Problem generating train pair data

[gdiazh]

Dear @wenbowen123, Hi thanks for your great work! I'm trying to generate data running the script blender_main.py (everything inside the docker) and using the default example 1.ply, I have just change the corresponding path in the config files and the background.

The data is generated but it always stops after 180 images, even though in the dataset_info.yml I have 'train_samples: 200000', I tried commenting some rm -rf commands in the 'blender_datset_generator.py' and changing the 'count' value every 180 images, but of course I cannot do this manually every time to generate the 200000 images. After the 180 generated images the output only says "Killed", do you know what this may be happening?

[wenbowen123]

did you monitor the memory usage? Seems like it's OOM.