I cloned the newest version, when I run the train script I get this warning:
WARNING: non-finite loss, ending training tensor([nan, nan, nan, nan], device='cuda:0')

Caching labels (20 found, 0 missing, 0 empty, 0 duplicate, for 20 images): 100%|█████████████████████| 20/20 [00:00<00:00, 465.89it/s]
Caching labels (10 found, 0 missing, 0 empty, 0 duplicate, for 10 images): 100%|████████████

Recreate steps:

• upload a file that is not a properly encoded JPG/PNG image to a COS bucket
• In cloud annotations select the bucket and annotation type "localization"

Result:

• A generic file icon is displayed in the carousel but no error is displayed.
• One can draw a bounding box in the editor

Expected result:

• The user is notified that one or more objects in the bucket are not of

Current Monitoring/Alerts section pertains to Gravity 5.5 and earlier

This task covers extending the training section to involve Gravity 6.X+

Also a new section should be added to the Monitoring/Alerts discuss setup of Prometheus to scrape metrics from a custom pod/deployment. Some customers have been looking at running a simple pod and would like to see the metrics scraped from it to the de

In the Kubernetes admin course, we mention CoreOS Tectonic, but it may soon be folded into OpenShift. The slides should be updated to clarify (and, while we're there, add a few other options).

This is in file slides/k8s/setup-selfhosted.md.

[Discussion in #446](https://github.com/jpetazzo/container.training/pull/446/files/3f40cc25a2f31e54a02d64e683deee1be9837e86#diff-2ffe1c2ff95f331d6a12f9ea

SUMMARY

It would be nice to include workbench documentation fo

Please forgive the provocative and overly broad title. ;)

Opening the very first examples in IDEA, we immediately get suggestions. For instance:

And you see the grey wrigglies indicating that IDEA thinks we should use isNotEmpty().

If this is not a conscious decision,