from conv_net import build_network
args = GetOptions()
if args['activation'] == 'rectify':
activation_function = rectify
elif args['activation'] == 'leaky_rectify':
activation_function = leaky_rectify
elif args['activation'] == 'very_leaky_rectify':
activation_function = very_leaky_rectify
else:
raise ValueError('Unknown activation function')
args['activation_function'] = activation_function
X, y, images_id = load_numpy_arrays(args['train_file'])
sample_size = y.shape[0] - y.shape[0] % args['batch_size']
X = X[:sample_size]
y = y[:sample_size]
print "Train:"
print "X.shape:", X.shape
print "y.shape:", y.shape
y_counts = np.unique(y, return_counts=True)[1]
print "y value counts: ", y_counts
print "pictures size: ", sqrt(X.shape[1]/3.)
args['dataset_ratio'] = 3.8
args['network'] = 'AlexNet'
args['batch_size'] = 141
__author__ = 'thiebaut'
__date__ = '07/11/15'
args = GetOptions()
if args['activation'] == 'rectify':
activation_function = rectify
elif args['activation'] == 'leaky_rectify':
activation_function = leaky_rectify
elif args['activation'] == 'very_leaky_rectify':
activation_function = very_leaky_rectify
else:
raise ValueError('Unknown activation function')
args['activation_function'] = activation_function
X, y, images_id = load_numpy_arrays(args['train_file'])
sample_size = y.shape[0] - y.shape[0] % args['batch_size']
X = X[:sample_size]
y = y[:sample_size]
print "Train:"
print "X.shape:", X.shape
print "y.shape:", y.shape
y_counts = np.unique(y, return_counts=True)[1]
print "y value counts: ", y_counts
print "pictures size: ", sqrt(X.shape[1]/3.)
# Compute over-sampling of class 1
dataset_ratio = float(y_counts[1])/y_counts[0]
print "Labels ratio: {:.2f}".format(dataset_ratio)
if args['initial_ratio'] is None:
#!/usr/bin/env python
# -*- coding: utf-8 -*-
from utils import make_submission_file
from utils import load_numpy_arrays
from datetime import date
import cPickle
import sys
conv_net = cPickle.load(open(str(sys.argv[1]),'rb'))
# ----- Test set ----
X_test, _, images_id_test = load_numpy_arrays('test.npz')
print "Test:"
print "X_test.shape:", X_test.shape
predictions = conv_net.predict_proba(X_test)
make_submission_file(predictions, images_id_test, output_filepath='submissions/submission_'+str(date.today)+'.csv')
