search

dvlab-research / LargeKernel3D

LargeKernel3D: Scaling up Kernels in 3D Sparse CNNs (CVPR 2023)
https://arxiv.org/abs/2206.10555
Apache License 2.0
188 stars 8 forks source link

runtime measurement #2

Closed konyul closed 1 year ago

konyul commented 1 year ago

How did you validate runtime of your model?? Thank you

yukang2017 commented 1 year ago

Thanks for your interest in our work. The runtime of our model is test on a NVIDIA 2080ti GPU. For nuScenes models, it is based on the original CenterPoint codebase. The implementation of SW-LK Conv is based on our improved version of spconv library for GPU acceleration. It will also be released.

Qidian213 commented 1 year ago

When will the code be released? Thanks.

yukang2017 commented 1 year ago

Thanks for your interest. I have released the core spatial sparse conv in this repo spconv-plus. The overall training code for this paper will be released latter.