Added the verification of num_max_dispatch_tokens_per_rank to the decode operator adaptation layer.#330

Yael-X merged 1 commit intosgl-project:mainsgl-project/sgl-kernel-npu:mainfrom

oagniqgnat:fix_decode_max_tokenoagniqgnat/sgl-kernel-npu:fix_decode_max_tokenCopy head branch name to clipboard