https://github.com/kuixu/3ddensenet.torch
3D DenseNet(torch version) for ModelNet40 dataset
https://github.com/kuixu/3ddensenet.torch
3d-densenet 3d-models 3d-volume 3dshapenet densenet mesh modelnet
Last synced: about 1 month ago
JSON representation
3D DenseNet(torch version) for ModelNet40 dataset
- Host: GitHub
- URL: https://github.com/kuixu/3ddensenet.torch
- Owner: kuixu
- Created: 2017-08-17T03:54:08.000Z (over 7 years ago)
- Default Branch: master
- Last Pushed: 2017-10-19T02:35:34.000Z (over 7 years ago)
- Last Synced: 2025-03-21T15:42:00.963Z (about 1 month ago)
- Topics: 3d-densenet, 3d-models, 3d-volume, 3dshapenet, densenet, mesh, modelnet
- Language: Lua
- Homepage:
- Size: 10.8 MB
- Stars: 44
- Watchers: 3
- Forks: 16
- Open Issues: 1
-
Metadata Files:
- Readme: README.md
- Contributing: CONTRIBUTING.md
Awesome Lists containing this project
README
3D DenseNet in torch
============================
3D DenseNet is using 3D Convolutional(VolumetricConvolution in torch), Pooling, BatchNormalization layers with 3D kernel. This implements is based on [DenseNet](https://github.com/liuzhuang13/DenseNet) and [fb.resnet.torch](https://github.com/facebook/fb.resnet.torch/). DenseNet introduced in the paper "Densely Connected Convolutional Networks" (CVPR 2017, Best Paper Award)
## Requirements
See the [installation instructions](https://github.com/facebook/fb.resnet.torch/blob/master/INSTALL.md) for a step-by-step guide.
- luarock install hdf5 nninit
- Download the [ModelNet40](http://3dshapenets.cs.princeton.edu/) dataset and hdf5 format ([Google Drive](https://drive.google.com/file/d/0B1S5mVkJgRwqRUZjakNuMi1yck0/view?usp=sharing), [Baidu Cloud Disk](https://pan.baidu.com/s/1hsN81qG))
## Dataset
1. Download data through above link;
2. and modify the file path in `train.list` and `test.list` file;
3. then modify the `datadir` variable in `examples/run_modelnet40.sh`.
## Training
See the [training recipes](https://github.com/facebook/fb.resnet.torch/blob/master/TRAINING.md) for addition examples.
For Modelnet40, just run shell `examples/run_modelnet40.sh 0,1`, `0,1` is the GPU ids with multi-GPU supported.
```bash
cd examples
./run_modelnet40_h5.sh 0,1
```
## Trained models
#### modelnet40_60x validation error rate
| Network | Top-1 error | Top-5 error |
| -------------- | ----------- | ----------- |
| Voxnet | 13.74 | 1.92 |
| DenseNet-20-12 | 12.99 | 2.03 |
| DenseNet-30-12 | 12.11 | 1.94 |
| DenseNet-30-16 | 11.08 | 1.61 |
| DenseNet-40-12 | 11.57 | 1.78 |
## Notes
This implementation differs from the ResNet paper in a few ways:
**3D Convolution**: We use the [VolumetricConvolution](https://github.com/torch/nn/blob/master/doc/convolution.md) to implement 3D Convolution.