def saveParameters(self):
"""
Saves the parameters as tiktorch Object
"""
self.code_path = str(self.code_path_textbox.text())
self.model_class_name = str(self.model_class_name_textbox.text())
self.state_path = str(self.state_path_textbox.text())
self.input_shape = [
int(x) for x in self.input_shape_textbox.text()[1:-1].replace(
", ", ",").split(",")
]
self.minimal_increment = [
int(x)
for x in self.minimal_increment_textbox.text()[1:-1].replace(
", ", ",").split(",")
]
self.model_init_kwargs = yaml.load(
str(self.model_init_kwargs_textbox.text())[1:-1].replace(
", ", "\n"))
self.model_path = str(self.model_path_textbox.text())
spec = TikTorchSpec(
code_path=self.code_path,
model_class_name=self.model_class_name,
state_path=self.state_path,
input_shape=self.input_shape,
minimal_increment=self.minimal_increment,
model_init_kwargs=self.model_init_kwargs,
)
buildface = BuildSpec(self.model_path)
buildface.build(spec)
def test_BuildUNet2d(self):
spec = TikTorchSpec(code_path='/home/jo/sfb1129/pretrained_net_constantin/ISBI2012_UNet_pretrained/model.py',
model_class_name='UNet2dGN',
state_path='/home/jo/sfb1129/pretrained_net_constantin/ISBI2012_UNet_pretrained/state.nn',
input_shape=(1, 572, 572),
minimal_increment=[32, 32],
model_init_kwargs={'in_channels': 1, 'out_channels': 1, 'initial_features': 64})
self.spec.validate()
build_spec = BuildSpec(build_directory='/home/jo/ISBI_UNet_pretrained', device='cpu')
build_spec.build(self.spec)
def test_BuilDUNet3d(self):
spec = TikTorchSpec(code_path='/home/jo/uni/master-models/master_models/models/dunet3D.py',
model_class_name='DUNet3D',
state_path='/home/jo/uni/master-models/master_models/results/dunet3D/trained_net/best_model_dunet3D.torch',
input_shape=[1, 512, 512],
minimal_increment=[32, 32],
model_init_kwargs={'in_channels': 1, 'out_channels': 1})
self.spec.validate()
build_spec = BuildSpec(build_directory='/home/jo/CREMI_DUNet_pretrained', device='cpu')
build_spec.build(self.spec)
def test_BuildDUNet2d(self):
spec = TikTorchSpec(code_path='/home/jo/config/model.py',
model_class_name='DUNet2D',
state_path='/home/jo/config/state.nn',
input_shape=[1, 512, 512],
minimal_increment=[32, 32],
model_init_kwargs={'in_channels': 1, 'out_channels': 1})
self.spec.validate()
build_spec = BuildSpec(build_directory='/home/jo/CREMI_DUNet_pretrained', device='cpu')
build_spec.build(self.spec)
def test_BuildDUNet2d(self):
spec = TikTorchSpec(
code_path="/home/jo/config/model.py",
model_class_name="DUNet2D",
state_path="/home/jo/config/state.nn",
input_shape=[1, 512, 512],
minimal_increment=[32, 32],
model_init_kwargs={
"in_channels": 1,
"out_channels": 1
},
)
self.spec.validate()
build_spec = BuildSpec(
build_directory="/home/jo/CREMI_DUNet_pretrained", device="cpu")
build_spec.build(self.spec)
def test_BuilDUNet3d(self):
spec = TikTorchSpec(
code_path=
"/home/jo/uni/master-models/master_models/models/dunet3D.py",
model_class_name="DUNet3D",
state_path=
"/home/jo/uni/master-models/master_models/results/dunet3D/trained_net/best_model_dunet3D.torch",
input_shape=[1, 512, 512],
minimal_increment=[32, 32],
model_init_kwargs={
"in_channels": 1,
"out_channels": 1
},
)
self.spec.validate()
build_spec = BuildSpec(
build_directory="/home/jo/CREMI_DUNet_pretrained", device="cpu")
build_spec.build(self.spec)
def test_BuildUNet2d(self):
spec = TikTorchSpec(
code_path=
"/home/jo/sfb1129/pretrained_net_constantin/ISBI2012_UNet_pretrained/model.py",
model_class_name="UNet2dGN",
state_path=
"/home/jo/sfb1129/pretrained_net_constantin/ISBI2012_UNet_pretrained/state.nn",
input_shape=(1, 572, 572),
minimal_increment=[32, 32],
model_init_kwargs={
"in_channels": 1,
"out_channels": 1,
"initial_features": 64
},
)
self.spec.validate()
build_spec = BuildSpec(build_directory="/home/jo/ISBI_UNet_pretrained",
device="cpu")
build_spec.build(self.spec)