I am working on my custom model. My custom model takes integer tensor and double tensor and returns double tensor.
However, in Barracuda, it seems like only the float type can be fed into the model. For example, when we create a tensor, it only takes a float array. public Tensor(TensorShape s, float[] srcData, string n = "")
Do you know how to feed integer tensors and double tensors into the model?
I also convert those int and double arrays into float arrays and feed them to the model. However, the output is different from what it is supposed to be.
I've also added my model here in case you want to investigate.
model
Hi,
I am working on my custom model. My custom model takes integer tensor and double tensor and returns double tensor. However, in Barracuda, it seems like only the float type can be fed into the model. For example, when we create a tensor, it only takes a float array. public Tensor(TensorShape s, float[] srcData, string n = "") Do you know how to feed integer tensors and double tensors into the model?