Reinforcement Learning
stable-baselines3
English
safe-rl
human-robot-interaction
benchmark
ppo
calibration
Instructions to use inq-android/eedgym-ckpts with libraries, inference providers, notebooks, and local apps. Follow these links to get started.
- Libraries
- stable-baselines3
How to use inq-android/eedgym-ckpts with stable-baselines3:
from huggingface_sb3 import load_from_hub checkpoint = load_from_hub( repo_id="inq-android/eedgym-ckpts", filename="{MODEL FILENAME}.zip", ) - Notebooks
- Google Colab
- Kaggle
EED Gym baseline and ablation checkpoints
This repository contains RL baseline and ablation study checkpoints for the Empathic Ethical Disobedience (EED) Gym environment.
40 zipped checkpoint files
Each folder corresponds to a trained baseline run and contains 5
.zipfiles (5 seeds)- Vanilla PPO
- PPO-LSTM
- Masked PPO
- Lagrangian PPO
The
ablationsfolder contains 4 subfolders with checkpoints for different ablation described in the paper, namely:- no affect ablation
- no clarify/alternative ablation
- no curriculum ablation
- no trust penalty ablation
Usage
Download with git lfs or via direct file links, e.g.:
wget https://huggingface.co/inq-android/eedgym-ckpts/resolve/main/ppo_vanilla/vanilla_seed0.zip
- Downloads last month
- 4