add pytorch implementation

.gitignore

@@ -1,4 +1,4 @@
 data/
 log/
 *.pyc
-
+.DS_Store

README.md

@@ -4,28 +4,22 @@ We propose a new neural network module dubbed EdgeConv suitable for CNN-based hi
 [[Project]](https://liuziwei7.github.io/projects/DGCNN) [[Paper]](https://arxiv.org/abs/1801.07829)     
 
 ## Overview
-`DGCNN` is the author's re-implementation of Dynamic Graph CNN, which achieves state-of-the-art performance on point-cloud-related high-level tasks including category classification, semantic segmentation and part segmentation. 
+`DGCNN` is the author's re-implementation of Dynamic Graph CNN, which achieves state-of-the-art performance on point-cloud-related high-level tasks including category classification, semantic segmentation and part segmentation.
 
-<img src='./misc/demo_teaser.png' width=800>
+<img src='./tensorflow/misc/demo_teaser.png' width=800>
 
 Further information please contact [Yue Wang](https://www.csail.mit.edu/person/yue-wang) and [Yongbin Sun](https://autoid.mit.edu/people-2).
 
-## Other Implementations
-* [pytorch-dgcnn](https://rusty1s.github.io/pytorch_geometric/build/html/modules/nn.html#torch_geometric.nn.conv.EdgeConv)
+## Author's Implementations
+
+The classification experiments in our paper are done with the pytorch implementation.
 
-## Requirements
-* [TensorFlow](https://www.tensorflow.org/)
+* [tensorflow-dgcnn](./tensorflow)
+* [pytorch-dgcnn](./pytorch)
 
-## Point Cloud Classification
-* Run the training script:
-``` bash
-python train.py
-```
-* Run the evaluation script after training finished:
-``` bash
-python evalutate.py
+## Other Implementations
+* [pytorch-geometric](https://rusty1s.github.io/pytorch_geometric/build/html/modules/nn.html#torch_geometric.nn.conv.EdgeConv)
 
-```
 
 ## Citation
 Please cite this paper if you want to use it in your work,

pytorch/README.md

@@ -0,0 +1,33 @@
+# Dynamic Graph CNN for Learning on Point Clouds (PyTorch)
+
+## Point Cloud Classification
+* Run the training script:
+
+
+``` 1024 points
+python main.py --exp_name=dgcnn_1024 --model=dgcnn --num_points=1024 --k=20 --use_sgd=True
+```
+
+``` 2048 points
+python main.py --exp_name=dgcnn_2048 --model=dgcnn --num_points=2048 --k=40 --use_sgd=True
+```
+
+* Run the evaluation script after training finished:
+
+``` 1024 points
+python main.py --exp_name=dgcnn_1024_eval --model=dgcnn --num_points=1024 --k=20 --use_sgd=True --eval=True --model_path=checkpoints/dgcnn_1024/models/model.t7
+```
+
+``` 2048 points
+python main.py --exp_name=dgcnn_2048_eval --model=dgcnn --num_points=2048 --k=40 --use_sgd=True --eval=True --model_path=checkpoints/dgcnn_2048/models/model.t7
+```
+
+* Run the evaluation script with pretrained models:
+
+``` 1024 points
+python main.py --exp_name=dgcnn_1024_eval --model=dgcnn --num_points=1024 --k=20 --use_sgd=True --eval=True --model_path=pretrained/model.1024.t7
+```
+
+``` 2048 points
+python main.py --exp_name=dgcnn_2048_eval --model=dgcnn --num_points=2048 --k=40 --use_sgd=True --eval=True --model_path=pretrained/model.2048.t7
+```

tensorflow/README.md

@@ -0,0 +1,15 @@
+# Dynamic Graph CNN for Learning on Point Clouds (TensorFlow)
+
+## Point Cloud Classification
+* Run the training script:
+
+``` bash
+python train.py
+```
+
+* Run the evaluation script after training finished:
+
+``` bash
+python evalutate.py
+
+```