def test_load_model(self):
os.environ['CUDA_VISIBLE_DEVICES'] = ''
img_path = os.path.join(base_path, '../test_data/shapes/pixels/*')
label_path = os.path.join(base_path, '../test_data/shapes/labels.ilp')
savepath = os.path.join(base_path, '../test_data/tmp')
os.makedirs(savepath, exist_ok=True)
# train initial model and save it
t = Session()
t.load_training_data(img_path, label_path)
t.make_model('convnet_for_unittest', (1, 100, 100))
t.train(max_epochs=3,
steps_per_epoch=2,
log_filename=os.path.join(savepath, 'log.csv'),
model_filename=os.path.join(savepath, 'model.h5'))
# initialize new sassion and load pretrained model
t2 = Session()
t2.load_training_data(img_path, label_path)
assert t2.model is None
t2.load_model(os.path.join(savepath, 'model.h5'))
assert t2.model is not None
# train model further
t2.train(max_epochs=3,
steps_per_epoch=2,
log_filename=os.path.join(savepath, 'log.csv'),
model_filename=os.path.join(savepath, 'model_continued.h5'))
def test_shape_data_fast(self):
# train a classifier and predict training data
os.environ['CUDA_VISIBLE_DEVICES'] = '2'
img_path = os.path.join(base_path, '../test_data/shapes/pixels/*')
label_path = os.path.join(base_path, '../test_data/shapes/labels.ilp')
savepath = os.path.join(base_path, '../test_data/tmp')
os.makedirs(savepath, exist_ok=True)
t = Session()
t.load_training_data(img_path, label_path)
t.make_model('convnet_for_unittest', (1, 100, 100))
t.train(max_epochs=3,
steps_per_epoch=4,
log_filename=os.path.join(savepath, 'log.csv'),
model_filename=os.path.join(savepath, 'model.h5'))
t.load_prediction_data(img_path, savepath)
t.predict()
artifacts = os.listdir(savepath)
assert 'log.csv' in artifacts
assert 'pixels_1_class_1.tif' in artifacts
assert 'pixels_1_class_2.tif' in artifacts
assert 'pixels_1_class_3.tif' in artifacts
assert 'pixels_2_class_1.tif' in artifacts
assert 'pixels_2_class_2.tif' in artifacts
assert 'pixels_2_class_3.tif' in artifacts
shutil.rmtree(savepath)
def test_set_augmentation(self):
img_path = os.path.join(base_path, '../test_data/shapes/pixels/*')
label_path = os.path.join(base_path, '../test_data/shapes/labels.ilp')
t = Session()
t.load_training_data(img_path, label_path)
t.make_model('unet_2d', (1, 572, 572))
t.set_augmentation('flip+rotate')
self.assertEqual(t.data.augmentation, {'rotate', 'flip'})
t.set_augmentation('flip+rotate+shear')
self.assertEqual(t.data.augmentation, {'rotate', 'flip', 'shear'})
t.set_augmentation('flip')
self.assertEqual(t.data.augmentation, {'flip'})
def test_set_normalization(self):
img_path = os.path.join(base_path, '../test_data/shapes/pixels/*')
label_path = os.path.join(base_path, '../test_data/shapes/labels.ilp')
t = Session()
t.load_training_data(img_path, label_path)
t.make_model('unet_2d', (1, 572, 572))
t.set_normalization('local')
assert t.data.normalize_mode == 'local'
assert t.data.global_norm_minmax is None
t.set_normalization('global_0+255')
assert t.data.normalize_mode == 'global'
assert t.data.global_norm_minmax == (0, 255)
def train_test_model_convnet():
img_path = os.path.join(base_path, '../test_data/shapes/pixels/*')
label_path = os.path.join(base_path, '../test_data/shapes/labels.h5')
savepath = os.path.join(base_path, '../test_data/tmp')
model_filename = os.path.join(savepath, 'model_convnet.h5')
if os.path.isfile(model_filename):
return model_filename
os.makedirs(savepath, exist_ok=True)
t = Session()
t.load_training_data(img_path, label_path)
t.make_model('convnet_for_unittest', (1, 100, 100))
t.train(max_epochs=2,
steps_per_epoch=2,
log_filename=os.path.join(savepath, 'log.csv'),
model_filename=model_filename)
return model_filename
def train_test_model_unet_2d_3channels_2classes():
img_path = os.path.join(base_path, '../test_data/shapes/pixels_rgb/*')
label_path = os.path.join(base_path,
'../test_data/shapes/labels_2classes.h5')
savepath = os.path.join(base_path, '../test_data/tmp')
model_filename = os.path.join(savepath,
'model_unet_2d_3channels_2classes.h5')
if os.path.isfile(model_filename):
return model_filename
os.makedirs(savepath, exist_ok=True)
t = Session()
t.load_training_data(img_path, label_path)
t.make_model('unet_2d', (1, 220, 220))
t.train(max_epochs=2,
steps_per_epoch=2,
log_filename=os.path.join(savepath, 'log.csv'),
model_filename=model_filename)
return model_filename
def test_shape_data(self):
# train a classifier and predict training data
os.environ['CUDA_VISIBLE_DEVICES'] = '2'
img_path = os.path.join(
base_path,
'../test_data/shapes/pixels/*')
label_path = os.path.join(
base_path,
'../test_data/shapes/labels.h5')
savepath = os.path.join(
base_path,
'../test_data/tmp')
os.makedirs(savepath, exist_ok=True)
t = Session()
t.load_training_data(img_path, label_path)
t.make_model('convnet_for_unittest', (1, 100, 100))
t.train(max_epochs=50,
steps_per_epoch=50,
log_filename=os.path.join(savepath, 'log.csv'),
model_filename=os.path.join(savepath, 'model.h5'))
t.load_prediction_data(img_path, savepath)
