configs/inference.yaml

# This implements the workflow for applying the network to a directory of images and saving the predicted segmentations.

imports:
- $import os
- $import torch
- $import glob

# pull out some constants from MONAI
image: $monai.utils.CommonKeys.IMAGE
pred: $monai.utils.CommonKeys.PRED

# hyperparameters for you to modify on the command line
batch_size: 1  # number of images per batch
num_workers: 0  # number of workers to generate batches with
num_classes: 4  # number of classes in training data which network should predict
device: $torch.device('cuda:0' if torch.cuda.is_available() else 'cpu')

# define various paths
bundle_root: .  # root directory of the bundle
ckpt_path: $@bundle_root + '/models/model.pt'  # checkpoint to load before starting
dataset_dir: $@bundle_root + '/test_data'  # where data is coming from
output_dir: './outputs'  # directory to store images to

# network definition, this could be parameterised by pre-defined values or on the command line
network_def:
  _target_: UNet
  spatial_dims: 3
  in_channels: 1
  out_channels: '@num_classes'
  channels: [8, 16, 32, 64]
  strides: [2, 2, 2]
  num_res_units: 2
network: $@network_def.to(@device)

# list all niftis in the input directory
file_pattern: '*.nii*'
data_list: '$list(sorted(glob.glob(os.path.join(@dataset_dir, @file_pattern))))'
# collect data dictionaries for all files
data_dicts: '$[{@image:i} for i in @data_list]'

# these transforms are used for inference to load and regularise inputs
transforms:
- _target_: LoadImaged
  keys: '@image'
  image_only: true
- _target_: EnsureChannelFirstd
  keys: '@image'
- _target_: ScaleIntensityd
  keys: '@image'

preprocessing:
  _target_: Compose
  transforms: $@transforms

dataset:
  _target_: Dataset
  data: '@data_dicts'
  transform: '@preprocessing'

dataloader:
  _target_: ThreadDataLoader  # generate data ansynchronously from inference
  dataset: '@dataset'
  batch_size: '@batch_size'
  num_workers: '@num_workers'

# should be replaced with other inferer types if training process is different for your network
inferer:
  _target_: SimpleInferer

# transform to apply to data from network to be suitable for loss function and validation
postprocessing:
  _target_: Compose
  transforms:
  - _target_: Activationsd
    keys: '@pred'
    softmax: true
  - _target_: AsDiscreted
    keys: '@pred'
    argmax: true
  - _target_: SaveImaged
    keys: '@pred'
    meta_keys: pred_meta_dict
    data_root_dir: '@dataset_dir'
    output_dir: '@output_dir'
    dtype: $None
    output_dtype: $None
    output_postfix: ''
    resample: false
    separate_folder: true

# inference handlers to load checkpoint, gather statistics
handlers:
- _target_: CheckpointLoader
  _disabled_: $not os.path.exists(@ckpt_path)
  load_path: '@ckpt_path'
  load_dict:
    model: '@network'
- _target_: StatsHandler
  name: null  # use engine.logger as the Logger object to log to
  output_transform: '$lambda x: None'

# engine for running inference, ties together objects defined above and has metric definitions
evaluator:
  _target_: SupervisedEvaluator
  device: '@device'
  val_data_loader: '@dataloader'
  network: '@network'
  inferer: '@inferer'
  postprocessing: '@postprocessing'
  val_handlers: '@handlers'

run:
- $@evaluator.run()