Finding callback __init__ error in learn.fit()

Hello, I have been trying to make a GAN run lately but I keep getting this error message:


And here is my GanLearner definition
ganlearner

I will also add that I didn’t use a custom DataBunch in for map_data, I created my databunch like this.
data

I don’t know if that might have had an effect in the error message I am getting.
My comprehension of this error is that it has to do with the initialization of a callback dealing with the Learn class (I might be wrong).
Your help is very much welcomed.

The transforms returned by get_transforms are to be applied on the dataset, not the dataloader level, this may be one of the reasons you get a bug.

My task actually to make an image to image translatior, to translate satellite images to their google maps images equivalent. I just noticed the ImageImageList itemlist in the library. I tried to use it and was able to create my databunch but the same error persists.

Update: for some reason data.one_batch() outputs a tensor but show_batch() actually shows images although they sometime appear with themselves as target.

It’s hard to say where the problem lies without seeing the whole code and being able to reproduce. Also, make sure you have the latest version of the library as it may be a bug that has been fixed.

Thank you for your reply Sylvain. Would it be possible for you to take a look at my notebook if it is not too much asking?