Repository navigation
Replies: 1 comment
|
The largest number that fits your GPU is the best. Something like 16 or 32 is a good start. |
0 replies
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
Uh oh!
There was an error while loading. Please reload this page.
Hello,
I'd like to ask if there are some heuristics to set
validation_batch_size?Normally validation set size is way smaller than training set. So having the same value for
batch_sizemight be sensible.Could you give me some advice on this?
Thank you in advance!
All reactions