search

ShichenLiu / SoftRas

Project page of paper "Soft Rasterizer: A Differentiable Renderer for Image-based 3D Reasoning"
MIT License
1.24k stars 156 forks source link

Memory consumption during data set loading during training #53

Closed zyz-notebooks closed 3 years ago

zyz-notebooks commented 4 years ago

Hello, ShichenLiu ! Thank you for your work. When I load data during training, it takes up a lot of my memory. My memory configuration is only 16g. How should I set it to avoid this kind of situation? I look forward to your reply

ShichenLiu commented 3 years ago

You can write your own dataloader: (1) save each image in the huge tensor as separate image files (.png); (2) in the dataloader, read the images on-the-fly in the getitem method