toml file and model training

[hd2326]

Greetings @iiSeymour !

I am trying to train my own bonito models, and it would be really appreciated if you could provide some insights into the following questions.

1. What exactly do "--save-ctc" do? Specifically, what does the yielded "chunks.npy" contain? Maybe normalized event signal levels?
2. What do the parameters in the toml file mean, specifically: [input] features [qscore] bias [qscore] scale [encoder] stride [encoder] winlen [encoder] scale [encoder] blank_score [global_norm] state_len
3. Can I use different toml config files (different deep learning architectures) when running "bonito basecaller --save-ctc" and "bonito train"?

Thank you very much in advance for your help!