An out-of-memory (OOM) error occurred on an H20 with 8 GPUs and 96GB of VRAM when using wan_14b_text_to_video_tensor_parallel.py.

[KBRASK]

[weitunglin]

Hello, have you solved the issue? could you give me some insights related to this problem? I also want to run the wan 14B model with multiple gpu tensor parallel.