Learning/13_CNN.ipynb

{
 "cells": [
  {
   "cell_type": "code",
   "execution_count": 55,
   "id": "d6e84c33",
   "metadata": {},
   "outputs": [],
   "source": [
    "import numpy as np\n",
    "import pandas as pd\n",
    "import tensorflow as tf\n",
    "from tensorflow.keras import datasets,layers,models\n",
    "import matplotlib.pyplot as plt\n"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "3615a839",
   "metadata": {},
   "source": [
    "# Loading the data"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 56,
   "id": "2fa02c40",
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "((50000, 32, 32, 3), (50000, 1))"
      ]
     },
     "execution_count": 56,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "(X_train,y_train),(X_test,y_test) = datasets.cifar10.load_data()\n",
    "X_train.shape , y_train.shape"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "b4adecf0",
   "metadata": {},
   "source": [
    "# Exploring about the data"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 57,
   "id": "db96eaec",
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "array([[ 59,  62,  63],\n",
       "       [ 43,  46,  45],\n",
       "       [ 50,  48,  43],\n",
       "       [ 68,  54,  42],\n",
       "       [ 98,  73,  52],\n",
       "       [119,  91,  63],\n",
       "       [139, 107,  75],\n",
       "       [145, 110,  80],\n",
       "       [149, 117,  89],\n",
       "       [149, 120,  93],\n",
       "       [131, 103,  77],\n",
       "       [125,  99,  76],\n",
       "       [142, 115,  91],\n",
       "       [144, 112,  86],\n",
       "       [137, 105,  79],\n",
       "       [129,  97,  71],\n",
       "       [137, 106,  79],\n",
       "       [134, 106,  76],\n",
       "       [124,  97,  64],\n",
       "       [139, 113,  78],\n",
       "       [139, 112,  75],\n",
       "       [133, 105,  69],\n",
       "       [136, 105,  74],\n",
       "       [139, 108,  77],\n",
       "       [152, 120,  89],\n",
       "       [163, 131, 100],\n",
       "       [168, 136, 108],\n",
       "       [159, 129, 102],\n",
       "       [158, 130, 104],\n",
       "       [158, 132, 108],\n",
       "       [152, 125, 102],\n",
       "       [148, 124, 103]], dtype=uint8)"
      ]
     },
     "execution_count": 57,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "X_train[0][0]"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 58,
   "id": "424c65a9",
   "metadata": {},
   "outputs": [],
   "source": [
    "classes = [\"airplane\",\"automobile\",\"bird\",\"cat\",\"deer\",\"dog\",\"frog\",\"horse\",\"ship\",\"truck\"]"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 59,
   "id": "b6ebdaf3",
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "(50000,)"
      ]
     },
     "execution_count": 59,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "y_train = y_train.reshape(y_train.shape[0])\n",
    "y_train.shape"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 60,
   "id": "12a86dc1",
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "(10000,)"
      ]
     },
     "execution_count": 60,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "y_test = y_test.reshape(y_test.shape[0])\n",
    "y_test.shape"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 61,
   "id": "251feb75",
   "metadata": {},
   "outputs": [],
   "source": [
    "def figure_plot(index):\n",
    "    plt.figure(figsize = (5,2))\n",
    "    plt.imshow(X_train[index])\n",
    "    plt.xlabel(classes[y_train[index]])"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 62,
   "id": "ebd40e71",
   "metadata": {},
   "outputs": [
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 500x200 with 1 Axes>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 500x200 with 1 Axes>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 500x200 with 1 Axes>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 500x200 with 1 Axes>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 500x200 with 1 Axes>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "import random\n",
    "for i in range(5):\n",
    "    j = random.randint(0, X_train.shape[0])\n",
    "    figure_plot(j)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 63,
   "id": "048884ff",
   "metadata": {},
   "outputs": [],
   "source": [
    "#Normalizing the data\n",
    "X_train = X_train/255\n",
    "X_test = X_test / 255"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 67,
   "id": "4bcac59f",
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "0.0"
      ]
     },
     "execution_count": 67,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "X_train.min()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 69,
   "id": "f3b6fb8e",
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "(50000, 32, 32, 3)"
      ]
     },
     "execution_count": 69,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "X_train.shape"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 74,
   "id": "43f4a3a0",
   "metadata": {},
   "outputs": [],
   "source": [
    "model = models.Sequential([\n",
    "    layers.Flatten(input_shape = (32,32,3)),\n",
    "    layers.Dense(2000,activation = 'relu'),\n",
    "    layers.Dense(1000,activation='relu'),\n",
    "    layers.Dense(10,activation='tanh')\n",
    "])"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 82,
   "id": "fb11bc43",
   "metadata": {},
   "outputs": [],
   "source": [
    "model.compile(loss = 'sparse_categorical_crossentropy',metrics = ['accuracy'])"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 83,
   "id": "a5439621",
