Is there a way to purge a learner from GPU in order to load a new learner and not stacking models on gpu?
I think it’s something like
learn.to("cpu")
if learn is your learner.
Or simply
del learn
if you don’t need it anymore. Google it though for precise syntax.
1 Like