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
One of Conv parameters is now called filter (after #873 was merged.
In my opinion, size better described what Flux expects to be supplied by the user: the size of the kernel. Now it reads filter, which, to me, sounds like Flux expects the user to supply the actual kernel here, not its size.
If others think it's fine, then I'm ok with that too.
The text was updated successfully, but these errors were encountered:
To address your question specifically, Keras uses kernel_size which seems really clear to me. size would probably be preferred to filter if describing the size of convolutional filters
Ah I see now, looking at the example I think it's reasonably clear what the argument is, but without the example could be somewhat confusing. Personal opinion, obviously.
One of
Conv
parameters is now calledfilter
(after #873 was merged.In my opinion,
size
better described what Flux expects to be supplied by the user: the size of the kernel. Now it readsfilter
, which, to me, sounds like Flux expects the user to supply the actual kernel here, not its size.If others think it's fine, then I'm ok with that too.
The text was updated successfully, but these errors were encountered: