chungyiweng / humannerf

HumanNeRF turns a monocular video of moving people into a 360 free-viewpoint video.
MIT License
791 stars 87 forks source link

Always “Produce empty images; reload the init model.” #43

Closed surheaven closed 2 years ago

surheaven commented 2 years ago

1663748709084 When I was training my own data, I found that I could only train 100iter per epoch, and each epoch prompted”Produce empty images; reload the init model.“

cshen11 commented 2 years ago

@surheaven Hi, I had the same issue "Produce empty images; reload the init model". This consistently happens on all zju-mocap datasets and only happens with adventure.yaml. The config with single_gpu.yaml had no problem.

Can you share how you resolved the issue?