Keras.ipynb

{
 "cells": [
  {
   "cell_type": "code",
   "execution_count": 1,
   "metadata": {},
   "outputs": [
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "Using TensorFlow backend.\n"
     ]
    }
   ],
   "source": [
    "from keras import models\n",
    "from keras import layers\n",
    "from keras.datasets import mnist"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## Exemplo MNIST"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 2,
   "metadata": {},
   "outputs": [],
   "source": [
    "(train_images, train_labels) , (test_images, test_labels) = mnist.load_data()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 3,
   "metadata": {},
   "outputs": [],
   "source": [
    "network = models.Sequential()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 4,
   "metadata": {},
   "outputs": [],
   "source": [
    "network.add(layers.Dense(512, activation='relu', input_shape=(28 * 28,)))\n",
    "network.add(layers.Dense(10, activation='softmax'))"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 5,
   "metadata": {},
   "outputs": [],
   "source": [
    "network.compile(optimizer=\"rmsprop\", loss=\"categorical_crossentropy\", metrics=[\"accuracy\"])"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 6,
   "metadata": {},
   "outputs": [],
   "source": [
    "digit = test_images[90]\n",
    "train_images = train_images.reshape((60000, 28 * 28))\n",
    "train_images = train_images.astype('float32') / 255\n",
    "\n",
    "test_images = test_images.reshape((10000, 28 * 28))\n",
    "test_images = test_images.astype('float32') / 255"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 7,
   "metadata": {},
   "outputs": [],
   "source": [
    "from keras.utils import to_categorical"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 8,
   "metadata": {},
   "outputs": [],
   "source": [
    "train_labels = to_categorical(train_labels)\n",
    "test_labels = to_categorical(test_labels)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 9,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Epoch 1/3\n",
      "60000/60000 [==============================] - 9s 156us/step - loss: 0.2546 - accuracy: 0.9262\n",
      "Epoch 2/3\n",
      "60000/60000 [==============================] - 1s 18us/step - loss: 0.1044 - accuracy: 0.9695\n",
      "Epoch 3/3\n",
      "60000/60000 [==============================] - 1s 18us/step - loss: 0.0683 - accuracy: 0.9795\n"
     ]
    },
    {
     "data": {
      "text/plain": [
       "<keras.callbacks.callbacks.History at 0x7efb76fabcc0>"
      ]
     },
     "execution_count": 9,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "network.fit(train_images, train_labels, epochs=3, batch_size=128)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 10,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "10000/10000 [==============================] - 0s 43us/step\n",
      "0.08201343419300393 0.9729999899864197\n"
     ]
    }
   ],
   "source": [
    "test_loss, test_acc = network.evaluate(test_images, test_labels)\n",
    "print(test_loss, test_acc)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 11,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 432x288 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "import matplotlib.pyplot as plt\n",
    "\n",
    "plt.imshow(digit, cmap=plt.cm.binary)\n",
    "plt.show()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 12,
   "metadata": {},
   "outputs": [],
   "source": [
    "resultados = network.predict(test_images)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 13,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "A probabilidade de ser 3 é 0.9924829602241516\n"
     ]
    }
   ],
   "source": [
    "for i, probabilidade in enumerate(resultados[90]):\n",
    "    if(probabilidade > .1):\n",
    "        print(\"A probabilidade de ser {0} é {1}\".format(i, probabilidade))"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### IMDB movies example"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 12,
   "metadata": {},
   "outputs": [],
   "source": [
    "from keras.datasets import imdb"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 13,
   "metadata": {},
   "outputs": [],
   "source": [
