yangyin2016 / ENet_PyTorch

Implementation of ENet (A Deep Neural Network Architecture for Real-Time Semantic Segmentation using PyTorch

ENet_PyTorch

Implementation of ENet: A Deep Neural Network Architecture for Real-Time Semantic Segmentation using PyTorch (v.1.10)

This implementation is validated on the CamVid dataset. The pre-trained of ENet trained with CamVid is available here

Dataset	Type	Classes	Input resolution	Batch size	Epochs	Mean IoU (%)	GFLOPS	Parameters
CamVid	Original	11	480x360	10	-	58.3	3.83	0.37M
CamVid	Implementation	11	480x360	10	100	58.6	2.34	0.35M

To Use

1. Clone the repository

git clone https://github.com/ntkhoa95/ENet_PyTorch.git
cd ENet_PyTorch

2. Download the CamVid datasets

Download the CamVid dataset and unzip to ./content/camvid/

wget https://www.dropbox.com/s/pxcz2wdz04zxocq/CamVid.zip?dl=1 -O CamVid.zip
unzip CamVid.zip

3. Use command to train the model

python init.py --mode train -iptr ./content/camvid/train/ -lptr ./content/camvid/trainannot/

4. Use command to test model

python init.py --mode test -m ./content/checkpoint/camvid/best_model.pth -i ./content/camvid/test/0001TP_008550.png

5. Use --help to get more commands

python init.py --help

Reference

1. A. Paszke, A. Chaurasia, S. Kim, and E. Culurciello. Enet: A deep neural network architecture for real-time semantic segmentation. arXiv preprint arXiv:1606.02147, 2016.