second training stage - Githubissues

rasd3 / 3D-CVF

[ECCV 2020] This is the official implementation of 3D-CVF: Generating Joint Camera and LiDAR Features Using Cross-View Spatial Feature Fusion for 3D Object Detection

125 stars 23 forks source link

second training stage #18

Open RickOnEarth opened 2 years ago

RickOnEarth commented 2 years ago

Has anyone run the second training stage successfully? Seems like the PointNet encoding part after RoI pooling is missing.

Eaphan commented 2 years ago

I have the same question with you. Can you give the complete code for reproducing the results claimed in paper? @rasd3