    "(train_data, train_labels) , (test_data, test_labels) = imdb.load_data(num_words=10000)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 97,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "8982"
      ]
     },
     "execution_count": 97,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "len(train_data)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 14,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "9999"
      ]
     },
     "execution_count": 14,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "max(max(seq) for seq in train_data)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 15,
   "metadata": {},
   "outputs": [],
   "source": [
    "word_index = imdb.get_word_index()\n",
    "reverse_word_index = dict([(value,key) for (key, value) in word_index.items()])\n",
    "\n",
    "decoded_review = ' '.join([\n",
    "    reverse_word_index.get(i -3, '?') for i in train_data[0]\n",
    "])"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 16,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "\"? this film was just brilliant casting location scenery story direction everyone's really suited the part they played and you could just imagine being there robert ? is an amazing actor and now the same being director ? father came from the same scottish island as myself so i loved the fact there was a real connection with this film the witty remarks throughout the film were great it was just brilliant so much that i bought the film as soon as it was released for ? and would recommend it to everyone to watch and the fly fishing was amazing really cried at the end it was so sad and you know what they say if you cry at a film it must have been good and this definitely was also ? to the two little boy's that played the ? of norman and paul they were just brilliant children are often left out of the ? list i think because the stars that play them all grown up are such a big profile for the whole film but these children are amazing and should be praised for what they have done don't you think the whole story was so lovely because it was true and was someone's life after all that was shared with us all\""
      ]
     },
     "execution_count": 16,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "decoded_review"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 17,
   "metadata": {},
   "outputs": [],
   "source": [
    "import numpy as np\n",
    "\n",
    "def vectorizes_sequences(sequences, dimension=10000):\n",
    "    results = np.zeros((len(sequences), dimension))\n",
    "    for i, sequence in enumerate(sequences):\n",
    "        results[i, sequence] = 1.\n",
    "    return results\n",
    "\n",
    "x_train = vectorizes_sequences(train_data)\n",
    "x_test = vectorizes_sequences(train_data)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 18,
   "metadata": {},
   "outputs": [],
   "source": [
    "y_train = np.asarray(train_labels).astype('float32')\n",
    "y_test = np.asarray(test_labels).astype('float32')"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### Creating the model "
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 20,
   "metadata": {},
   "outputs": [],
   "source": [
    "from keras import models\n",
    "from keras import layers"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 98,
   "metadata": {},
   "outputs": [],
   "source": [
    "model = models.Sequential()\n",
    "model.add(layers.Dense(16, activation ='relu', input_shape=(10000,)))\n",
    "model.add(layers.Dense(16, activation='relu'))\n",
    "model.add(layers.Dense(1, activation='sigmoid'))"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 84,
   "metadata": {},
   "outputs": [],
   "source": [
    "model.compile(optimizer='rmsprop',\n",
    "             loss='binary_crossentropy',\n",
    "             metrics=['accuracy'])"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 85,
   "metadata": {},
   "outputs": [],
   "source": [
    "from keras import optimizers\n",
    "\n",
    "model.compile(optimizer=optimizers.RMSprop(lr=0.001),\n",
    "             loss='binary_crossentropy',\n",
    "             metrics=['accuracy'])"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 86,
   "metadata": {},
   "outputs": [],
   "source": [
    "from keras import losses\n",
    "from keras import metrics\n",
    "\n",
    "model.compile(optimizer=optimizers.RMSprop(lr=0.001),\n",
