Weizhi-Zhong / IP_LAP

CVPR2023 talking face implementation for Identity-Preserving Talking Face Generation With Landmark and Appearance Priors

Geek Repo:Geek Repo

Github PK Tool:Github PK Tool

Segmentation fault

Gpwner opened this issue · comments

commented

在训练video_render的时候使用8*V100/16GB 的时候会提示“Segmentation fault”;但当把batch-size该成16的时候就可以正常训练了,请问这个是正常的吗 ?如果想尽可能地利用我的8张V100 应该如何修改代码呢 ?

Hi, thanks for your interest.
I think this is normal, as the memory of a single V100 is 16GB.

The question of how to fully utilize multi-GPU memory is related to engineering optimization.
I am sorry that I am not able to answer.
Maybe you can refer to this link or others.