Simple way to get at raw probabilities/logits via python bindings?

DeepSpeech
#1

I’m able to run the default python interface and get the top-scoring character sequences no problem. I was just wondering what’s the best/most-direct way of getting at say the top-N results per-character?

#2

Currently, it’s not exposed, but there’s a bug already open to expose more informations, I’d suggest you chime in there :slight_smile:

1 Like
#3

Can I use pre-trained model with DeepSpeech.py? see if this helps

1 Like
#4

Yes I believe that’s what I’m looking for thank you.