    "             loss=losses.binary_crossentropy,\n",
    "             metrics=[metrics.binary_accuracy])"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 87,
   "metadata": {},
   "outputs": [],
   "source": [
    "x_val = x_train[:10000]\n",
    "partial_x_train = x_train[10000:]\n",
    "\n",
    "y_val = y_train[:10000]\n",
    "partial_y_train = y_train[10000:]"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 99,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Train on 15000 samples, validate on 10000 samples\n",
      "Epoch 1/20\n",
      "15000/15000 [==============================] - 1s 68us/step - loss: 0.5082 - acc: 0.7853 - val_loss: 0.3834 - val_acc: 0.8611\n",
      "Epoch 2/20\n",
      "15000/15000 [==============================] - 1s 62us/step - loss: 0.2979 - acc: 0.9064 - val_loss: 0.3017 - val_acc: 0.8903\n",
      "Epoch 3/20\n",
      "15000/15000 [==============================] - 1s 62us/step - loss: 0.2173 - acc: 0.9321 - val_loss: 0.2953 - val_acc: 0.8801\n",
      "Epoch 4/20\n",
      "15000/15000 [==============================] - 1s 62us/step - loss: 0.1717 - acc: 0.9444 - val_loss: 0.2824 - val_acc: 0.8867\n",
      "Epoch 5/20\n",
      "15000/15000 [==============================] - 1s 62us/step - loss: 0.1418 - acc: 0.9541 - val_loss: 0.2931 - val_acc: 0.8833\n",
      "Epoch 6/20\n",
      "15000/15000 [==============================] - 1s 62us/step - loss: 0.1185 - acc: 0.9634 - val_loss: 0.2914 - val_acc: 0.8864\n",
      "Epoch 7/20\n",
      "15000/15000 [==============================] - 1s 62us/step - loss: 0.0982 - acc: 0.9715 - val_loss: 0.3098 - val_acc: 0.8811\n",
      "Epoch 8/20\n",
      "15000/15000 [==============================] - 1s 62us/step - loss: 0.0821 - acc: 0.9763 - val_loss: 0.3263 - val_acc: 0.8820\n",
      "Epoch 9/20\n",
      "15000/15000 [==============================] - 1s 61us/step - loss: 0.0688 - acc: 0.9819 - val_loss: 0.3502 - val_acc: 0.8811\n",
      "Epoch 10/20\n",
      "15000/15000 [==============================] - 1s 62us/step - loss: 0.0564 - acc: 0.9854 - val_loss: 0.3732 - val_acc: 0.8768\n",
      "Epoch 11/20\n",
      "15000/15000 [==============================] - 1s 62us/step - loss: 0.0467 - acc: 0.9883 - val_loss: 0.4007 - val_acc: 0.8752\n",
      "Epoch 12/20\n",
      "15000/15000 [==============================] - 1s 64us/step - loss: 0.0382 - acc: 0.9919 - val_loss: 0.4315 - val_acc: 0.8737\n",
      "Epoch 13/20\n",
      "15000/15000 [==============================] - 1s 63us/step - loss: 0.0291 - acc: 0.9941 - val_loss: 0.4633 - val_acc: 0.8715\n",
      "Epoch 14/20\n",
      "15000/15000 [==============================] - 1s 62us/step - loss: 0.0272 - acc: 0.9940 - val_loss: 0.4930 - val_acc: 0.8705\n",
      "Epoch 15/20\n",
      "15000/15000 [==============================] - 1s 63us/step - loss: 0.0155 - acc: 0.9983 - val_loss: 0.5327 - val_acc: 0.8688\n",
      "Epoch 16/20\n",
      "15000/15000 [==============================] - 1s 64us/step - loss: 0.0174 - acc: 0.9970 - val_loss: 0.5579 - val_acc: 0.8678\n",
      "Epoch 17/20\n",
      "15000/15000 [==============================] - 1s 62us/step - loss: 0.0099 - acc: 0.9995 - val_loss: 0.6130 - val_acc: 0.8658\n",
      "Epoch 18/20\n",
      "15000/15000 [==============================] - 1s 61us/step - loss: 0.0118 - acc: 0.9977 - val_loss: 0.6306 - val_acc: 0.8670\n",
      "Epoch 19/20\n",
      "15000/15000 [==============================] - 1s 63us/step - loss: 0.0057 - acc: 0.9998 - val_loss: 0.7105 - val_acc: 0.8586\n",
      "Epoch 20/20\n",
      "15000/15000 [==============================] - 1s 62us/step - loss: 0.0059 - acc: 0.9993 - val_loss: 0.6975 - val_acc: 0.8631\n"
     ]
    }
   ],
   "source": [
    "model.compile(optimizer='rmsprop',\n",
    "             loss='binary_crossentropy',\n",
    "             metrics=['acc'])\n",
    "\n",
    "history = model.fit(partial_x_train,\n",
    "                   partial_y_train,\n",
    "                   epochs=20,\n",
    "                   batch_size=512,\n",
    "                   validation_data=(x_val,y_val))"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 81,
   "metadata": {},
   "outputs": [],
   "source": [
    "history_dict = history.history"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 28,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "dict_keys(['val_loss', 'val_acc', 'loss', 'acc'])"
