search

hollow-503 / UniM2AE

[ECCV2024] UniM2AE: Multi-modal Masked Autoencoders with Unified 3D Representation for 3D Perception in Autonomous Driving
Apache License 2.0
38 stars 0 forks source link

Pretraining #3

Closed konyul closed 10 months ago

konyul commented 11 months ago

Hi, Thanks for sharing such a great work. Did you use swint-nuImages checkpoint for BEVFusion-SST or UniM2AE?

hollow-503 commented 10 months ago

Sorry for the late reply. We used the swint-nuImages checkpoint for fair comparison with other SOTA methods. It should be clarified that we demonstrate the effectiveness of UniM2AE on camera encoder in our paper. We used it just for the alignment with other SOTA methods.

konyul commented 10 months ago

Thank you for the reply