extracting values from inference results

[Finfalter]

Could you sketch the way how to further process the inference (extract the result) from a model? To explain what I mean: I took the model of tensorflow's hello_world example and got some results, for example

[[0.8327713]], shape=[1, 1], strides=[1, 1], layout=CFcf (0xf), dynamic ndim=2

The underlying model is supposed to have learned sine(). The input in this case was 0.94529176, so the output is ok with inferred 0.8327713. The underlying variable of the displayed value is of type

ndarray::ArrayBase<ndarray::ViewRepr<&f32>, ndarray::dimension::dim::Dim<ndarray::dimension::dynindeximpl::IxDynImpl>>

I can display the entire output as shown above but in order to further process it, I have to extract it as a f32.
How would I do that?

[kali]

Hey!

The array_view methods, like you tried, are a bridge to ndarray ArrayBase. You can access values by indexing (using []) like shown here: https://docs.rs/ndarray/0.15.4/ndarray/struct.ArrayBase.html#indexing-and-dimension.

[Finfalter]

It works, thanks a lot!