[Docs] Batched inference in existing models?

[wouterwln]

📚 The doc issue

I can not find in the documentation if it is possible to do inference on batches of images (or an np.ndarray of rank 4) using the MMPoseInferencer. I can imagine it is beneficial for users to run inference in batches because of the obvious performance gains. But it is not clear how and if this is possible

Suggest a potential alternative/fix

A clear tutorial on how to do batched inference.