Parallel inference for Wan2.1-I2V-14B-480P after LoRA training

[jxjx12138]

Wonderful job! Will Wan2.1 support parallel multi-GPU inference after LoRA training, like https://github.com/modelscope/DiffSynth-Studio/blob/main/examples/wanvideo/wan_14b_text_to_video_tensor_parallel.py?