# -*- coding: utf-8 -*-
from traffic_signs import load_data
import numpy as np
from sklearn.model_selection import train_test_split
from keras.models import Sequential
from keras.layers.core import Dense, Activation, Flatten
from keras.layers.convolutional import Conv2D, MaxPooling2D
#from keras.utils import to_categorical
#---------------------------- Q3 -----------------------------------------
X, y = load_data("GTSRB_subset_2")
X = X - np.min(X)
X = X / np.max(X)
y1 = np.int32(y == 0)
y2 = np.transpose(np.array([y, y1]))
X_train, X_test, y_train, y_test = train_test_split(X, y2, test_size=0.2)
#---------------------------- Q4 -----------------------------------------
model = Sequential()
N = 32 # Number of feature maps
w, h = 5, 5 # Conv. window size
plt.ylabel('w$_1$')
plt.title('Optimization path')
plt.subplot(212)
plt.plot(100.0 * np.array(accuracies), linewidth=2)
plt.ylabel('Accuracy / %')
plt.xlabel('Iteration')
plt.tight_layout()
plt.savefig("log_loss_minimization.pdf", bbox_inches="tight")
####################Q4####################
import traffic_signs
from sklearn.metrics import accuracy_score
import sklearn.model_selection
X, y = traffic_signs.load_data(".")
X_flat = []
for matrix in X:
X_flat.append(matrix.ravel())
X = np.float64(X_flat)
clf_list = [LogisticRegression(solver='liblinear'), SVC(gamma='scale')]
clf_name = ['LR', 'SVC']
C_range = [1e-5, 1e-4, 1e-3, 1e-2, 1e-1, 1]
X_train, X_test, y_train, y_test = sklearn.model_selection.train_test_split(
X, y, test_size=0.2, random_state=0)
for clf, name in zip(clf_list, clf_name):
for C in C_range:
for penalty in ["l1", "l2"]:
clf.C = C
clf.penalty = penalty
clf.fit(X_train, y_train)
"""
Created on Mon Feb 4 12:52:14 2019
@author: Asier
"""
import traffic_signs
import numpy as np
from sklearn.model_selection import train_test_split
from sklearn.svm import SVC
from sklearn.model_selection import GridSearchCV
from sklearn.model_selection import RandomizedSearchCV
C_range = [0.00001, 0.0001, 0.001, 0.01, 0.1, 1]
features = []
X, y = traffic_signs.load_data('./data')
#### Load or Extract features
#features = traffic_signs.extract_lbp_features(X)
features = np.load('./data/features.npy')
#features = np.array(features)
#features = features.reshape(features.shape[0], features.shape[1]*features.shape[2])
#X = X.reshape(X.shape[0], X.shape[1]*X.shape[2])
## Spliting data
X_train, X_test, y_train, y_test = train_test_split(features, y)
parameters = {'kernel': ('linear', 'rbf'), 'C': C_range}
svc = SVC(gamma="scale")
import traffic_signs
import numpy as np
from sklearn.model_selection import train_test_split
from keras.models import Sequential
from keras.layers import Dense, Flatten
from keras.layers import Conv2D, MaxPooling2D
from keras.utils import to_categorical
from keras.losses import categorical_crossentropy
from keras.optimizers import SGD
num_classes = 2
batch_size = 32
epochs = 20
X, y = traffic_signs.load_data('./input')
X_norm = ((X - np.amin(X)) / (np.amax(X) - np.amin(X)))
X_train, X_test, y_train, y_test = train_test_split(X_norm, y)
y_train = to_categorical(y_train, num_classes)
y_test = to_categorical(y_test, num_classes)
model = Sequential()
model.add(
Conv2D(32,
kernel_size=(5, 5),
activation='relu',
input_shape=X_norm.shape[1:],
padding='same')