调用IDataView时不能确定成员特性的MakePredictionFunction类型

Question

我试图使用新的Microsoft.ML 0.6.0来制作预测函数。

当我叫"model.AsDynamic.MakePredictionFunction“时，我收到

"System.ArgumentOutOfRangeException：‘无法确定成员特性的IDataView类型’“。

代码：

using System;
using Microsoft.ML;
using Microsoft.ML.Runtime.Api;
using Microsoft.ML.Runtime.Data;
using Microsoft.ML.Trainers;
using Microsoft.ML.StaticPipe;

namespace MachineLearning
{
    class MLTest
    {
        public void Run()
        {
            var env = new LocalEnvironment();
            var reader = TextLoader.CreateReader(env, ctx => (label: ctx.LoadBool(0), features: ctx.LoadFloat(1, 3)));
            var traindata = reader.Read(new MultiFileSource("train.txt"));
            var bctx = new BinaryClassificationContext(env);
            var est = reader.MakeNewEstimator()
                .Append(x => (x.label, prediction: bctx.Trainers.Sdca(x.label, x.features.Normalize())));
            var model = est.Fit(traindata);

            //FAILS: System.ArgumentOutOfRangeException: 'Could not determine an IDataView type for member features'
            var predictionFunct = model.AsDynamic.MakePredictionFunction<Issue, Prediction>(env);

        }

        public class Issue
        {
            public float label;
            public Vector<float> features; //what is wrong?
        }

        public class Prediction
        {
            [ColumnName("prediction.predictedLabel")]
            public bool PredictionLabel;

            [ColumnName("prediction.probability")]
            public float Probability;

            [ColumnName("prediction.score")]
            public float Score;
        }
    }
}

文件train.txt包含：

1   0   0   0
1   0   1   0
1   0   0   1
1   0   1   1
0   1   1   1
0   1   0   1
0   1   1   0
0   1   0   0

看起来错误在课堂上是“问题”，但到底是什么错了呢？谢谢

Answer 1

你在尝试模式理解以生成数据视图和读取。通过使用启动数组，您可以使用列类型来设置数组的大小。

(这是.10 of ML.NET的版本)

public class Issue
{
        public float label;
        public float[] features; //change this
}

使用SchemaDefinition进行运行时类型映射提示

var inputSchemaDefinition = SchemaDefinition.Create(typeof(Issue), SchemaDefinition.Direction.Both);
inputSchemaDefinition["features"].ColumnType = new VectorType(NumberType.R4, 4);

然后创建引擎：

var predictionEngine = model.CreatePredictionEngine<InputSchema, Prediction>(model as IHostEnvironment, inputSchemaDefinition, outputSchemaDefinition);