search

Fictionarry / ER-NeRF

[ICCV'23] Efficient Region-Aware Neural Radiance Fields for High-Fidelity Talking Portrait Synthesis
https://fictionarry.github.io/ER-NeRF/
MIT License
894 stars 124 forks source link

Result of model training #11

Open linhnv97 opened 11 months ago

linhnv97 commented 11 months ago

Thank you for your project! After training model with new video input, I got the result like video below. Please tell me what is the reason

https://github.com/Fictionarry/ER-NeRF/assets/140048495/c72b36d9-be8e-42b3-9e37-220d740f6642

Fictionarry commented 11 months ago

Seems something goes wrong during training. Please notice the error info in the log, for example, nan loss or something else. And check whether the validation images in the workspace dir look good.

linhnv97 commented 11 months ago

Thank you for your reply, I have the error related "nan loss"