The render result is completely black with no person

Hi:
I tried training with the zju_mocap dataset over 18,000 times to get a model, but the rendered result was completely black with no person.
Can you give me some guidance？
Thank you！