1_CapasDensas.ipynb

{
 "cells": [
  {
   "cell_type": "code",
   "execution_count": 15,
   "metadata": {},
   "outputs": [],
   "source": [
    "import numpy as np\n",
    "import matplotlib.pyplot as plt\n",
    "from keras.datasets import mnist\n",
    "from keras.models import Sequential\n",
    "from keras.layers import Dense\n",
    "from keras.optimizers import SGD\n",
    "from keras.utils import to_categorical\n",
    "%matplotlib inline"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## MNIST dataset"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 5,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "5"
      ]
     },
     "execution_count": 5,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "(X_train, y_train), (X_test, y_test) = mnist.load_data()\n",
    "y_train[0]"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 7,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "image/png": "iVBORw0KGgoAAAANSUhEUgAAAP8AAAEICAYAAACQ6CLfAAAABHNCSVQICAgIfAhkiAAAAAlwSFlzAAALEgAACxIB0t1+/AAAADl0RVh0U29mdHdhcmUAbWF0cGxvdGxpYiB2ZXJzaW9uIDIuMi4yLCBodHRwOi8vbWF0cGxvdGxpYi5vcmcvhp/UCwAAB5hJREFUeJzt3U+I13Uex/HvZ3Njsc0GLwqBiB1GMmIuKYisKyEiKOxUh5XtZHRqoJOXbsHaBmoH1w5zCrqIx9SF8qAWbCBIfy7DurC3Yi5LTaVlovPdS7ft955x/M38xnk9HuDp5dfv9+CTr/qZ39j6vu+APL8Z9QMAoyF+CCV+CCV+CCV+CCV+CCV+CCV+flVr7Wpr7XZr7eYvP26M+pkYLvFTmer7/ve//Bgf9cMwXOKHUOKn8rfW2n9ba/9srf1x1A/DcDVf28+vaa3t6rpupuu6O13X/bnrujNd1030ff+fkT4YQyN+FqW19mHXdf/o+/7vo34WhsMf+1msvuu6NuqHYHjEz/9prY211g601n7XWlvXWvtL13V/6Lruw1E/G8OzbtQPwKr0267r/tp13fau6+51Xfevruv+1Pf9v0f6VAyVv/NDKH/sh1Dih1Dih1Dih1Ar+q/9rTX/ugjLrO/7RX09hjc/hBI/hBI/hBI/hBI/hBI/hBI/hBI/hBI/hBI/hBI/hBI/hBI/hBI/hBI/hBI/hBI/hBI/hBI/hBI/hBI/hBI/hBI/hBI/hBI/hBI/hBI/hBI/hBI/hBI/hBI/hBI/hBI/hBI/hBI/hBI/hBI/hBI/hFo36gdgeT3yyCPl/sQTTyzr/aempgZu69evL68dHx8v99dee63cT548OXA7cuRIee3t27fL/e233y73N998s9xXA29+CCV+CCV+CCV+CCV+CCV+CCV+COWcfwVs2bKl3B999NFy3717d7nv2bNn4DY2NlZe++KLL5b7KH311Vflfvr06XKfnJwcuP3www/ltV9++WW5f/zxx+X+MPDmh1Dih1Dih1Dih1Dih1Dih1Ct7/uVu1lrK3ezFTQxMVHuly9fLvfl/ljtajU/P1/uR48eLfebN28u+d6zs7Pl/u2335b7jRs3lnzv5db3fVvMz/Pmh1Dih1Dih1Dih1Dih1Dih1Dih1DO+Ydg48aN5X7t2rVy37Zt2zAfZ6gWeva5ubly37dv38Dtzp075bWpX//woJzzAyXxQyjxQyjxQyjxQyjxQyjxQyjfunsIvvnmm3I/duxYuR86dKjcP//883Jf6FtYV7744oty379/f7nfunWr3Hfs2DFwe/3118trWV7e/BBK/BBK/BBK/BBK/BBK/BBK/BDK5/lXgQ0bNpT7Qv+d9PT09MDtlVdeKa99+eWXy/3s2bPlzurj8/xASfwQSvwQSvwQSvwQSvwQSvwQyuf5V4Hvv//+ga7/7rvvlnztq6++Wu7nzp0r9/n5+SXfm9Hy5odQ4odQ4odQ4odQ4odQ4odQPtK7Bjz22GMDtwsXLpTX7t27t9wPHjxY7pcuXSp3Vp6P9AIl8UMo8UMo8UMo8UMo8UMo8UMo5/xr3FNPPVXun332WbnPzc2V+5UrV8r9+vXrA7d33323vHYlf2+uJc75gZL4IZT4IZT4IZT4IZT4IZT4IZRz/nCTk5Pl/t5775X7448/vuR7v/HGG+X+/vvvl/vs7OyS772WOecHSuKHUOKHUOKHUOKHUOKHUOKHUM75KT3zzDPl/s4775T7888/v+R7T09Pl/vx48fL/euvv17yvR9mzvmBkvghlPghlPghlPghlPghlPghlHN+HsjY2Fi5Hz58eOC20PcKaK0+rr58+XK579+/v9zXKuf8QEn8EEr8EEr8EEr8EEr8EMpRHyPz888/l/u6devK/e7du+V+4MCBgdvVq1fLax9mjvqAkvghlPghlPghlPghlPghlPghVH2QSrxnn3223F966aVyf+655wZuC53jL2RmZqbcP/nkkwf69dc6b34IJX4IJX4IJX4IJX4IJX4IJX4I5Zx/jRsfHy/3qampcn/hhRfKffPmzff9TIt17969cp+dnS33+fn5YT7OmuPND6HED6HED6HED6HED6HED6HED6Gc8z8EFjpLP3LkyMBtoXP8rVu3LuWRhuL69evlfvz48XI/f/78MB8njjc/hBI/hBI/hBI/hBI/hBI/hHLUtwI2bdpU7k8//XS5nzlzpty3b99+3880LNeuXSv3EydODNw++OCD8lofyV1e3vwQSvwQSvwQSvwQSvwQSvwQSvwQyjn/Im3cuHHgNj09XV47MTFR7tu2bVvSMw3Dp59+Wu6nTp0q948++qjcf/rpp/t+JlaGNz+EEj+EEj+EEj+EEj+EEj+EEj+Eijnn37VrV7kfO3as3Hfu3Dlwe/LJJ5f0TMPy448/DtxOnz5dXvvWW2+V+61bt5b0TKx+3vwQSvwQSvwQSvwQSvwQSvwQSvwQKuacf3Jy8oH2BzEzM1PuFy9eLPe7d++We/WZ+7m5ufJacnnzQyjxQyjxQyjxQyjxQyjxQyjxQ6jW9/3K3ay1lbsZhOr7vi3m53nzQyjxQyjxQyjxQyjxQyjxQyjxQyjxQyjxQyjxQyjxQyjxQyjxQyjxQyjxQyjxQyjxQyjxQyjxQyjxQyjxQyjxQ6gV/dbdwOrhzQ+hxA+hxA+hxA+hxA+hxA+hxA+hxA+hxA+hxA+hxA+hxA+hxA+hxA+hxA+hxA+hxA+hxA+hxA+hxA+hxA+hxA+hxA+h/gdqY2RigYrulQAAAABJRU5ErkJggg==\n",
      "text/plain": [
       "<Figure size 432x288 with 1 Axes>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "plt.imshow(X_train[0], 'gray')\n",
    "plt.title(y_train[0])\n",
    "plt.axis('off');"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 9,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "(-1.0, 1.0)"
      ]
     },
     "execution_count": 9,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "## Preprocesamiento\n",
