dmlc--dgl
esang
3fef5d27d3
* publish pct * add train_cls * add readme * update opt for point transformer * update the example index * update for comments Co-authored-by: Tong He <hetong007@gmail.com>
30 行
1.1 KiB
Markdown
30 行
1.1 KiB
Markdown
PCT
|
||
====
|
||
|
||
This is a reproduction of the paper: [PCT: Point cloud transformer](http://arxiv.org/abs/2012.09688).
|
||
|
||
# Performance
|
||
| Task | Dataset | Metric | Score - Paper | Score - DGL (Adam) | Time(s) - DGL |
|
||
|-----------------|------------|----------|------------------|-------------|-------------------|
|
||
| Classification | ModelNet40 | Accuracy | 93.2 | 92.1 | 740.0 |
|
||
| Part Segmentation | ShapeNet | mIoU | 86.4 | 85.6 | 390.0 |
|
||
|
||
+ Time(s) are the average training time per epoch, measured on EC2 g4dn.12xlarge instance w/ Tesla T4 GPU.
|
||
+ We run the code with the preprocessing used in [PointNet++](../pointnet). We can only get 84.5 for classification if we use the preprocessing described in the paper:
|
||
> During training, a random translation in [−0.2, 0.2], a random anisotropic scaling in [0.67, 1.5] and a random input dropout were applied to augment the input data.
|
||
|
||
|
||
# How to Run
|
||
|
||
For point cloud classification, run with
|
||
|
||
```python
|
||
python train_cls.py
|
||
```
|
||
|
||
For point cloud part-segmentation, run with
|
||
|
||
```python
|
||
python train_partseg.py
|
||
```
|