You signed in with another tab or window. Reload to refresh your session.You signed out in another tab or window. Reload to refresh your session.You switched accounts on another tab or window. Reload to refresh your session.Dismiss alert
Hello, I used RNNT training on the Chinese speech recognition library of more than 300 hours (the encoder did pretrain, but the decoder is a random initialization parameter). After training dozens of epoch, the loss first quickly dropped from more than 1000 to 60. Then slowly dropped to more than 20, but the SER of inference has risen from 2 to 20. Is this normal? It seems that you mentioned this phenomenon elsewhere.
Thank you very much!
The text was updated successfully, but these errors were encountered:
Hello, I used RNNT training on the Chinese speech recognition library of more than 300 hours (the encoder did pretrain, but the decoder is a random initialization parameter). After training dozens of epoch, the loss first quickly dropped from more than 1000 to 60. Then slowly dropped to more than 20, but the SER of inference has risen from 2 to 20. Is this normal? It seems that you mentioned this phenomenon elsewhere.
Thank you very much!
The text was updated successfully, but these errors were encountered: