Does it make sense to have a half precision mode for cnn?
The docs mention here
https://docs.fast.ai/callbacks.fp16.html
The use of the class, learner.to.fp_16, to double the speed.
How can this be applied to the class, cnn_learner which has no attribute ‘.to.fp_16’
dipam7
(Dipam)
2
Hey, I remember using it with a cnn_learner
. Link to notebook
aklbg
3
Learner(data, model, metrics=[accuracy]).to_fp16() not to.fp16()
I realize there is a Learner.to_fp16
. I am asking about a cnn_learner. These docs
https://docs.fast.ai/vision.learner.html
do not indicate that attribute. How could we modify cnn_learner to use half precision?
KarlH
(Karl)
5
cnn_learner
is just a function that returns a Learner
object.
1 Like