FastSurferCNN.inference

class FastSurferCNN.inference.Inference(cfg, device, ckpt='', lut=None)

Model evaluation class to run inference using FastSurferCNN.

Attributes

permute_order	(Dict[str, Tuple[int, int, int, int]]) Permutation order for axial, coronal, and sagittal
device	(Optional[torch.device])) Device specification for distributed computation usage.
default_device	(torch.device) Default device specification for distributed computation usage.
cfg	(yacs.config.CfgNode) Configuration Node
model_parallel	(bool) Option for parallel run
model	(torch.nn.Module) Neural network model
model_name	(str) Name of the model
alpha	(Dict[str, float]) Alpha values for different planes.
post_prediction_mapping_hook	Hook for post prediction mapping.

Methods

setup_model([cfg, device])	Set up the model.
set_cfg(cfg)	Set the configuration node.
to([device])	Move and/or cast the parameters and buffers.
load_checkpoint(ckpt)	Load the checkpoint and set device and model.
eval(init_pred, val_loader, *[, out_scale, out])	Perform prediction and inplace-aggregate views into pred_prob.
run(init_pred, img_filename, orig_data, ...)	Run the loaded model on the data (T1) from orig_data and img_filename (for messages only) with scale factors orig_zoom.

eval(init_pred, val_loader, *, out_scale=None, out=None)

Perform prediction and inplace-aggregate views into pred_prob.

Parameters:

init_predtorch.Tensor

Initial prediction.

val_loaderDataLoader

Validation loader.

out_scaleOptional

Output scale (Default value = None).

outOptional[torch.Tensor]

Previous prediction tensor (Default value = None).

Returns:

torch.Tensor

Prediction probability tensor.

get_cfg()

Return the configurations.

Returns:

yacs.config.CfgNode

Configuration node.

get_device()

Return the device.

Returns:

torch.device

The device used for computation.

get_max_size()

Return the max size.

Returns:

int | tuple[int, int]

The maximum size, either a single value or a tuple (width, height).

get_model_height()

Return the model height.

Returns:

int

The height of the model.

get_model_width()

Return the model width.

Returns:

int

The width of the model.

get_modelname()

Return the model name.

Returns:

str

The name of the model.

get_num_classes()

Return the number of classes.

Returns:

int

The number of classes.

get_plane()

Return the plane.

Returns:

str

The plane used in the model.

load_checkpoint(ckpt)

Load the checkpoint and set device and model.

Parameters:

ckptUnion[str, os.PathLike]

String or os.PathLike object containing the name to the checkpoint file.

run(init_pred, img_filename, orig_data, orig_zoom, out=None, out_res=None, batch_size=None)

Run the loaded model on the data (T1) from orig_data and img_filename (for messages only) with scale factors orig_zoom.

Parameters:

init_predtorch.Tensor

Initial prediction.

img_filenamestr

Original image filename.

orig_datanpt.NDArray

Original image data.

orig_zoomnpt.NDArray

Original zoom.

outOptional[torch.Tensor]

Updated output tensor (Default = None).

out_resOptional[int]

Output resolution (Default value = None).

batch_sizeint

Batch size (Default = None).

Returns:

toch.Tensor

Prediction probability tensor.

set_cfg(cfg)

Set the configuration node.

Parameters:

cfgyacs.config.CfgNode

Configuration node.

setup_model(cfg=None, device=None)

Set up the model.

Parameters:

cfgyacs.config.CfgNode

Configuration Node (Default value = None).

devicetorch.device

Device specification for distributed computation usage. (Default value = None).

to(device=None)

Move and/or cast the parameters and buffers.

Parameters:

deviceOptional[torch.device]

The desired device of the parameters and buffers in this module (Default value = None).