Hello.  Running locally, I would like to run my algorithm train an ML.Net model, save the trained .mdl, then run the algorithm again with the loaded .mdl.  ML.Net has some functionality to save the trained object to a .mdl file.  I see QuantConnect has the ObjectStore which can save/load strings and some serialized types.  But it seems that the trained model object by ML.Net is not serializable, and the ObjectStore does not support saving of files.  When just trying to save/load the model using Ml.Net's MlContext.Model.Save, the file is not found on the consecutive algorithm run using MlContext.Model.Load method.  Is this not possible or am I just missing something?

Author