[tabular] mention persist and compile in deploy tutorial #2717
Conversation
|
Job PR-2717-a175686 is done. |
| ```{.python .input} | ||
| predictor_clone_opt.compile_models() | ||
| ``` |
There was a problem hiding this comment.
This will probably cause the tutorial to return False in the later check that the predictions are equivalent. Instead, have the equivalency checks prior to the compile_models step, then predict again and mention that predictions might not be exactly the same, but should be very close.
There was a problem hiding this comment.
Moved to bottom part of the tutorial and mentioned the potential difference.
|
Job PR-2717-0298008 is done. |
|
Job PR-2717-64c6a22 is done. |
Issue #, if available:
Description of changes:
This PR updates the tabular model deployment tutorial with some context on
persist_modelsandcompile_modelsfor efficient real-time prediction.By submitting this pull request, I confirm that you can use, modify, copy, and redistribute this contribution, under the terms of your choice.