Have you tried reinstalling/upgrading fastai and using the original library function? I just tried it in the ‘02_production’ notebook from the fastai book on Colab and it worked for me. It looks like they pushed a fix to the deployed version.
You are right! I did not find time to check the original function, which is now working again.
Hello, not sure if this is related to the same issue but when using
plot_top_losses() form an interp object built on Datablock implementation (not the midlevel API) of the siamese network tutorial, i get the following error :
fastcore/dispatch.py:191, in retain_type(new, old, typ, as_copy)
189 typ = old if isinstance(old,type) else type(old)
190 # Do nothing the new type is already an instance of requested type (i.e. same type)
→ 191 if typ==NoneType or isinstance(new, typ): return new
192 return retain_meta(old, cast(new, typ), as_copy=as_copy)
==TypeError: isinstance() arg 2 must be a type, a tuple of types, or a union==
is there a way for these functions to return the handles to the figures they plot? I would like to automatically save my figures without having to go manually through them and right-click+save them all.
I think there used to be an input argument called “return_fig” that you could set to “True” to get the handles out of the funcitons, but it does not seem to recognise it anymore.
I am using Colab, python 3.7.