Explicitly pass expert_tensor_parallel_size to initialize_model_parallel#537

ananthsub merged 2 commits into

NVIDIA-NeMo:mainNVIDIA-NeMo/Megatron-Bridge:mainfrom

nathan-az:pass-expert-tp-to-initnathan-az/Megatron-Bridge:pass-expert-tp-to-initCopy head branch name to clipboard