pythonpytorchfloating-pointtype-conversiontensor

How to convert torch tensor to float?


I am using flask to do inference and I am getting this result. Is their any way to convert this tensor into float because I want to use this result to display in a react app:

{
  result: {
    predictions: "tensor([[-3.4333]], grad_fn=<AddmmBackward>)"
  }
}

Solution

  • From Torch.Tensor.item docs:

    x = torch.tensor([1.0])
    
    print((x.item())
    

    output:

    1.0
    

    type check:

    print(type(x.item())
    

    output:

    float