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Epoch 1/5\n",
      "1563/1563 [==============================] - 103s 66ms/step - loss: 5.0449 - accuracy: 0.0998 - val_loss: 2.3026 - val_accuracy: 0.1001\n",
      "Epoch 2/5\n",
      "1563/1563 [==============================] - 100s 64ms/step - loss: 2.3026 - accuracy: 0.0998 - val_loss: 2.3026 - val_accuracy: 0.1001\n",
      "Epoch 3/5\n",
      "1563/1563 [==============================] - 101s 65ms/step - loss: 2.3026 - accuracy: 0.0998 - val_loss: 2.3026 - val_accuracy: 0.1001\n",
      "Epoch 4/5\n",
      "1563/1563 [==============================] - 100s 64ms/step - loss: 2.3026 - accuracy: 0.0998 - val_loss: 2.3026 - val_accuracy: 0.1001\n",
      "Epoch 5/5\n",
      "1563/1563 [==============================] - 103s 66ms/step - loss: 2.3026 - accuracy: 0.0998 - val_loss: 2.3026 - val_accuracy: 0.1001\n"
     ]
    },
    {
     "data": {
      "text/plain": [
       "<keras.callbacks.History at 0x17f59b50310>"
      ]
     },
     "execution_count": 83,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "model.fit(X_train,y_train,validation_data = (X_test,y_test),epochs = 5)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 86,
   "id": "4379c434",
   "metadata": {},
   "outputs": [],
   "source": [
    "#CNN\n",
    "cnn = models.Sequential([\n",
    "    layers.Conv2D(filters=30,kernel_size = (2,2),activation='relu',input_shape = (32,32,3)),\n",
    "    layers.MaxPooling2D((2,2)),\n",
    "    \n",
    "    layers.Conv2D(filters=60,kernel_size = (2,2),activation='relu'),\n",
    "    layers.MaxPooling2D((2,2)),\n",
    "    \n",
    "    layers.Flatten(),\n",
    "    layers.Dense(64,activation='relu'),\n",
    "    layers.Dense(10,activation = 'softmax')\n",
    "])"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 90,
   "id": "e5f649b2",
   "metadata": {},
   "outputs": [],
   "source": [
    "cnn.compile(optimizer='adam',loss='sparse_categorical_crossentropy', metrics=['accuracy'])"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 94,
   "id": "71b9a5da",
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Epoch 1/5\n",
      "1563/1563 [==============================] - 20s 13ms/step - loss: 0.7295 - accuracy: 0.7451 - val_loss: 0.8992 - val_accuracy: 0.6909\n",
      "Epoch 2/5\n",
      "1563/1563 [==============================] - 18s 11ms/step - loss: 0.6948 - accuracy: 0.7576 - val_loss: 0.9018 - val_accuracy: 0.6847\n",
      "Epoch 3/5\n",
      "1563/1563 [==============================] - 18s 11ms/step - loss: 0.6586 - accuracy: 0.7707 - val_loss: 0.9185 - val_accuracy: 0.6861\n",
      "Epoch 4/5\n",
      "1563/1563 [==============================] - 18s 12ms/step - loss: 0.6267 - accuracy: 0.7820 - val_loss: 0.9067 - val_accuracy: 0.6947\n",
      "Epoch 5/5\n",
      "1563/1563 [==============================] - 18s 12ms/step - loss: 0.5953 - accuracy: 0.7928 - val_loss: 0.9408 - val_accuracy: 0.6950\n"
     ]
    },
    {
     "data": {
      "text/plain": [
       "<keras.callbacks.History at 0x17f74505c90>"
      ]
     },
     "execution_count": 94,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "cnn.fit(X_train,y_train,epochs=5,validation_data = (X_test, y_test) )"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 95,
   "id": "1abf6969",
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "313/313 [==============================] - 1s 5ms/step - loss: 0.9408 - accuracy: 0.6950\n"
     ]
    },
    {
     "data": {
      "text/plain": [
       "[0.9408475756645203, 0.6949999928474426]"
      ]
     },
     "execution_count": 95,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "cnn.evaluate(X_test,y_test)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 98,
   "id": "3acc2132",
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "313/313 [==============================] - 2s 8ms/step\n"
     ]
    }
   ],
   "source": [
    "y_pred = cnn.predict(X_test)\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 110,
   "id": "78964e1b",
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "8"
      ]
     },
     "execution_count": 110,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "np.argmax(y_pred[0])"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 114,
   "id": "9b04330e",
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "8\n",
      "8\n",
      "0\n",
      "0\n",
      "6\n"
     ]
    }
   ],
   "source": [
    "for element in range(5):\n",
    "    print(np.argmax(y_pred[element]))"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 107,
   "id": "d5854abd",
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "[0]"
      ]
     },
     "execution_count": 107,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "\n",
    "y_pred_class = [np.argmax(y_pred[element] for element in range(5))]\n",
    "y_pred_class"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 116,
   "id": "16de0455",
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "array([3, 8, 8, 0, 6], dtype=uint8)"
      ]
     },
     "execution_count": 116,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "y_test[:5]"
   ]
  }
 ],
 "metadata": {
  "kernelspec": {
   "display_name": "Python 3 (ipykernel)",
   "language": "python",
   "name": "python3"
  },
  "language_info": {
   "codemirror_mode": {
    "name": "ipython",
    "version": 3
   },
   "file_extension": ".py",
   "mimetype": "text/x-python",
   "name": "python",
   "nbconvert_exporter": "python",
   "pygments_lexer": "ipython3",
   "version": "3.10.6"
  }
 },
 "nbformat": 4,
 "nbformat_minor": 5
}