Here is other article about ML.NET concepts.
Today i write about MLContext.
MLContext
ML.NET has a lot of component and for the developer navigate it in the middle, could be complicated.
MLContext it’s the object that seems Caronte. It drive our soul in all ML.NET’s objects.
In fact It give us the visibility of all components and operations.
Let see:
MLContext.Data that contains the methods about the creation of IDataView interface:
– Load/Save data in file.
– Save memory data in cache.
– Filter data.
MLContext.Model thath contains the methods about the models or the transformer:
– Save and load
– Create prediction function
MLContext.Transforms that contains estimatorsfor the non-prevision operations.
– Categorical
– Text for natural language processing
– Conversion
– Vectorial Prediction
– Normalization
– Column operation (rename, delete, clone, concatenate, etc.)
MLContext.BinaryClassification, MulticlassClassification, Ranking and others, that are catalogues of learning algorithm for specific machine learning tasks:
– TrainTestSplit e CrossValidate
– Trainers that containing various specific trainers for activities.
MLContext.Log that’s a stream of text messages on long-running processes ML.NET is running.
References: