[doc] update inference related docs from `mp_size` to `tensor_parallel` for TP (#5048)
`mp_size` field is deprecated in flavor of `tensor_parallel`/`tp`
https://github.com/microsoft/DeepSpeed/blob/master/deepspeed/inference/engine.py
so update related docs that are still sticking to `mp_size`
Co-authored-by: Michael Wyatt <michaelwyatt@microsoft.com>