search

mileyan / pseudo_lidar

(CVPR 2019) Pseudo-LiDAR from Visual Depth Estimation: Bridging the Gap in 3D Object Detection for Autonomous Driving
https://mileyan.github.io/pseudo_lidar/
MIT License
967 stars 215 forks source link

Evaluating based on the pretrained AVOD weights #19

Open ghost opened 4 years ago

ghost commented 4 years ago

Please forgive my naivety here!

Based on the README, there's a link to the weights you have trained from AVOD using the pseudo-LiDAR point clouds.

However, the code provided by the AVOD team only has ways to evaluate or run inference using checkpoints.

Am I missing something here? Or is it just not possible to run their experiments using your weights.

Meaning I would have to either train their model, or write code to load the weights and run the inference that way, rather than the way they've coded it?

Am I missing something?

sarimmehdi commented 4 years ago

Hello. I ran into the same issue. Can you tell me whether you were able to fix it?

mileyan commented 4 years ago

Hi, I have made an avod. It can directly run pseudo-lidar. Please try it https://github.com/mileyan/avod_pl.