chu

i did not install the tensorflow-gpu in my conda environment, but when i set GPU_COUNT = 1, it seems to use my gpu to inference, why??? 

Sorry bro I did not use pybullet and cannot help you, I changed to use blenderproc with my project 从 Windows 版邮件发送 发件人: Walter ***@***.***> 发送时间: 2022年9月19日 20:26 收件人: ***@***.***>...

you can use bop_toolkit to generate the [masks](https://github.com/thodan/bop_toolkit/blob/master/scripts/calc_gt_masks.py), see this example it may help you

> My dataset includes RGB image and depth image but I don't have modal segmasks(I'm not sure what it it). So how can I test on my dataset see this...