Fei Xia

It's because space7 doesn't exist in your dataset, did you download the dataset?

- Reward: alive_score is a reward function to prevent agent from tipping over; progress is the difference of the potential function for two consecutive timesteps (dense reward); obstacle distance penalize...

Can you plot your reward curve during your training process? This would be insightful! Thanks.

The code was tested with `pytorch==0.3.1` and needs to be adapted for higher pytorch version.

These are the partition files that I used. [test.txt](https://github.com/fxia22/pointnet.pytorch/files/3914859/test.txt) [train.txt](https://github.com/fxia22/pointnet.pytorch/files/3914860/train.txt) [trainval.txt](https://github.com/fxia22/pointnet.pytorch/files/3914861/trainval.txt) [val.txt](https://github.com/fxia22/pointnet.pytorch/files/3914862/val.txt)

@RyanCV the original EMD code was developed for 6D point cloud (xyz + normal), so there is 6. You can set the last 3 dimensions to 0 for xyz only...

In the example I used it for object part segmentation so T-net is still relevant. If you use it for segmentation for point cloud in the wild feel free to...

thanks for the feedback, I will double check.

Can you let me know your software stack version? @WangZixuan

thanks, that's helpful, what about pytorch version? @WangZixuan