monailabel.tasks.infer.basic_infer module

class monailabel.tasks.infer.basic_infer.BasicInferTask(path, network, type, labels, dimension, description, model_state_dict='model', input_key='image', output_label_key='pred', output_json_key='result', config=None, load_strict=True, roi_size=None, preload=False, train_mode=False, skip_writer=False)[source]

Bases: monailabel.interfaces.tasks.infer_v2.InferTask

Basic Inference Task Helper

Parameters

  • path (Union[None, str, Sequence[str]]) – Model File Path. Supports multiple paths to support versions (Last item will be picked as latest)

  • network (Optional[Any]) – Model Network (e.g. monai.networks.xyz). None in case if you use TorchScript (torch.jit).

  • type (Union[str, InferType]) – Type of Infer (segmentation, deepgrow etc..)

  • labels (Union[str, None, Sequence[str], Dict[Any, Any]]) – Labels associated to this Infer

  • dimension (int) – Input dimension

  • description (str) – Description

  • model_state_dict (str) – Key for loading the model state from checkpoint

  • input_key (str) – Input key for running inference

  • output_label_key (str) – Output key for storing result/label of inference

  • output_json_key (str) – Output key for storing result/label of inference

  • config (Optional[Dict[str, Any]]) – K,V pairs to be part of user config

  • load_strict (bool) – Load model in strict mode

  • roi_size – ROI size for scanning window inference

  • preload – Preload model/network on all available GPU devices

  • train_mode – Run in Train mode instead of eval (when network has dropouts)

  • skip_writer – Skip Writer and return data dictionary

__init__(path, network, type, labels, dimension, description, model_state_dict='model', input_key='image', output_label_key='pred', output_json_key='result', config=None, load_strict=True, roi_size=None, preload=False, train_mode=False, skip_writer=False)[source]
Parameters

  • path (Union[None, str, Sequence[str]]) – Model File Path. Supports multiple paths to support versions (Last item will be picked as latest)

  • network (Optional[Any]) – Model Network (e.g. monai.networks.xyz). None in case if you use TorchScript (torch.jit).

  • type (Union[str, InferType]) – Type of Infer (segmentation, deepgrow etc..)

  • labels (Union[str, None, Sequence[str], Dict[Any, Any]]) – Labels associated to this Infer

  • dimension (int) – Input dimension

  • description (str) – Description

  • model_state_dict (str) – Key for loading the model state from checkpoint

  • input_key (str) – Input key for running inference

  • output_label_key (str) – Output key for storing result/label of inference

  • output_json_key (str) – Output key for storing result/label of inference

  • config (Optional[Dict[str, Any]]) – K,V pairs to be part of user config

  • load_strict (bool) – Load model in strict mode

  • roi_size – ROI size for scanning window inference

  • preload – Preload model/network on all available GPU devices

  • train_mode – Run in Train mode instead of eval (when network has dropouts)

  • skip_writer – Skip Writer and return data dictionary

add_cache_transform(t, data, keys=('image', 'image_meta_dict'), hash_key=('image_path', 'model'))[source]
clear()[source]
clear_cache()[source]
config()[source]
Return type

Dict[str, Any]

detector(data=None)[source]
Return type

Optional[Callable]

get_path(validate=True)[source]
inferer(data=None)[source]
Return type

Inferer

info()[source]
Return type

Dict[str, Any]

inverse_transforms(data=None)[source]

Provide List of inverse-transforms. They are normally subset of pre-transforms. This task is performed on output_label (using the references from input_key)

Parameters

data – current data dictionary/request which can be helpful to define the transforms per-request basis

Return one of the following.

  • None: Return None to disable running any inverse transforms (default behavior).

  • Empty: Return [] to run all applicable pre-transforms which has inverse method

  • list: Return list of specific pre-transforms names/classes to run inverse method

For Example:

return [
    monai.transforms.Spacingd,
]
Return type

Optional[Sequence[Callable]]

is_valid()[source]
Return type

bool

abstract post_transforms(data=None)[source]

Provide List of post-transforms

Parameters

data –

current data dictionary/request which can be helpful to define the transforms per-request basis

For Example:

return [
    monai.transforms.EnsureChannelFirstd(keys='pred', channel_dim='no_channel'),
    monai.transforms.Activationsd(keys='pred', softmax=True),
    monai.transforms.AsDiscreted(keys='pred', argmax=True),
    monai.transforms.SqueezeDimd(keys='pred', dim=0),
    monai.transforms.ToNumpyd(keys='pred'),
    monailabel.interface.utils.Restored(keys='pred', ref_image='image'),
    monailabel.interface.utils.ExtremePointsd(keys='pred', result='result', points='points'),
    monailabel.interface.utils.BoundingBoxd(keys='pred', result='result', bbox='bbox'),
]

Return type

Sequence[Callable]

abstract pre_transforms(data=None)[source]

Provide List of pre-transforms

Parameters

data –

current data dictionary/request which can be helpful to define the transforms per-request basis

For Example:

return [
    monai.transforms.LoadImaged(keys='image'),
    monai.transforms.EnsureChannelFirstd(keys='image', channel_dim='no_channel'),
    monai.transforms.Spacingd(keys='image', pixdim=[1.0, 1.0, 1.0]),
    monai.transforms.ScaleIntensityRanged(keys='image',
        a_min=-57, a_max=164, b_min=0.0, b_max=1.0, clip=True),
]

Return type

Sequence[Callable]

run_detector(data, convert_to_batch=True, device='cuda')[source]

Run Detector over pre-processed Data. Derive this logic to customize the normal behavior. In some cases, you want to implement your own for running chained inferers over pre-processed data

Parameters

  • data (Dict[str, Any]) – pre-processed data

  • convert_to_batch – convert input to batched input

  • device – device type run load the model and run inferer

Returns

updated data with output_key stored that will be used for post-processing

run_inferer(data, convert_to_batch=True, device='cuda')[source]

Run Inferer over pre-processed Data. Derive this logic to customize the normal behavior. In some cases, you want to implement your own for running chained inferers over pre-processed data

Parameters

  • data (Dict[str, Any]) – pre-processed data

  • convert_to_batch – convert input to batched input

  • device – device type run load the model and run inferer

Returns

updated data with output_key stored that will be used for post-processing

run_invert_transforms(data, pre_transforms, names)[source]
run_post_transforms(data, transforms)[source]
run_pre_transforms(data, transforms)[source]
set_loglevel(level)[source]
writer(data, extension=None, dtype=None)[source]

You can provide your own writer. However, this writer saves the prediction/label mask to file and fetches result json

Parameters

  • data (Dict[str, Any]) – typically it is post processed data

  • extension – output label extension

  • dtype – output label dtype

Return type

Tuple[Any, Any]

Returns

tuple of output_file and result_json

class monailabel.tasks.infer.basic_infer.CallBackTypes(value)[source]

Bases: str, enum.Enum

An enumeration.

INFERER = 'INFERER'
INVERT_TRANSFORMS = 'INVERT_TRANSFORMS'
POST_TRANSFORMS = 'POST_TRANSFORMS'
PRE_TRANSFORMS = 'PRE_TRANSFORMS'
WRITER = 'WRITER'

previous

monailabel.tasks.infer package

next

monailabel.tasks.infer.bundle module