Open flassTer opened 5 years ago
Hello, I was wondering, how would you be able to compute the average confidence level of the transcripted text output by the acoustic models? For example where do you calculate the soft-max of each character that makes up the predicted word?
Hello, I was wondering, how would you be able to compute the average confidence level of the transcripted text output by the acoustic models? For example where do you calculate the soft-max of each character that makes up the predicted word?