How to support topk>2 (topk=6 is needed in our experiment) in MoE model?

[Eutenacity]

we find that

@staticmethod
def supports_config(config: DSMoEConfig) -> bool:
if config.input_dtype != config.output_dtype:
return False

    if config.input_dtype != torch.float16 and config.input_dtype != torch.bfloat16:
        return False

    if config.top_k != 1 and config.top_k != 2:
        return False

    return True

in "deepspeed.inference.v2.modules.implementations.moe.cutlass_multi_gemm.DSMultiGemmMoE"

It means DSMoE only supoort topk=1 or topk =2.

But our experiment need support for topk=6

[YunxinLi]

Hello, we need a problem when we set ``WORLD_SIZE = 8
EP_WORLD_SIZE = 8
EXPERTS = 8'' for deepspeed moe layer. The forward propagation of the model will always stay on the moe calculation and cannot pass the moe layer.