t.predict()
# read prediction images and compare with validation data
def read_images(image_nr, class_nr):
if class_nr == 1:
shape = 'circles'
if class_nr == 2:
shape = 'triangles'
filename = os.path.join(
savepath,
'pixels_{}_class_{}.tif'.format(image_nr,
class_nr))
print(filename)
prediction_img = np.squeeze(imread(filename))
filename = os.path.join(
savepath,
'../shapes/val/{}_{}.tiff'.format(
shape,
image_nr))
print(filename)
val_img = np.squeeze(skimage.io.imread(filename))
return prediction_img, val_img
prediction_img, val_img = read_images(1, 1)
accuracy = np.mean(prediction_img[val_img > 0][:])
print(accuracy)
prediction_img, val_img = read_images(1, 2)
accuracy = np.mean(prediction_img[val_img > 0][:])
print(accuracy)
prediction_img, val_img = read_images(2, 1)
accuracy = np.mean(prediction_img[val_img > 0][:])
print(accuracy)
prediction_img, val_img = read_images(2, 2)
accuracy = np.mean(prediction_img[val_img > 0][:])
print(accuracy)
shutil.rmtree(savepath)
def main(args):
if args['--cpu']:
# deactivate gpu for tensorflow
os.environ['CUDA_VISIBLE_DEVICES'] = ''
if args['--gpu']:
# define gpu hardware
os.environ['CUDA_VISIBLE_DEVICES'] = args['--gpu']
image_path = os.path.abspath(os.path.expanduser(args['<image_path>']))
model_name = args['<network>']
if args['deploy']:
output_path = args['<output_path>']
predict_example_image = not args['--skip-predict']
assert os.path.isfile(model_name)
exp = DeepimagejExporter(model_name, output_path, image_path)
exp.export_as_deepimagej(author=args['--author'],
version=args['--version'],
url=args['--url'],
credit=args['--credit'],
reference=args['--reference'],
size=args['--size'],
applymodel=predict_example_image)
return
norm_string = args['--normalize']
s = Session()
if args['train']:
label_path = os.path.abspath(os.path.expanduser(args['<label_path>']))
aug_string = args['--augment']
max_epochs = int(args['--epochs'])
steps_per_epoch = int(args['--steps'])
log_filename = args['--csvfile']
model_export_filename = args['--file']
valfraction = float(args['--valfraction'])
if args['--batchsize']:
batch_size = int(args['--batchsize'])
else:
batch_size = None
s.load_training_data(image_path, label_path, batch_size)
models_available = ['unet_2d', 'unet_multi_z', 'convnet_for_unittest']
if os.path.isfile(model_name):
s.load_model(model_name)
elif model_name in models_available:
size_xy = 572
if model_name == 'unet_2d' or model_name == 'convnet_for_unittest':
size_z = 1
if model_name == 'unet_multi_z':
size_z = 5
if model_name == 'convnet_for_unittest':
size_xy = 100
s.make_model(model_name, (size_z, size_xy, size_xy))
s.set_normalization(norm_string)
s.set_augmentation(aug_string)
if valfraction > 0:
s.define_validation_data(valfraction)
s.train(max_epochs=max_epochs,
steps_per_epoch=steps_per_epoch,
log_filename=log_filename,
model_filename=model_export_filename)
f = h5py.File(model_export_filename, 'r+')
lbl_map_list = [[key, val] for key, val in s.lbl_map.items()]
f.create_dataset('lbl_map', data=lbl_map_list)
if args['predict']:
output_path = args['<output_path>']
assert os.path.isfile(model_name), '<network> must be a h5 model file'
multichannel = not (args['--split'])
s.load_prediction_data(image_path, output_path)
s.load_model(model_name)
s.set_normalization(norm_string)
s.predict(multichannel=multichannel)
def main(args):
s = Session()
if args['--cpu']:
# deactivate gpu for tensorflow
os.environ['CUDA_VISIBLE_DEVICES'] = ''
if args['--gpu']:
# define gpu hardware
os.environ['CUDA_VISIBLE_DEVICES'] = args['--gpu']
image_path = os.path.abspath(os.path.expanduser(args['<image_path>']))
model_name = args['<network>']
norm_string = args['--normalize']
if args['train']:
label_path = os.path.abspath(os.path.expanduser(args['<label_path>']))
aug_string = args['--augment']
max_epochs = int(args['--epochs'])
steps_per_epoch = int(args['--steps'])
log_filename = args['--csvfile']
model_export_filename = args['--file']
valfraction = float(args['--valfraction'])
s.load_training_data(image_path, label_path)
models_available = ['unet_2d',
'unet_multi_z',
'convnet_for_unittest']
if os.path.isfile(model_name):
s.load_model(model_name)
elif model_name in models_available:
size_xy = 572
if model_name == 'unet_2d' or model_name == 'convnet_for_unittest':
size_z = 1
if model_name == 'unet_multi_z':
size_z = 5
if model_name == 'convnet_for_unittest':
size_xy = 100
s.make_model(model_name, (size_z, size_xy, size_xy))
s.set_normalization(norm_string)
s.set_augmentation(aug_string)
if valfraction > 0:
s.define_validation_data(valfraction)
s.train(max_epochs=max_epochs,
steps_per_epoch=steps_per_epoch,
log_filename=log_filename,
model_filename=model_export_filename)
if args['predict']:
output_path = args['<output_path>']
assert os.path.isfile(model_name), '<network> must be a h5 model file'
s.load_prediction_data(image_path, output_path)
s.load_model(model_name)
s.set_normalization(norm_string)
s.predict()