inference model onnx with dynamic input sizes in ml.net

[hoangdung]

I want to infer model.onnx with animation size using ml.net or any other library on .net.
I used ml.net and it seems to require declaring a fixed input size and the image must be resized to that size before being added to the model.

 

[Skydiver]

I think that you need to step back a little bit and provide more background about what you are trying to do. As I understand things, you want to create a learning model. A learning model of what?

You make mention of fixed input size and images. Why does it matter if the image sizes need to be resized to that fixed size? That a common enough graphics operation that shouldn't be a hindrance.