First batch can't access any element of self.train_ds

Part 1 (2019)
1

I am working with NIH chest X-ray data set and trying to use the methods outlined in the Coco example. I have built a img2bbox dictionary with the following format

{‘00013118_008.png’: [[547.0192168, 225.0847458, 626.20565748, 311.86440682],
[‘Atelectasis’]],
‘00014716_007.png’: [[131.5434984, 686.1016949, 445.0350238, 871.5932203],
[‘Atelectasis’]]}

I am then trying to build an ObjectItemList

data=(ObjectItemList.from_df(df=df_bb,path=path,folder=‘images’,cols=0)
.split_by_rand_pct(valid_pct=.2,seed=0)
.label_from_func(get_y_func)
.databunch(bs=16, collate_fn=bb_pad_collate)
)

I keep getting the following warning which I believe is linked to the creation of the databunch.

C:\Users\aszei\Anaconda3\envs\New FastAI\lib\site-packages\fastai\basic_data.py:262: UserWarning: There seems to be something wrong with your dataset, for example, in the first batch can’t access any element of self.train_ds.
Tried: 728,128,444,647,245…
warn(warn_msg)

Any help with this will be greatly appreciated. Thanks.