business-science / modeltime

Modeltime unlocks time series forecast models and machine learning in one framework
https://business-science.github.io/modeltime/
Other
522 stars 79 forks source link

Modeltime in Production #134

Open AlbertoAlmuinha opened 2 years ago

AlbertoAlmuinha commented 2 years ago

Hi @mdancho84 ,

I've been thinking that it would be useful for Modeltime to incorporate some function like "deploy_model" to deploy the models directly to some of the main clouds (AWS, Azure...). Some packages like Pycaret are doing it and I think this functionality could be very interesting for Modeltime to be used by more people.

We would have to think about how to do it, but I think it could be worthwhile. Let me know your thoughts about this :)

spsanderson commented 2 years ago

I think this would be an amazing idea, I also think maybe being able to export the model to sql would be good

Sent from my iPhone www.spsanderson.com

On Oct 2, 2021, at 8:38 PM, Alberto González Almuiña @.***> wrote:

 Hi @mdancho84 ,

I've been thinking that it would be useful for Modeltime to incorporate some function like "deploy_model" to deploy the models directly to some of the main clouds (AWS, Azure...). Some packages like Pycaret are doing it and I think this functionality could be very interesting for Modeltime to be used by more people.

We would have to think about how to do it, but I think it could be worthwhile. Let me know your thoughts about this :)

— You are receiving this because you are subscribed to this thread. Reply to this email directly, view it on GitHub, or unsubscribe.

AlbertoAlmuinha commented 2 years ago

@spsanderson I would first focus on being able to deploy the model in some cloud. About SQL there is already some Tidymodels package but each model would need a special treatment, so it would be more complex. Take a look: https://github.com/tidymodels/tidypredict

spsanderson commented 2 years ago

@AlbertoAlmuinha thanks for the link

mdancho84 commented 2 years ago

This is interesting!