fix: enhance inference adapter with raw result copying

[A-Artemis]

What does this PR do?

Model.infer_async passed get_raw_result (numpy views into the InferRequest output tensors) instead of copy_raw_result. Detection/inst-seg postprocess mutate those buffers in place, np.round(boxes, out=boxes), boxes -= …, boxes *= …, labels += shift, from inside async callbacks, with up to 16 requests being recycled. Classification/semantic segmentation only read.

Fixes this by exposing copy_raw_result() method.

Before submitting

• This PR fixes a typo or improves the docs (you can dismiss the other checks if that's the case).
• Did you make sure to update the documentation with your changes?
• Did you write any new necessary tests?