save_best_only=True,
save_weights_only=True)
hist = model.fit(train_set_R1,
Y_train,
validation_data=(test_set_R1, Y_test),
batch_size=16,
nb_epoch=jumEpoch,
shuffle=True,
verbose=1,
callbacks=[checkpointer])
# Evaluate the model
# load best model
model.load_weights(nama_filenya)
Y_pred = model.predict(test_set_R1, batch_size=8)
#print(Y_pred)
k_val = 1
Y_pred_label = []
for idt in range(len(Y_pred)):
Y_pred_label.append(np.argmax(Y_pred[idt]))
print Y_test.shape
print Y_pred.shape
print np.array(Y_pred_label).shape
print np.argmax(Y_test, axis=1)
print("Skor Model:")
accScore = accuracy_score(np.argmax(Y_test, axis=1), Y_pred_label)
print(accScore)
cohennya = cohen_kappa_score(np.argmax(Y_test, axis=1), Y_pred_label)
print("kohen kappa:")
metrics=['accuracy'])
validation_datagen = ImageDataGenerator(
featurewise_center=True,
featurewise_std_normalization=True,
zca_whitening=True)
validation_datagen.fit(X_train)
generator = validation_datagen.flow(X_evaluation,
Y_evaluation,
batch_size=1,
shuffle=False)
total_correct = 0
for sample_idx in range(X_evaluation.shape[0]):
(X, y) = generator.next()
preds = model.predict(X)
predicted_label = np.argmax(preds)
actual_label = y_evaluation[sample_idx] # np.argmax(y) #
# output the predicted label/image
y_evaluation[sample_idx] = predicted_label
cv2.imwrite(
"predicted_images/" + str(predicted_label) + "/" +
str(sample_idx) + ".png",
np.transpose((X_evaluation[sample_idx] + 1.) / 2. * 255.,
[1, 2, 0]))
# if predicted_label == actual_label:
# total_correct += 1
# print('Status: ', predicted_label == actual_label, \
df_pred = pd.DataFrame(list(zip(word)), columns=['Name'])
pred_img = df_pred['Name']
pred_img = np.array(list(pred_img))
#Normalize the data
pred_img = pred_img / 255
img_shape = pred_img.shape[0]
#reshape data to fit model
pred_img = pred_img.reshape(img_shape, 64, 64, 1)
model_path = Path + 'fm_cnn_BN16.h5'
# load the saved best model weights
model = load_model(model_path)
model.compile(loss=keras.losses.categorical_crossentropy,
optimizer=keras.optimizers.RMSprop(),
metrics=['accuracy'])
# predict outputs on validation images
prediction = model.predict(pred_img)
prediction = np.argmax(prediction, axis=1)
with open(Path + 'encoder4.pickle', 'rb') as handle:
label_encoder = pickle.load(handle)
letter = label_encoder.inverse_transform(prediction)
words = ''.join(letter)
print(words)
space = ' '
sent.append(words)
sent.append(space)
sentences = ''.join(sent)
img = norm_img(np.asarray(img_resize(img, 64)))
x_train.append(img)
imgs = np.asarray(x_train)
img_array -= 1
# vae losses / training
if (load_images_at_start):
imgs = x_train[img_array]
else:
imgs = np.asarray(x_train)
imgs_vaeimp = discriminator_vaeimp.predict(imgs)
enc_loss = vae_model.train_on_batch(imgs, imgs_vaeimp)
# discriminator losses / training
disc_loss1 = discriminator2.train_on_batch(imgs, np.ones((minibatch, 1)))
lcode = encoder.predict(imgs)[2]
lc_img = decoder.predict(lcode)
disc_loss2 = discriminator2.train_on_batch(lc_img, np.zeros(
(minibatch, 1)))
noise = np.random.normal(0, 1, (minibatch, latent_dim))
gen_imgs = decoder.predict(noise)
disc_loss3 = discriminator2.train_on_batch(gen_imgs,
np.zeros((minibatch, 1)))
# generator losses / training
img_array = np.random.randint(1, 202600, batch_size)
if (load_images_at_start):
img_array = img_array - 1
img_array2 = img_array[0:minibatch]
imgs = x_train[img_array2]
else:
