hailanyi / VirConv

Virtual Sparse Convolution for Multimodal 3D Object Detection
https://arxiv.org/abs/2303.02314
Apache License 2.0
266 stars 39 forks source link

Questions about the "def roi_grid_pool_mm" in the VirConv-L? #40

Closed weilli closed 1 year ago

weilli commented 1 year ago

Thanks for your excellent job! I have some questions about the code. I thought the VirConv-L is the early fusion. Why does the function "roi_grid_pool_mm exist" in the TEDMHead?

hailanyi commented 1 year ago

Just for code simplicity. The feature will be replaced by early fused feature if no image-only feature provided. Please refer to line 608 to 611 in ted_head.py .