search

noahzn / Lite-Mono

[CVPR2023] Lite-Mono: A Lightweight CNN and Transformer Architecture for Self-Supervised Monocular Depth Estimation
MIT License
540 stars 61 forks source link

the CPU utilization has been very high, but the GPU utilization has been very low #136

Closed lshhan closed 6 months ago

lshhan commented 6 months ago

Hello, I would like to ask if the data loading part of the code is performed on the CPU? During my experiment, the CPU utilization has been very high, but the GPU utilization has been very low. What is the reason for this? cpu image

noahzn commented 6 months ago

Hi, yes, the data loading is performed on CPUs. Then, they are moved to GPUs (https://github.com/noahzn/Lite-Mono/blob/main/trainer.py#L269)

noahzn commented 6 months ago

I'm now closing this issue as there is no response.