I've found this library very useful for things aside from just model training. Really it can be used for any python script that is being deployed to cloud infrastructure. I've found it particularly useful for deploying containerised applications which gives me peace of mind that I'll just get a slack message if something has gone wrong. Whether or not it's ML related.
So on that note, I wonder if you would consider broadening the scope of the library to also accommodate other types of script. As far as I can tell, this just means modifying some of the messages to be more generic instead of referring specifically to "training". Potentially there could even be different categories of deployment corresponding to different vocabulary for the messages.
Is this something you would consider? Or maybe there is another library to use in these non-ML circumstances that you're aware of?
amin-nejad
commented
2 years ago
I've found this library very useful for things aside from just model training. Really it can be used for any python script that is being deployed to cloud infrastructure. I've found it particularly useful for deploying containerised applications which gives me peace of mind that I'll just get a slack message if something has gone wrong. Whether or not it's ML related.
So on that note, I wonder if you would consider broadening the scope of the library to also accommodate other types of script. As far as I can tell, this just means modifying some of the messages to be more generic instead of referring specifically to "training". Potentially there could even be different categories of deployment corresponding to different vocabulary for the messages.
Is this something you would consider? Or maybe there is another library to use in these non-ML circumstances that you're aware of?