Datablock api summary method fails for me when I try to do image classification using a df

[sophia]

Hi, so I wanted to do image classification from a dataframe using the datablock api of fastai2.
(All image files (3500 for train and valid together) are in the same folder).

I have been using for image classification fastai1 before, so I don’t have much experience with fastai2 yet.

I am using a df with the column names= “name”, “label”, “split0”, “split1”, etc. (splitx is like “is_valid” with being True or False for every image).

So I got the datablock working, and I can alse use show_batch, but using the summary function does not work on the datablock I created.

So I was wondering what is going wrong here? - Does “summary” work the same for a datablock when you use a dataframe? What do I have to do differently? Or is there just something wrong?

Maybe someone can help

[sophia]

Ok so after looking it for some time, I just wasn’t very clever. There is supposed to be given the source or in this case the respective df as input in the summary brackets.

So in my case it is supposed to be datablock1.summary(df1).

Sorry for post (was lost for some time).