Regarding the dimension of query and key about axial-deeplab HOT 4 CLOSED

No, the "half" here is just an architectural hyper-parameter, which controls efficiency-accuracy trade-off. Empirically, we found "half" is a good trade-off in our cases, but "one" or "quarter" might also work well.

from axial-deeplab.

@csrhddlam Thanks. I guess "one" is supposed to give the best accuracy? Do you have an estimate of how better it is compared to "half"?

from axial-deeplab.

Yes, more channels usually leads to better accuracy, but we did not study much about it. Personally, I won't expect much improvement by switching from "half" to "one".

from axial-deeplab.

Thanks for the answer.

from axial-deeplab.