Adaptive-Texture-Filtering-for-Single-Domain-Generalized-Segmentation

This repository provides the official PyTorch implementation of the following paper:
Adaptive Texture Filtering for Single Domain Generalized Segmentation

Congifuration Environment

• python 3.7
• pytorch 1.8.2
• torchvision 0.9.2
• cuda 11.1

Before start

Download the GTA5, SYNTHIA, Cityscapes, BDD-100K, Mapillary datasets.

Training and test

Run like this: python main.py --gta5_data_path /dataset/GTA5 --city_data_path /dataset/cityscapes --cuda_device_id 0
Notice:
-First phase training: load the model from model_firststage.py
-Second phase training: load the model from model.py