Dataset Viewer issue

#8
by lil-mdellabitta - opened

The dataset viewer is not working.

Error details:

Error code:   JobManagerCrashedError

Is there a way to get more details on what it crashed doing? Is this a result of the dataset being too large and things timing out?

cc @albertvillanova @lhoestq @severo .

Usually it's because the job ran out of memory. I launched a refresh of the job, it should be processed soon

Unfortunately it failed again, we'll investigate that. Sorry for the inconvenience

Thanks for paying attention to this!

i ran out of memory myself a couple of times.
the work-around for me was to filter remotely:
from datasets import load_dataset
dataset = load_dataset('harvard-lil/cold-cases', split='train', streaming=True)
train_data = dataset['train']
filtered_dataset = dataset.filter(lambda example: example['court_jurisdiction'] == 'Indiana, IN')

Sign up or log in to comment