whylogs
whylogs copied to clipboard
Track NaN and Infinity separately from `null`
Description
When a system produces NaN or infinity
, that often indicates bad data or bugs.
Suggestions
Track these under n
(count) metric. Maybe under a namespace of counts?
Related
- Tensorflow tracks the number of
NaN
s: https://github.com/tensorflow/metadata/blob/770c81d26e1bd09faecb2b7940466ad9ddde5483/tensorflow_metadata/proto/v0/statistics.proto#L501
This issue is stale. Remove stale label or it will be closed tomorrow.
still relevant
This issue is stale. Remove stale label or it will be closed tomorrow.