TorchSeg
This project aims at providing a fast, modular reference implementation for semantic segmentation models using PyTorch.
Highlights
- Modular Design: easily construct customized semantic segmentation models by combining different components.
- Distributed Training: >60% faster than the multi-thread parallel method(nn.DataParallel), we use the multi-processing parallel method.
- Multi-GPU training and inference: support different manners of inference.
- Provides pre-trained models and implement different semantic segmentation models.
Prerequisites
- PyTorch 1.0
pip3 install torch torchvision
- Easydict
pip3 install easydict
- Apex
- Ninja
sudo apt-get install ninja-build
- tqdm
pip3 install tqdm
Updates
v0.1.1 (05/14/2019)
- Release the pre-trained models and all trained models
- Add PSANet for ADE20K
- Add support for CamVid, PASCAL-Context datasets
- Start only supporting the distributed training manner
Model Zoo
Pretrained Model
Supported Model
Performance and Benchmarks
SS:Single Scale MSF:Multi-scale + Flip
PASCAL VOC 2012
Methods | Backbone | TrainSet | EvalSet | Mean IoU(ss) | Mean IoU(msf) | Model |
---|---|---|---|---|---|---|
FCN-32s | R101_v1c | train_aug | val | 71.26 | - | |
DFN(paper) | R101_v1c | train_aug | val | 79.67 | 80.6* | |
DFN(ours) | R101_v1c | train_aug | val | 79.40 | 81.40 | GoogleDrive |
80.6*: this result reported in paper is further finetuned on train dataset.
Cityscapes
Non-real-time Methods
Methods | Backbone | OHEM | TrainSet | EvalSet | Mean IoU(ss) | Mean IoU(msf) | Model |
---|---|---|---|---|---|---|---|
DFN(paper) | R101_v1c | β | train_fine | val | 78.5 | 79.3 | |
DFN(ours) | R101_v1c | β | train_fine | val | 79.09 | 80.41 | GoogleDrive |
DFN(ours) | R101_v1c | β | train_fine | val | 79.16 | 80.53 | GoogleDrive |
BiSeNet(paper) | R101_v1c | β | train_fine | val | - | 80.3 | |
BiSeNet(ours) | R101_v1c | β | train_fine | val | 79.09 | 80.39 | GoogleDrive |
BiSeNet(paper) | R18 | β | train_fine | val | 76.21 | 78.57 | |
BiSeNet(ours) | R18 | β | train_fine | val | 76.28 | 78.00 | GoogleDrive |
BiSeNet(paper) | X39 | β | train_fine | val | 70.1 | 72 | |
BiSeNet(ours)* | X39 | β | train_fine | val | 70.32 | 72.06 | GoogleDrive |
Real-time Methods
Methods | Backbone | OHEM | TrainSet | EvalSet | Mean IoU | Model |
---|---|---|---|---|---|---|
BiSeNet(paper) | R18 | β | train_fine | val | 74.8 | |
BiSeNet(ours) | R18 | β | train_fine | val | 74.83 | GoogleDrive |
BiSeNet(paper) | X39 | β | train_fine | val | 69 | |
BiSeNet(ours)* | X39 | β | train_fine | val | 68.51 | GoogleDrive |
BiSeNet(ours)*: because we didn't pre-train the Xception39 model on ImageNet in PyTorch, we train this experiment from scratch. We will release the pre-trained Xception39 model in PyTorch and the corresponding experiment.
ADE
Methods | Backbone | TrainSet | EvalSet | Mean IoU(ss) | Accuracy(ss) | Model |
---|---|---|---|---|---|---|
PSPNet(paper) | R50_v1c | train | val | 41.68 | 80.04 | |
PSPNet(ours) | R50_v1c | train | val | 41.65 | 79.74 | GoogleDrive |
PSPNet(paper) | R101_v1c | train | val | 41.96 | 80.64 | |
PSPNet(ours) | R101_v1c | train | val | 42.89 | 80.55 | GoogleDrive |
PSANet(paper) | R50_v1c | train | val | 41.92 | 80.17 | |
PSANet(ours)* | R50_v1c | train | val | 41.67 | 80.09 | GoogleDrive |
PSANet(paper) | R101_v1c | train | val | 42.75 | 80.71 | |
PSANet(ours) | R101_v1c | train | val | 43.04 | 80.56 | GoogleDrive |
PSANet(ours)*: The original PSANet in the paper constructs the attention map with over-parameters, while we only predict the attention map with the same size of the feature map. The performance is almost similar to the original one.
To Do
- offer comprehensive documents
- support more semantic segmentation models
- Deeplab v3 / Deeplab v3+
- DenseASPP
- EncNet
- OCNet
Training
- create the config file of dataset:
train.txt
,val.txt
,test.txt
file structureοΌ(split withtab
)path-of-the-image path-of-the-groundtruth
- modify the
config.py
according to your requirements - train a network:
Distributed Training
We use the official torch.distributed.launch
in order to launch multi-gpu training.
This utility function from PyTorch spawns as many Python processes as the number
of GPUs we want to use, and each Python process will only use a single GPU.
For each experiment, you can just run this script:
export NGPUS=8
python -m torch.distributed.launch --nproc_per_node=$NGPUS train.py
Inference
In the evaluator, we have implemented the multi-gpu inference base on the multi-process. In the inference phase, the function will spawns as many Python processes as the number of GPUs we want to use, and each Python process will handle a subset of the whole evaluation dataset on a single GPU.
- evaluate a trained network on the validation set:
python3 eval.py
- input arguments:
usage: -e epoch_idx -d device_idx [--verbose ] [--show_image] [--save_path Pred_Save_Path]
Disclaimer
This project is under active development. So things that are currently working might break in a future release. However, feel free to open issue if you get stuck anywhere.
Citation
The following are BibTeX references. The BibTeX entry requires the url LaTeX package.
Please consider citing this project in your publications if it helps your research.
@misc{torchseg2019,
author = {Yu, Changqian},
title = {TorchSeg},
howpublished = {\url{https://github.com/ycszen/TorchSeg}},
year = {2019}
}
Please consider citing the DFN in your publications if it helps your research.
@inproceedings{yu2018dfn,
title={Learning a Discriminative Feature Network for Semantic Segmentation},
author={Yu, Changqian and Wang, Jingbo and Peng, Chao and Gao, Changxin and Yu, Gang and Sang, Nong},
booktitle={Proceedings of the IEEE Conference on Computer Vision and Pattern Recognition (CVPR)},
year={2018}
}
Please consider citing the BiSeNet in your publications if it helps your research.
@inproceedings{yu2018bisenet,
title={Bisenet: Bilateral segmentation network for real-time semantic segmentation},
author={Yu, Changqian and Wang, Jingbo and Peng, Chao and Gao, Changxin and Yu, Gang and Sang, Nong},
booktitle={European Conference on Computer Vision},
pages={334--349},
year={2018},
organization={Springer}
}
Why this name, Furnace?
Furnace means the Alchemical Furnace. We all are the Alchemist, so I hope everyone can have a good alchemical furnace to practice the Alchemy. Hope you can be a excellent alchemist.