pmarcelino

# Extract features
import os, shutil
from keras.preprocessing.image import ImageDataGenerator

datagen = ImageDataGenerator(rescale=1./255)
batch_size = 32

def extract_features(directory, sample_count):
    features = np.zeros(shape=(sample_count, 7, 7, 512))  # Must be equal to the output of the convolutional base
    labels = np.zeros(shape=(sample_count))

pmarcelino

pmarcelino

# Build model
import sklearn
from sklearn.cross_validation import train_test_split
from sklearn.grid_search import GridSearchCV
from sklearn.svm import LinearSVC

X_train, y_train = svm_features.reshape(300,7*7*512), svm_labels

param = [{
          "C": [0.01, 0.1, 1, 10, 100]

pmarcelino

# Concatenate training and validation sets
svm_features = np.concatenate((train_features, validation_features))
svm_labels = np.concatenate((train_labels, validation_labels))

pmarcelino

# Define model
from keras import models
from keras import layers
from keras import optimizers

epochs = 100

model = models.Sequential()
model.add(layers.GlobalAveragePooling2D(input_shape=(7,7,512)))
model.add(layers.Dense(1, activation='sigmoid'))

pmarcelino

# Define model
from keras import models
from keras import layers
from keras import optimizers

epochs = 100

model = models.Sequential()
model.add(layers.Flatten(input_shape=(7,7,512)))
model.add(layers.Dense(256, activation='relu', input_dim=(7*7*512)))

pmarcelino

# Create smaller dataset for Dogs vs. Cats
import os, shutil

original_dataset_dir = '/Users/macbook/dogs_cats_dataset/train/'

base_dir = '/Users/macbook/book/dogs_cats/data'
if not os.path.exists(base_dir):
    os.mkdir(base_dir)

# Create directories

pmarcelino

pmarcelino

pmarcelino

{
 "cells": [
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "# Missing Data"
   ]
  },
  {