zhangyp15 / OccFormer

[ICCV 2023] OccFormer: Dual-path Transformer for Vision-based 3D Semantic Occupancy Prediction
https://arxiv.org/abs/2304.05316
Apache License 2.0
324 stars 22 forks source link

The influence of not using the c++ extension which provided in the original implementation of Deformable DETR? #16

Open Fanqyu opened 10 months ago

Fanqyu commented 10 months ago

Hi there, it's a great work. I'd like to know if there is any impact, such as slow training or inference speed, etc, for only using the "CPU version of multi-scale deformable attention".