test

ESPNet: Towards Fast and Efficient Semantic Segmentation on the Embedded Devices

This folder contains the python scripts for running our pretrained models on the Cityscape dataset.

Getting Started

We provide the pretrained weights for ESPNet and ESPNet-C. Recall that ESPNet is the same as ESPNet-C, but with light weight decoder.

Pre-requisites:

• By default, we expect all images inside the ./data directory. If they are in different directory, please change the data_dir argument in the VisualizeResults.py file.

• Also, if the image format is different (e.g. jpg), please change in the VisualizeResults.py file.

This can be done using the below command:

python VisualizeResults.py --data_dir <data_dir> --img_extn <image extension>

Running ESPNet-C models

To run the ESPNet-C models, execute the following commands

python VisualizeResults.py --modelType 2 --p 2 --q 3

Here, p and q are the depth multipliers. Our models only support p=2 and q=3,5,8

Running ESPNet models

To run the ESPNet models, execute the following commands

python VisualizeResults.py --modelType 1 --p 2 --q 3

Here, p and q are the depth multipliers. Our models only support p=2 and q=3,5,8