gtnet.predict module

gtnet.predict.predict(argv=None)

Get network predictions for each sequence in Fasta file

Parameters:

argv (Namespace, default=sys.argv) – The command-line arguments to use for running this command

gtnet.predict.run_torchscript_inference(fastas, model, conf_models, window, step, vocab, seqs=False, n_chunks=10000, device=device(type='cpu'), logger=None)

Run Torchscript inference

Parameters:

• fastas (str) – The path to the Fasta file with sequences to do inference on

• model (RecursiveScriptModule) – The Torchscript model to run inference with

• conf_models (dict) – A dictionary with the confidence model for each taxonomic level. Each model should be a RecursiveScriptModule. The expected keys in this dict are ‘domain’, ‘phylum’, ‘class’, ‘order’, ‘family’, ‘genus’ and ‘species’.

• window (int) – The length of the sliding window to use for doing inference

• step (int) – The length of the step of the sliding window to use for doing inference

• vocab (str) – The vocabulary used for training model

• n_chunks (int, default=10000) – The length of the step of the sliding window to use for doing inference

• device (device, default=torch.device('cpu')) – The Pytorch device to run inference on

• logger (Logger) – The Python logger to use when running inference