Jupyter kernel died everytime I run "show_batch"

Part 1 (2020)
1

hi everyone, when I am running the below code from 02_production.ipynb:

dls = bears.dataloaders(path)  # this line can execute normally

dls.valid.show_batch(max_n=4, nrows=1)  # this line make jupyter kernel died

my jupyter kernel always died. if any one have an idea what this issue is?

2

If you running locally it might be your system running out of memory.

3

Hi! I get the same when trying to load UNet keras model on remote Jupyter notebook. The system, I guess, was fine. What else could it be?

4

not really, when I run dls.valid.show_batch，the code haven’t train any mode yet, just build some data structure. probably some bug on jupyter, cause I try the same code in Pycharm，finish with code 0, seems nothing error happen in pycharm.

5

I meet the issue on local jupyter server, not only show_batch method, but also show results method will cause jupyter kernel crashed. It is really weird.

6

Seems to be having the same issue with data.show_batch() when I run it locally. But does work on google vertex ai workbench.