[Question] Does deepspeed support model parallelism via API "PipeModelDataParallelTopology"?

[chenyaofo]

Describe the Question I just notice code in https://github.com/HuangLK/llama-deepspeed/blob/faedea514b11c18c695e1b2a6adb63b102ef001c/models/llama_pipeline_model.py#L174

It appears that the code is utilizing the PipeModelDataParallelTopology API. I'm curious to know if it supports model parallelism during training.

I have extensively searched the Internet, documents, and issues, but unfortunately, I have been unable to find an answer to this question. Furthermore, it seems that there are no available documents providing an explanation or usage guide for this particular API.