Possibilities for training on longer samples #4
Unanswered
SashaBurashnikova
asked this question in
Q&A
Replies: 1 comment 1 reply
-
Hi! If you mean training on longer sequences, there shouldn't really be any limit as long as they fit in memory (you might just want to tune the batch size and learning rate). On the other hand, if you mean running the pre-trained model on longer samples, I would indeed expect the quality to degrade. But either way, I haven't tried this, so I cannot speak from experience. |
Beta Was this translation helpful? Give feedback.
1 reply
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
-
Hello,
Your results sounds very nice, thank you for sharing them in git!) i'm just wondering is it possible to apply your model to samples with longer duration? If yes, what is the relevant time-limit for samples in case of keeping the current output quality?
Beta Was this translation helpful? Give feedback.
All reactions