What is Trustvis?

How To Use it?

Environment Configuration

1. create conda environment

$ cd Vis
$ conda create -n trustvis python=3.7
$ (trustvis) conda activate visualizer

2. install pyTorch and CUDA For setting up PyTorch on that conda environment, use the guidelines provided at PyTorch's official local installation page. This guide will help you select the appropriate configuration based on your operating system, package manager, Python version, and CUDA version.

3. install requirements

$ (trustvis) pip install -r requirements.txt

evaluate subject model

$ conda activate trustvis
$(trustvis) python subject_model_eval.py

The trainig dynamic performance will be store in /training_dynamic/Model/subject_model_eval.json

Run trustvis

$ cd trustvis

train baseline visualization model

$ conda activate trustvis
$(trustvis) python base.py --epoch epoch_num --content_path training_dynamic folder's path

• the vis model will be store in /training_dynamic/Model/Epoch_{epoch_number}/base.pth
• the vis result will be store in /training_dynamic/Base/***.png
• the evaluation resulte wiil be store in /training_dynamic/Model/base_eval.json

train proxy only visualization model(for ablation study)

$(trustvis) proxy.py --epoch epoch_num --content_path training_dynamic folder's path

• the vis model will be store in /training_dynamic/Model/Epoch_{epoch_number}/proxy.pth
• the vis result will be store in /training_dynamic/Proxy/***.png
• the evaluation resulte wiil be store in /training_dynamic/Model/proxy_eval.json

train active learning only visualization model(for ablation study)

⚠️ proxy only visualization model should be trained and saved

$(trustvis) al_base.py --epoch epoch_num --content_path training_dynamic folder's path

• the vis model will be store in /training_dynamic/Model/Epoch_{epoch_number}/al_base.pth
• the vis result will be store in /training_dynamic/al_base/***.png
• the evaluation resulte wiil be store in /training_dynamic/Model/al_base_eval.json

train trustvis(proxy based + active learning)

⚠️ proxy only visualization model should be trained and saved

$(trustvis) al_proxy.py --epoch epoch_num --content_path training_dynamic folder's path

• the vis model will be store in /training_dynamic/Model/Epoch_{epoch_number}/trustvis.pth
• the vis result will be store in /training_dynamic/Trust_al/***.png
• the evaluation resulte wiil be store in /training_dynamic/Model/trustvis_al_eval.json