deepdisc.inference.predictors

Functions

return_predictor_transformer(cfg[, checkpoint])	This function returns a trained model and its config file.
get_predictions(dataset_dict, imreader, key_mapper, ...)	Returns indices for matched pairs of ground truth and detected objects in an image
get_predictions_new(dataset_dict, predictor)	Returns indices for matched pairs of ground truth and detected objects in an image

Module Contents

return_predictor_transformer(cfg, checkpoint=None)

This function returns a trained model and its config file.

Used for models with lazy config files. Also assumes a cascade roi head structure.

Parameters:

cfg (.py file) – a LazyConfig

Return type:

torch model

get_predictions(dataset_dict, imreader, key_mapper, predictor)

Returns indices for matched pairs of ground truth and detected objects in an image

Parameters:

• dataset_dict (dictionary) – The dictionary metadata for a single image

• imreader (ImageReader object) – An object derived from ImageReader base class to read the images.

• key_mapper (function) – The key_mapper should take a dataset_dict as input and return the key used by imreader

• predictor (AstroPredictor) – The predictor object used to make predictions on the test set

Returns:

• matched_gts (list(int)) – The indices of matched objects in the ground truth list

• matched_dts (list(int)) – The indices of matched objects in the detections list

• outputs (list(Intances)) – The list of detected object Instances

get_predictions_new(dataset_dict, predictor)

Returns indices for matched pairs of ground truth and detected objects in an image

Parameters:

• dataset_dict (dictionary) – The dictionary metadata for a single image

• predictor (AstroPredictor) – The predictor object used to make predictions on the test set

Returns:

outputs – The list of detected object Instances

Return type:

list(Intances)