def check_data():
dataset = data_input_fn(FLAGS, 'NOT NEEDED')
for imgs, gts in dataset:
print(imgs.shape, gts.shape)
for ii in range(imgs.shape[0]):
dat = np.concatenate((imgs[ii], gts[ii]), axis=1)
plt.imshow(dat)
plt.show()
def check_data():
dataset = data_input_fn(FLAGS)
COUNT = 10
counter = 0
for imgs, gts in dataset:
print(imgs.shape, gts.shape)
for ii in range(imgs.shape[0]):
counter += 1
dat = np.concatenate((imgs[ii], gts[ii]), axis=1)
plt.imshow(dat)
plt.show()
if counter >= COUNT:
break
if counter >= COUNT:
break
def eval_in_fn():
FLAGS.MODE = 'eval'
return data_input_fn(FLAGS, 'eval')
def train_in_fn():
return data_input_fn(FLAGS, 'train')