num_targets in params file vs --ti option in basenji_test.py

[FaizRizvi]

Hi,
I am trying to understand how basenji works. We are trying to train a model on a small dataset and produce bw files of prediction for all five tracks. I do not understand the role of num_targets and its relation with --ti. For example. We had a dataset that included 2 files to train a model and we wanted to make predictions on all five tracks, but it failed to do so. Could you possibly explain the roles of num_targets in the params file and the --ti option in basenji_test.py. Many Thanks!
Best,
Faiz

[davek44]

This changed a bit between the tf1 and tf2 versions, and I know you're moving to tf2 from the other thread, so I'll hold off to give specific advice. Generally, num_targets in the parameters file is meant to help guide the model construction during training. I wrote the --ti option in basenji_test.py to allow me to write bigwigs for just a few of the many targets. So you can ignore it if you want to write all targets.