# instantiate datagenerator class
test_set = Dataloader(image_paths=test_imagepath,
mask_paths=test_maskpath,
vector_paths=test_vector_path,
image_size=input_size,
numclasses=num_of_Classes,
channels=[3, 3],
palette=palette,
seed=47)
# evaluation parameters
BS = 1
# initialize data generators with threading lock
testgen = threadsafe_iter(
test_set.data_gen(should_augment=False, batch_size=BS))
No_of_test_images = len(os.listdir(test_imagepath))
print("Number of Test Images = {}".format(No_of_test_images))
eval = evaluation(input_size,
num_of_Classes,
palette,
pretrained_unet='output/Checkpoint_Weights.h5',
pretrained_fcn='models/FCN_weights.h5',
pretrained_segnet='models/SegNet_weights.h5',
unet=True,
fcn=False,
segnet=False)
eval.evaluate(testgenerator=testgen,
numclasses=num_of_Classes,
channels=[3, 3],
palette=palette,
seed=47)
# Build Model
model_FCN = FCN.build((352, 352, 3), num_of_Classes, pretrained_weights="FCN_weights.h5")
model_FCN.summary()
# learning parameters
INIT_LR = 0.001
EPOCHS = 100
BS = 2
# initialize data generators
traingen = threadsafe_iter(train_set.data_gen(should_augment=True, batch_size=BS))
valgen = threadsafe_iter(val_set.data_gen(should_augment=False, batch_size=BS))
# Print INFO
No_of_train_images = len(os.listdir(dataset_path + '/train_frames/train'))
No_of_val_images = len(os.listdir(dataset_path + '/val_frames/val'))
print("Number of Training Images = {}".format(No_of_train_images))
# optimizer
opt = Adam(lr=INIT_LR, decay=0.0000001)
# early stopping parameters and Saving Training Data
weights_path = os.getcwd() + '/output/weights.h5'
checkpoint = ModelCheckpoint(weights_path, monitor='val_loss', verbose=1, save_best_only=True, mode='min')
csvlogger = CSVLogger('./log_FCN.out', append=True, separator=';')
earlystopping = EarlyStopping(monitor='val_loss', verbose=1, min_delta=0.001, patience=50, mode='min', baseline=0.5)