Translating Raw Data into a deep-learning trainable format

[CCranney]

Hi,

Thank you so much for putting together this curated list of nanopore software. I'm investigating ways to convert raw nanopore data (.fast5 files) and associated sequences (.fasta files) into a format that can be fed into a deep learning algorithm. SACall, for example, requires input to be numpy-specific signal and label files - but fails to demonstrate how it made the conversion from this dataset to numpy arrays.

Do you know if those kinds of software packages exist and how that would be best accomplished?