You signed in with another tab or window. Reload to refresh your session.You signed out in another tab or window. Reload to refresh your session.You switched accounts on another tab or window. Reload to refresh your session.Dismiss alert
Is your feature request related to a problem? Please describe.
For models with sparse features, passing a default value to the model server ballons the packet size by as much as 80%. Maybe we can define the default value of the features in model config?
Describe the solution you'd like
Define default value in model config. If the feature is not passed, use default value.
Describe alternatives you've considered
Use ensemble model. First model does the msg parsing and creating a full feature list, then pass it to the main model. This is however still expensive and adds to deserialization and serialization.
The text was updated successfully, but these errors were encountered:
Is your feature request related to a problem? Please describe.
For models with sparse features, passing a default value to the model server ballons the packet size by as much as 80%. Maybe we can define the default value of the features in model config?
Describe the solution you'd like
Define default value in model config. If the feature is not passed, use default value.
Describe alternatives you've considered
Use ensemble model. First model does the msg parsing and creating a full feature list, then pass it to the main model. This is however still expensive and adds to deserialization and serialization.
The text was updated successfully, but these errors were encountered: