# validation dataset
dataset_val = u.get_dataset(coco_path, 'val')
gen_val = prepare(dataset_val, epochs, batch_size, input_shape, output_shape)
# fuu = next(gen_val)
# import ipdb; ipdb.set_trace()
# train dataset
dataset_train = u.get_dataset(coco_path, 'train')
gen_train = prepare(dataset_train, epochs, batch_size, input_shape,
output_shape)
callback = ModelSaveBestAvgAcc(filepath="model-{epoch:02d}-{avgacc:.2f}.hdf5",
verbose=True,
cond=filter_val('fmeasure'))
losses = []
for i in range(0, 1):
losses.append(binary_focal_loss(gamma=2.))
input_tensor = layers.Input(shape=input_shape)
model = get_model(input_tensor=input_tensor)
x = layers.Multiply()([input_tensor, model.output])
x = model(x)
# import ipdb; ipdb.set_trace()
model = models.Model(inputs=input_tensor, outputs=x)
model.compile(optimizer=opt.Adam(lr=1e-4),
fuu = next(gen_val)
# from keras.models import load_model
# model = load_model(
# '_demo-v14-model-09-0.97.hdf5',
# custom_objects={'precision': precision, 'evaluate': evaluate})
import ipdb
ipdb.set_trace()
# train dataset
dataset_train = u.get_dataset(coco_path, 'train')
gen_train = prepare(dataset_train, epochs, batch_size, input_shape,
os.path.join(coco_path, 'train_output'))
callback = ModelSaveBestAvgAcc(filepath="model-{epoch:02d}-{avgacc:.2f}.hdf5",
verbose=True,
cond=filter_val('evaluate'))
losses = []
losses.append(mse)
# for i in range(0, 1):
# losses.append(binary_focal_loss(gamma=2.))
# import ipdb; ipdb.set_trace()
model = get_model(input_shape, dataset_val.num_classes)
model.compile(optimizer=opt.Adam(lr=1e-4),
loss=losses,
metrics=['accuracy', precision, evaluate])
model.summary()