ERROR calculating ambient loss

[enrico310786]

trafficstars

Hi,

I am trying to train the model on my own video. After some hours after launching the command

CUDA_VISIBLE_DEVICES=0 python tasks/run.py --config=egs/datasets/May/lm3d_radnerf_sr.yaml --exp_name=motion2video_nerf/may_head --reset

i got this warning/error. It's been going on like this for about two hours now. Is it normal? I have to correct the code eliminating the item() string?

Thanks for the help

[enrico310786]

The training is finished, but the total loss is none. Can I continue with the Torso NeRF training? Or is there something to correct?

[Net-Maker]

Hello, I met the same problem this morning. You can fine the answer in #24 .Hope it could solve your problem.

[Net-Maker]

By the way,the nan loss leads to a totally wrong result,you can check your tensorboard to find a checkpoint that can be use.

[RayDean]

same problem, how did you solved this problem in detail? thanks.

[Net-Maker]

just modified the file "--config=egs/datasets/May/lm3d_radnerf_sr.yaml",find the ”ambient_loss_target“ and increase the value