Different behavior when calling Fit() on a transformer chain and on an IDataLoaderEstimator

[yaeldekel]

The following code runs without errors:

            var loader = ml.Data.CreateTextLoader<InputData>(hasHeader: true, dataSample: file);
            var data = loader.Load(file);

            // Pipeline.
            var pipeline = ml.BinaryClassification.Trainers.GeneralizedAdditiveModels();

            // Train.
            var model = pipeline.Fit(data);

However, the following code fails with schema mismatch exception:

            var loader = ml.Data.CreateTextLoader<InputData>(hasHeader: true, dataSample: file);

            // Define the same pipeline starting with the loader.
            var pipeline = loader.Append(ml.BinaryClassification.Trainers.GeneralizedAdditiveModels());

           // Train
            var model = pipeline.Fit(file);

This may also be related to issue #1969 .

[yaeldekel]

By the way, this is the definition of InputData:

        private class InputData
        {
            [LoadColumn(0)]
            public float Label { get; set; }
            [LoadColumn(9, 14)]
            [VectorType(6)]
            public float[] Features { get; set; }
        }

and this is the dataset: https://github.com/dotnet/machinelearning/blob/master/test/data/adult.tiny.with-schema.txt.

[Ivanidzo4ka]

Does it complain on expected bool got float?
If it does I have PR which forces check input schema for first case. #2804

[yaeldekel]

I think your PR may fix this issue as well as #1969, will try it again once it is merged.