    "print(X_train.min(), X_train.max())\n",
    "\n",
    "# Normalizamos\n",
    "X_train = (X_train - 127.5) / 127.5\n",
    "X_test = (X_test - 127.5) / 127.5\n",
    "print(X_train.min(), X_train.max())"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 10,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "(60000, 28, 28)"
      ]
     },
     "execution_count": 10,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "X_train.shape"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 12,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "((60000, 784), (10000, 784))"
      ]
     },
     "execution_count": 12,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "# Las imagenes por defecto son de 28x28\n",
    "# Para Fully-Connected Networks usaremos una vector de 784 (28x28)\n",
    "X_train = X_train.reshape((-1, 784))\n",
    "X_test = X_test.reshape((-1, 784))\n",
    "\n",
    "X_train.shape, X_test.shape"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 13,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "5\n",
      "[0. 0. 0. 0. 0. 1. 0. 0. 0. 0.]\n"
     ]
    }
   ],
   "source": [
    "## One Hot encoder\n",
    "\n",
    "print(y_train[0])\n",
    "\n",
    "# Aplicamos la funcion \"to_categorical\" para llevar a un vector de dimension 10\n",
    "y_train = to_categorical(y_train, num_classes=10)\n",
    "y_test = to_categorical(y_test, num_classes=10)\n",
    "\n",
    "print(y_train[0])"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## Construyendo el modelo en Keras"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 16,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "_________________________________________________________________\n",
      "Layer (type)                 Output Shape              Param #   \n",
      "=================================================================\n",
      "dense_1 (Dense)              (None, 10)                7850      \n",
      "=================================================================\n",
      "Total params: 7,850\n",
      "Trainable params: 7,850\n",
      "Non-trainable params: 0\n",
      "_________________________________________________________________\n"
     ]
    }
   ],
   "source": [
    "model = Sequential()\n",
    "\n",
    "model.add(Dense(10, activation='softmax', input_shape=(784,)))\n",
    "\n",
    "model.compile(loss='categorical_crossentropy', optimizer=SGD(1e-2), metrics=['accuracy'])\n",
    "\n",
    "model.summary()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 18,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Train on 60000 samples, validate on 10000 samples\n",
      "Epoch 1/5\n",
      "60000/60000 [==============================] - 2s 40us/step - loss: 1.0251 - acc: 0.7161 - val_loss: 0.6108 - val_acc: 0.8455\n",
      "Epoch 2/5\n",
      "60000/60000 [==============================] - 1s 15us/step - loss: 0.5549 - acc: 0.8527 - val_loss: 0.4839 - val_acc: 0.8702\n",
      "Epoch 3/5\n",
      "60000/60000 [==============================] - 1s 14us/step - loss: 0.4728 - acc: 0.8716 - val_loss: 0.4318 - val_acc: 0.8846\n",
      "Epoch 4/5\n",
      "60000/60000 [==============================] - 1s 15us/step - loss: 0.4333 - acc: 0.8805 - val_loss: 0.4002 - val_acc: 0.8911\n",
      "Epoch 5/5\n",
      "60000/60000 [==============================] - 1s 16us/step - loss: 0.4085 - acc: 0.8864 - val_loss: 0.3820 - val_acc: 0.8944\n"
     ]
    }
   ],
   "source": [
    "log = model.fit(X_train, y_train, validation_data=(X_test, y_test), epochs=5, batch_size=256)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 19,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 432x288 with 1 Axes>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "plt.plot(log.history['loss'], label='train_set')\n",
    "plt.plot(log.history['val_loss'], label='test_set')\n",
    "plt.legend();"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "- Experimentar cambiando el learning rate\n",
    "- Aumentando las capas de la red"
   ]
  }
 ],
 "metadata": {
  "kernelspec": {
   "display_name": "Python 3",
   "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": "3.5.5"
  }
 },
 "nbformat": 4,
 "nbformat_minor": 2
}