accuracy: 0.4400\n",
            "Epoch 24/50\n",
            "1563/1563 [==============================] - 8s 5ms/step - loss: 1.0289 - accuracy: 0.6470 - val_loss: 1.3469 - val_accuracy: 0.5252\n",
            "Epoch 25/50\n",
            "1563/1563 [==============================] - 8s 5ms/step - loss: 1.0140 - accuracy: 0.6523 - val_loss: 1.3474 - val_accuracy: 0.5253\n",
            "Epoch 26/50\n",
            "1563/1563 [==============================] - 9s 5ms/step - loss: 0.9974 - accuracy: 0.6589 - val_loss: 1.3656 - val_accuracy: 0.5212\n",
            "Epoch 27/50\n",
            "1563/1563 [==============================] - 9s 5ms/step - loss: 0.9821 - accuracy: 0.6647 - val_loss: 1.3613 - val_accuracy: 0.5232\n",
            "Epoch 28/50\n",
            "1563/1563 [==============================] - 9s 5ms/step - loss: 0.9623 - accuracy: 0.6719 - val_loss: 1.3741 - val_accuracy: 0.5261\n",
            "Epoch 29/50\n",
            "1563/1563 [==============================] - 9s 6ms/step - loss: 0.9495 - accuracy: 0.6758 - val_loss: 1.3851 - val_accuracy: 0.5186\n",
            "Epoch 30/50\n",
            "1563/1563 [==============================] - 9s 6ms/step - loss: 0.9302 - accuracy: 0.6838 - val_loss: 1.3953 - val_accuracy: 0.5240\n",
            "Epoch 31/50\n",
            "1563/1563 [==============================] - 9s 5ms/step - loss: 0.9132 - accuracy: 0.6886 - val_loss: 1.3345 - val_accuracy: 0.5301\n",
            "Epoch 32/50\n",
            "1563/1563 [==============================] - 8s 5ms/step - loss: 0.9003 - accuracy: 0.6926 - val_loss: 1.3679 - val_accuracy: 0.5248\n",
            "Epoch 33/50\n",
            "1563/1563 [==============================] - 9s 6ms/step - loss: 0.8841 - accuracy: 0.7006 - val_loss: 1.4438 - val_accuracy: 0.5042\n",
            "Epoch 34/50\n",
            "1563/1563 [==============================] - 8s 5ms/step - loss: 0.8694 - accuracy: 0.7061 - val_loss: 1.4343 - val_accuracy: 0.5007\n",
            "Epoch 35/50\n",
            "1563/1563 [==============================] - 8s 5ms/step - loss: 0.8546 - accuracy: 0.7107 - val_loss: 1.3266 - val_accuracy: 0.5449\n",
            "Epoch 36/50\n",
            "1563/1563 [==============================] - 8s 5ms/step - loss: 0.8373 - accuracy: 0.7173 - val_loss: 1.3653 - val_accuracy: 0.5405\n",
            "Epoch 37/50\n",
            "1563/1563 [==============================] - 8s 5ms/step - loss: 0.8211 - accuracy: 0.7242 - val_loss: 1.5287 - val_accuracy: 0.4776\n",
            "Epoch 38/50\n",
            "1563/1563 [==============================] - 9s 6ms/step - loss: 0.8089 - accuracy: 0.7278 - val_loss: 1.3524 - val_accuracy: 0.5399\n",
            "Epoch 39/50\n",
            "1563/1563 [==============================] - 9s 6ms/step - loss: 0.7909 - accuracy: 0.7362 - val_loss: 1.4491 - val_accuracy: 0.5104\n",
            "Epoch 40/50\n",
            "1563/1563 [==============================] - 9s 6ms/step - loss: 0.7761 - accuracy: 0.7415 - val_loss: 1.3902 - val_accuracy: 0.5328\n",
            "Epoch 41/50\n",
            "1563/1563 [==============================] - 8s 5ms/step - loss: 0.7592 - accuracy: 0.7452 - val_loss: 1.5202 - val_accuracy: 0.4974\n",
            "Epoch 42/50\n",
            "1563/1563 [==============================] - 9s 6ms/step - loss: 0.7474 - accuracy: 0.7517 - val_loss: 1.4863 - val_accuracy: 0.5115\n",
            "Epoch 43/50\n",
            "1563/1563 [==============================] - 9s 6ms/step - loss: 0.7356 - accuracy: 0.7568 - val_loss: 1.3284 - val_accuracy: 0.5497\n",
            "Epoch 44/50\n",
            "1563/1563 [==============================] - 9s 6ms/step - loss: 0.7159 - accuracy: 0.7628 - val_loss: 1.3780 - val_accuracy: 0.5367\n",
            "Epoch 45/50\n",
            "1563/1563 [==============================] - 8s 5ms/step - loss: 0.7042 - accuracy: 0.7659 - val_loss: 1.4472 - val_accuracy: 0.5221\n",
            "Epoch 46/50\n",
            "1563/1563 [==============================] - 9s 6ms/step - loss: 0.6907 - accuracy: 0.7722 - val_loss: 1.3283 - val_accuracy: 0.5545\n",
            "Epoch 47/50\n",
