search

songw-zju / Meta-RangeSeg

The official implementation of "Meta-RangeSeg: LiDAR Sequence Semantic Segmentation Using Multiple Feature Aggregation" (RA-L with IROS 2022)
MIT License
45 stars 6 forks source link

Questions about residual image #9

Closed ayoung1002 closed 1 year ago

ayoung1002 commented 1 year ago

Hi, thank you for your great works!! :+1: :)

Could you tell me about 'residual' param? Can I set residual to FALSE?

songw-zju commented 1 year ago

Thanks for your interest and sorry for the later reply. The residual parameter refers to whether to use the residual image during training. It's set to True for multi-scan task and False for single-scan task.

ayoung1002 commented 1 year ago

Thanks for your reply!

If I set residual False, weight has an error! (different kernel size) (Btw, I success to infer your code when setting residual 'True'.) Can you update another pre-trained weights for single scan task?

songw-zju commented 1 year ago

Hi @ayoung1002, sorry for the late reply. I have retrained the single scan network with the clean code in this repository and update the pre-trained weights with the predictions on valid/test set.