def getModel(num_img, img_height, img_width, output_dim, weights_path):
"""
Initialize model.
# Arguments
img_width: Target image widht.
img_height: Target image height.
num_img: Target images per block
output_dim: Dimension of model output (number of classes).
weights_path: Path to pre-trained model.
# Returns
model: A Model instance.
"""
model = nets.resnet50(num_img, img_height, img_width, output_dim)
if weights_path:
try:
model.load_weights(weights_path)
print("Loaded model from {}".format(weights_path))
except:
print("Impossible to find weight path. Returning untrained model")
return model
plt.imshow(X0[0][i - 1])
plt.show()
else:
for i in range(1, columns * rows + 1):
fig.add_subplot(rows, columns, i)
if args['mode'] == 'Grey':
plt.imshow(X0[i - 1][:, :, 0], cmap='gray')
else:
plt.imshow(X0[i - 1])
plt.show()
# if args['mode'] == 'Grey':
# model = get_model(args['mode'], n_sequence, n_output, dim, channels=1)
# else:
# model = get_model(args['mode'], n_sequence, n_output, dim, channels=3)
single_model = resnet50(Input(shape=(64, 168, 64, 3)), num_classes=3)
# model = MobileNetV3_Large((n_sequence, dim[0], dim[1], n_channels), n_output).build()
# model = MobileNetV3_Small((n_sequence, dim[0], dim[1], n_channels), n_output).build()
# model = ShuffleNetV2(input_shape = (n_sequence, dim[0], dim[1], n_channels), classes=n_output)
# model = Effnet((n_sequence, dim[0], dim[1], n_channels), n_output)
# model = iotnet3((16,96,96,3), 1, n=4, k=0.7)
# Load weight of unfinish training model(optional)
load_model = True
start_epoch = 0
if load_model:
weights_path = 'save_weight/weight-41-0.98-0.64-0.45690.hdf5' # name of model
start_epoch = 41
single_model.load_weights(weights_path)
# model = load_model(weights_path)