I have a Trained Model with Keras and Tensorflow Backend (Keras 2.2.4 Tensorflow 1.13.1) and i want to use that Model in Visual Studio with ML.Net.
Therefore i converted my Model to ONNX with winmltools.convert_keras (I tired it with a Tensorflow 2.0 model but i got the No module named 'tensorflow.tools.graph_transforms' error).
Now i finaly managed to load the model with:
string outName = "dense_6";
string inName = "conv2d_9_input";
string imgFolder = "path\to\Testimage
string pathToMLFile = "path\to\.onnx"
var pipeline = mlContext.Transforms.LoadImages(outputColumnName: outName, imageFolder: imgFolder, inputColumnName: inName)
.Append(mlContext.Transforms.ResizeImages(outName, 299, 299, inName))
.Append(mlContext.Transforms.ExtractPixels(outputColumnName: outName, interleavePixelColors: true, offsetImage: 117))
.Append(mlContext.Transforms.ApplyOnnxModel(outputColumnName: outName, inputColumnName: inName, modelFile: pathToMLFile, fallbackToCpu: true));
But now i need a IDataView to "Fit" the model (from my understanding it is needed to Initialize it?) Therefore i load an empty IDataView with:
var data = mlContext.Data.LoadFromEnumerable(new List<ImageNetData>());
with ImageNetData being
public class ImageNetData
{
[ColumnName("conv2d_9_input")]
[ImageType(299, 299)]
public Bitmap Image { get; set; }
[ColumnName("dense_6")]
public string Label;
}
now i get the error:
Schema mismatch for input column 'conv2d_9_input': expected String, got Image<299, 299>
Parametername: inputSchema
My model with Netron:
Why does it want a String? And if i remove the [ImageType(299,299)] and change the Bitmap to string i get:
Schema mismatch for input column 'conv2d_9_input': expected Image, got String
I hope my Problem is understandable.
Update: After the hint from Gopal Vashishtha i changed the input/output names:
var pipeline = mlContext.Transforms.LoadImages(outputColumnName: "image_object", imageFolder: imgFolder, inputColumnName: inName)
.Append(mlContext.Transforms.ResizeImages(outputColumnName: "image_object_resized", imageWidth: 299, imageHeight: 299, inputColumnName: "image_object"))
.Append(mlContext.Transforms.ExtractPixels(outputColumnName: "input", inputColumnName: "image_object_resized", interleavePixelColors: true, offsetImage: 117, scaleImage: 1 / 255f))
.Append(mlContext.Transforms.ApplyOnnxModel(outputColumnName: outName, inputColumnName: "input", modelFile: pathToMLFile + "\\Converted.onnx", fallbackToCpu: true))
But i still have the same Error. Might there be a Problem with the way i trained the Model? I used a numpy-array as image and 0/1/2 for the corresponding Classes.
