Result Gap between CTC and RNN-T Multi-Task Training #973
-
Between the non-streaming results of vanilla and CTC, it is clear that the model with CTC performed better for test-other. However, is there any idea or intuition about why the models trained with a secondary CTC loss performed slightly worse for test-clean?
|
Beta Was this translation helpful? Give feedback.
Answered by
yfyeung
Apr 3, 2023
Replies: 1 comment 1 reply
-
These gaps are so slight, I prefer to believe they are caused by noise. |
Beta Was this translation helpful? Give feedback.
1 reply
Answer selected by
marcoyang1998
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
These gaps are so slight, I prefer to believe they are caused by noise.