      ]
     },
     "execution_count": 28,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "history_dict.keys()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 36,
   "metadata": {},
   "outputs": [],
   "source": [
    "import matplotlib.pyplot as plt"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 37,
   "metadata": {},
   "outputs": [],
   "source": [
    "loss_values = history_dict['loss']\n",
    "val_loss_values = history_dict['val_loss']\n",
    "acc = history_dict['acc']"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 38,
   "metadata": {},
   "outputs": [],
   "source": [
    "epochs = range(1, len(acc) + 1)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 42,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 432x288 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "plt.plot(epochs, loss_values, 'bo', label='Training Loss')\n",
    "plt.plot(epochs, val_loss_values, 'b', label='Validation loss')\n",
    "plt.title('Training and validation loss')\n",
    "plt.xlabel('Epochs')\n",
    "plt.ylabel('Loss')\n",
    "plt.legend()\n",
    "\n",
    "plt.show()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 44,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "<Figure size 432x288 with 0 Axes>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "plt.clf()\n",
    "\n",
    "acc_values = history_dict['acc']\n",
    "val_acc_values = history_dict['val_acc']"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 45,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 432x288 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "plt.plot(epochs, acc_values, 'bo', label='Training Acc')\n",
    "plt.plot(epochs, val_acc_values, 'b', label='Validation Acc')\n",
    "plt.title('Training and validation Accuracy')\n",
    "plt.xlabel('Epochs')\n",
    "plt.ylabel('Accuracy')\n",
    "plt.legend()\n",
    "\n",
    "plt.show()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 71,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "array([[9.9999559e-01],\n",
       "       [1.3602984e-05],\n",
       "       [1.8877820e-06],\n",
       "       ...,\n",
       "       [9.3752506e-06],\n",
       "       [9.9999249e-01],\n",
       "       [1.7884157e-03]], dtype=float32)"
      ]
     },
     "execution_count": 71,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "model.predict(x_test)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## Reuters multiclass news classification "
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 89,
   "metadata": {},
   "outputs": [],
   "source": [
    "from keras.datasets import reuters"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 90,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Downloading data from https://s3.amazonaws.com/text-datasets/reuters.npz\n",
      "2113536/2110848 [==============================] - 5s 2us/step\n"
     ]
    }
   ],
   "source": [
    "(train_data, train_labels) , (test_data, test_labels) = reuters.load_data(num_words=10000)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 93,
   "metadata": {},
   "outputs": [],
   "source": [
    "import numpy as np\n",
    "\n",
    "def vectorizes_sequences(sequences, dimension=10000):\n",
    "    results = np.zeros((len(sequences), dimension))\n",
    "    for i, sequence in enumerate(sequences):\n",
    "        results[i, sequence] = 1.\n",
    "    return results"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 94,
   "metadata": {},
   "outputs": [],
   "source": [
    "x_train = vectorizes_sequences(train_data)\n",
    "x_test = vectorizes_sequences(test_data)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 96,
   "metadata": {},
   "outputs": [],
   "source": [
    "def to_one_hot(labels, dimension=46):\n",
    "    results = np.zeros((len(labels), dimension))\n",
    "    for i, label in enumerate(labels):\n",
    "        results[i, label] = 1\n",
    "    return results\n",
    "\n",
    "one_hot_train_labels = to_one_hot(train_labels)\n",
    "one_hot_train_labels = to_one_hot(test_labels)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": []
  }
 ],
 "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": "3.6.9"
  }
 },
 "nbformat": 4,
 "nbformat_minor": 4
}