oktoberfest.pr.DLomix

class oktoberfest.pr.DLomix(model_type, model_path, output_path, batch_size, download=False)

A class for interacting with DLomix models locally for inference.

Parameters:

• model_type (str)

• model_path (Path)

• output_path (Path)

• batch_size (int)

• download (bool)

__init__(model_type, model_path, output_path, batch_size, download=False)

Initialize a DLomix predictor from name or path of pre-loaded weights.

Parameters:

• model_type (str) – Type of model (intensity or irt)

• model_path (Path) – Path of model file

• output_path (Path) – Directory to save processed data for predictor to for reuse

• batch_size (int) – Batch size to use for inference

• download (bool) – Whether to download a new baseline model from GitHub to the model_path

Raises:

NotImplementedError – if a retention time predictor is requested

Methods

predict(data, dataset_name[, keep_dataset])	Create predictions for dataset using Keras model.
predict_xl(data, dataset_name[, keep_dataset])	Perform inference on the xl data using the DLOmix model.