how to run inference in a single image

[legan78]

Note that the method predicts sequences from a sequences of skeletons taken as inputs. You will need more than 1 image (50) from which you need to extract 3d skeletons.