Skip to content

Latest commit

 

History

History
88 lines (65 loc) · 2.79 KB

README.md

File metadata and controls

88 lines (65 loc) · 2.79 KB
Raw

Road Segmentation: HighwayToCIL

Eiman Alnuaimi, Alessandro Cabodi, Dimitri Francolla, Rafael Wanner

Segmentation Example

This repository contains the developed code for the Road Segmentation Project of the Computational Intelligence Lab 2023.

Get Started

The easiest way to get started is to create a conda environment. Here is a guide to install Anaconda. Once installed run the following:

conda create -n cil python=3.11
conda activate cil
conda install pytorch torchvision torchaudio pytorch-cuda=11.7 -c pytorch -c nvidia
pip install -r requirements.txt

Download the models

We provide checkpoints for the DLinkNet101_JPU and NL_DLinkNet34.

wget -O DLinkNet101_JPU.pth https://polybox.ethz.ch/index.php/s/SDHy5sud4bv6Paw/download
wget -O NL_DLinkNet34.pth https://polybox.ethz.ch/index.php/s/5xlz1eghH3qoI4s/download

Download the dataset

wget -O dataset.zip https://polybox.ethz.ch/index.php/s/gdZSy2zFOwIoGSI/download
unzip dataset.zip

Run the models

In order to run the different models, you can use the run.py script:

python3 run.py ACTION MODEL FLAGS
ACTION Description
train trains the specified model
predict predicts segmentation masks for the provided images
MODEL
Unet
ResUnet
LinkNet34
DLinkNet34
DLinkNet101
NL_DLinkNet34
DLinkNet101_JPU
FLAGS for ACTION train Description
--data_path specify where the data for training is located
--loss specify the loss, options = {"BCE", "FocalLoss", "HybridLoss"}
--num_epochs specify the maximum number of epochs to train
--batch_size specify the batch size to train on
FLAGS for ACTION predict Description
--data_path specify where the data to segment is located
--checkpoint specify the checkpoint for the model
--ensemble specify if ensembling should be used

Examples:

  1. Train the Unet using the BinaryCrossEntropyLoss for a maximum of 100 epochs with batch size 8:
python3 run.py train Unet --data_path dataset/kaggle --loss BCE --num_epochs 100 --batch_size 8
  1. Predict segmentation masks using the NL_DLinkNet34 and DLinkNet34 in ensemble mode
python3 run.py predict DLinkNet34 NL_DLinkNet34 --data_path dataset/kaggle/test/images --checkpoint checkpoints/DLinkNet34.pth checkpoints/NL_DLinkNet34.pth--ensemble