Request for Memory Optimization Suggestions for Running image_to_3d.py on RTX 4060 (8GB VRAM)

[Vincia-Jun]

Dear Developers,

Thank you for your excellent work on this project. I've been attempting to run the script python apps/image_to_3d.py on my personal laptop, which is equipped with an RTX 4060 graphics card with 8GB of VRAM. Unfortunately, I encountered an "Out of memory" error during execution.

I'm writing to kindly ask if there are any additional configurations or optimization suggestions that could help reduce the memory usage, allowing the script to run successfully on my RTX 4060 with 8GB VRAM.

Any guidance you can provide would be greatly appreciated. Thank you for your time and assistance.

Best regards,
Jim

[HochCC]

Hi @Vincia-Jun,
The full apps/image_to_3d.py process requires at least 15GB of GPU memory. However, if you set DelightingModel toNone and set project_delight to False, the memory requirement drops to around 8GB. This setting is still quite risky, but you could give it a try if needed.