Model Parallel Training

[AKMourato]

Greetings.
Is there any functional module to perform model parallel training?
I'm not referring to multi-gpu training, where each GPU receives a replica of the model and different samples, but instead split up the model into different GPUs and therefore each GPU receives parts of a sample (i.e: https://pytorch.org/tutorials/intermediate/model_parallel_tutorial.html).

Resizing 4k images in some applications can reduce by a good amount the performance, therefore this would be a nice feature to have.
Thanks