Original Repo The Sematic Segmentation Model Predicts 19 Classes of Cityscape Dataset. This code works on cpu as well. Requirements pip install -r requirement.txt Pretrained Weigths Cityscape Sample Command to Run Inference on Folder of Images python test.py --model_path weights/hardnet70_cityscapes_model.pkl --input sample_input/ --output sample_output/ Prediction Sample