            "1563/1563 [==============================] - 9s 6ms/step - loss: 0.6748 - accuracy: 0.7790 - val_loss: 1.4121 - val_accuracy: 0.5354\n",
            "Epoch 48/50\n",
            "1563/1563 [==============================] - 8s 5ms/step - loss: 0.6621 - accuracy: 0.7831 - val_loss: 1.3266 - val_accuracy: 0.5645\n",
            "Epoch 49/50\n",
            "1563/1563 [==============================] - 8s 5ms/step - loss: 0.6466 - accuracy: 0.7894 - val_loss: 1.4880 - val_accuracy: 0.5268\n",
            "Epoch 50/50\n",
            "1563/1563 [==============================] - 8s 5ms/step - loss: 0.6340 - accuracy: 0.7937 - val_loss: 1.3499 - val_accuracy: 0.5575\n"
          ]
        }
      ]
    },
    {
      "cell_type": "code",
      "source": [
        "predict=model.predict(x_test_final)\n",
        "predict"
      ],
      "metadata": {
        "id": "CFy30WfJWZgB",
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "outputId": "39177816-0f3c-45fe-e284-067823bea3a6"
      },
      "execution_count": null,
      "outputs": [
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "313/313 [==============================] - 1s 2ms/step\n"
          ]
        },
        {
          "output_type": "execute_result",
          "data": {
            "text/plain": [
              "array([[3.3224330e-03, 1.9329352e-02, 3.8593464e-02, ..., 1.6646140e-06,\n",
              "        3.2347683e-03, 5.7412189e-04],\n",
              "       [3.3585029e-03, 6.1551388e-02, 1.5280006e-03, ..., 5.8888298e-05,\n",
              "        5.5828285e-01, 3.7484932e-01],\n",
              "       [7.5017947e-01, 3.9295983e-02, 8.2944371e-03, ..., 6.2856367e-03,\n",
              "        1.1238943e-01, 8.1357837e-02],\n",
              "       ...,\n",
              "       [1.0566115e-03, 3.1086963e-04, 7.9004776e-01, ..., 4.3784366e-03,\n",
              "        6.6039478e-03, 1.2723636e-03],\n",
              "       [2.3839254e-02, 1.5184009e-02, 2.1616353e-01, ..., 3.6207479e-03,\n",
              "        2.4312749e-04, 7.9129974e-04],\n",
              "       [2.8184419e-03, 7.0123626e-03, 1.2057094e-01, ..., 3.0232379e-01,\n",
              "        5.2212537e-03, 6.6904617e-03]], dtype=float32)"
            ]
          },
          "metadata": {},
          "execution_count": 9
        }
      ]
    },
    {
      "cell_type": "code",
      "source": [
        "plt.plot(results.history['accuracy'])\n",
        "plt.plot(results.history['val_accuracy'])\n",
        "plt.legend(['Training', 'Validation'])\n",
        "plt.title('Training and Validation losses')\n",
        "plt.xlabel('epoches')\n",
        "plt.ylabel('accuracy')\n"
      ],
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 312
        },
        "id": "lmjrVNQjfLqN",
        "outputId": "eecf6221-5d62-4a4d-ee94-48a2d204f4bf"
      },
      "execution_count": null,
      "outputs": [
        {
          "output_type": "execute_result",
          "data": {
            "text/plain": [
              "Text(0, 0.5, 'accuracy')"
            ]
          },
          "metadata": {},
          "execution_count": 10
        },
        {
          "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(results.history['loss'])\n",
        "plt.plot(results.history['val_loss'])\n",
        "plt.legend(['Training', 'Validation'])\n",
        "plt.title('Training and Validation losses')\n",
        "plt.xlabel('epoches')\n",
        "plt.ylabel('loss')\n"
      ],
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 312
        },
        "outputId": "fbc06f13-b980-47c1-e40c-87425cce3bac",
        "id": "6wk1fqDTys0e"
      },
      "execution_count": null,
      "outputs": [
        {
          "output_type": "execute_result",
          "data": {
            "text/plain": [
              "Text(0, 0.5, 'loss')"
            ]
          },
          "metadata": {},
          "execution_count": 11
        },
        {
          "output_type": "display_data",
          "data": {
            "text/plain": [
              "<Figure size 432x288 with 1 Axes>"
            ],
            "image/png": "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\n"
          },
          "metadata": {
            "needs_background": "light"
          }
        }
      ]
    }
  ]
}