[STC] STC loss ascends while training #21
Comments
Hi, As a sanity check, you can output Just to make sure have you made sure |
Hello. @vineelpratap Now I'm trying to train with ctc(gtn_application). Thank you for your advice.
|
When I changed thalf 16000 -> 8000, and model to output word level directly, STC loss increases slightly after 8k step. More details about by model is,
|
Hello, I'm training ASR model with STC Loss and letter-to-word encoder like below.
But when I progress training, STC Loss ascended and became 'Inf' after 12000 step.
Is there any miss in my implementation?
Any help would be appreciated.
Thank you.
training args:
The text was updated successfully, but these errors were encountered: