Symbolic SGD (binary)ΒΆ

The documentation is generated based on the sources available at dotnet/machinelearning and released under MIT License.

Type: binaryclassifiertrainer Aliases: SymbolicSGD, SymSGD Namespace: Microsoft.ML.Trainers.SymSgd Assembly: Microsoft.ML.HalLearners.dll Microsoft Documentation: Symbolic SGD (binary)

Description

Parameters

Name Short name Default Description
l2Regularization l2 0 L2 regularization
learningRate lr   Learning rate
memorySize accelMemBudget 1024 The acceleration memory budget in MB
numberOfIterations iter 50 Number of passes over the data.
numberOfThreads nt   Degree of lock-free parallelism. Determinism not guaranteed. Multi-threading is not supported currently.
positiveInstanceWeight piw 1 Apply weight to the positive class, for imbalanced data
shuffle shuf True Shuffle data?
tolerance tol 0.0001 Tolerance for difference in average loss in consecutive passes.
updateFrequency freq   The number of iterations each thread learns a local model until combining it with the global model. Low value means more updated global model and high value means less cache traffic.