bird_classifier_train.py

from __future__ import division, print_function, absolute_import

import tflearn
from tflearn.data_utils import shuffle
from tflean.layers.core import input_data, dropout, fully_connected
from tflearn.layers.conv import conv_2d, max_pool_2d
from tflearn.layers.estimator import regression
from tflearn.data_preprocessing import ImageProcessing
from tflearn.data_augmentation import ImageAugmentation
import pickle


X, Y, X_test, Y_test = pickle.load(open("full_dataset.pkl","rb"))

X, Y = shuffle(X,Y)

img_prep = ImageProcessing()
img_prep.add_featurewise_zero_center()
img_prep.add_featurewise_stdnorm()


img_aug = ImageAugmentation()
img_aug.add_random_flip_leftright()
img_aug.add_random_rotation(max_angle=25.)
img_aug.add_random_blur(sigma_max=3.)


#define network architecture

network = input_data(shape=[None,32,32,3],
					  data_preprocessing=img_prep,
					  data_augmentation=img_aug)


network = conv_2d(network, 32, 3, activation='relu')

network = max_pool_2d(network,2)

network = conv_2d(network, 64, 3, activation='relu')

network = conv_2d(network, 64, 3, activation='relu')

network = max_pool_2d(network,2)

network = fully_connected(network, 512, activation='relu')

network = dropout(network, 0.5)

network = fully_connected(network,2, activation='softmax')

network = regression(network, optimizer='adam',
					 loss='categorical_crossentropy',
					 learning_rate=0.001)

model = tflearn.DNN(network, tensorboard_verbose=0, checkpoint_path='bird-classifier.tfl.ckpt')

model.fit(X, Y, n_epoch=100, shuffle=True, validation_set=(X_test,Y_testY),
			show_metric=True, batch_size=96,
			snapshot_epoch=True,
			run_id='bird-classifier')

model.save("bird-classifier.tfl")

def unpickle(file):
    import cPickle
    with open(file, 'rb') as fo:
        dict = cPickle.load(fo)
    return dict