rca22 / lightgbm.net Goto Github PK

I've created a .NET Web API project and written an API handler that successfully calls Booster.PredictForMat(...).

I've taken great care to ensure that each call to the API gets access to a Booster instance that is guaranteed to only be used with that particular API client, i.e. using object pools to have short-term leases on instances of Booster.

Do I actually need to ensure that each client has its own Booster instance, or can I share a single Booster instance across all clients?

Having one instance would vastly improve the runtime performance of my API, but I need to be sure it's a safe thing to do.

Hello,

I have trained a model using LightGBM CLI, and I am interested in using this model from a C# program. Unfortunately, I haven't been able to find an example of how to do this. Could you please point me in the right direction or provide some guidance?

I assume the code for using the model in C# should resemble something like this:

Booster booster = new Booster(@"path to my model from LightGBM CLI");
double[] desiredOutput = null;
var inputData = LoadInputs(@"path to my testing data", out desiredOutput);
var prediction = booster.Predict(inputData);
Console.WriteLine("Prediction: " + prediction);
Console.ReadLine();

Your assistance is greatly appreciated.

Thanks,
Wil

After loading previously saved model a call to ContinueTraining results in exception: No existing booster to train.

Is this a bug or a feature?

Thanks

Hi,

Would you please include GPU Binaries in the nuget package then Microsoft ML.net team would be able to add GPU support to their wrapper.

Thanks