Problem in running fastai based segmentation on Kaggle

Part 1 (2019)
1

Hi,
I am fairly new and after going through the course Part 1 of Deep Learning for Coders, I practiced on Kaggle. I was working on Steel Defect Detection Competition and got stuck in the lr_find step.

The error I am getting is
RuntimeError: Assertion `cur_target >= 0 && cur_target < n_classes' failed.
You can view my kernal here.

I tried searching online, but could not find a solution that can get rid of this problem. Can you guys have a look and suggest something that I may try?

Thanks!

2

check what data.c shows and make sure mask values are all between 0 and that value.