add MLflow logging example #892
Conversation
Addressing MLflow code example to integrate easily Nixtla#825 Nixtla#825
|
Check out this pull request on See visual diffs & provide feedback on Jupyter Notebooks. Powered by ReviewNB |
|
|
|
Hello @jmoralez , Cheers! |
Transferring change captured via Fixing bug for issue Nixtla#825 (PR now closed).
To address the first 4 changes suggested via PR 892.
Adding code to load the saved model and make predictions using MLFlow.
This change makes clear that the latest run ID model is being called. Otherwise, an specific model run_id can be loaded.
As discussed, removed lines of code to publish this examples as an MLFLow integration to train, track and monitor neuralforecast models.
Refining notebook code as suggested.
Names converted to lower case, thank you. Co-authored-by: José Morales <jmoralz92@gmail.com>
|
Hello @jmoralez , I just included all suggestions except for a suggestion to remove the #Registering model code section. I explained why on my comment to resolve that suggestion. Thank you for all these suggestion! :) |
Removing signature step as suggested.
Changing dataframe to concat.
Removed redundant step to simplify code as suggested.
|
Hey @cargecla1, I just pushed some changes. Please let me know if they look good to you, if they do we can merge this. |
Hello @jmoralez , These changes look great, thank you for these additions! Thank you very much for your support! |
|
Thanks a lot for the contribution! |
Addressing MLflow code example to integrate easily
Fixes #825