search

quantumiracle / Popular-RL-Algorithms

PyTorch implementation of Soft Actor-Critic (SAC), Twin Delayed DDPG (TD3), Actor-Critic (AC/A2C), Proximal Policy Optimization (PPO), QT-Opt, PointNet..
Apache License 2.0
1.14k stars 129 forks source link

please evaluate the performance over multiple seeds #47

Closed thlautenschlaeger closed 3 years ago

thlautenschlaeger commented 3 years ago

In the README down below, the performance evaluations are done only on a single seed. Please evaluate over multiple seeds because as of now the results can be random and are thus not comparable.

quantumiracle commented 3 years ago

I appreciate your suggestion. However, as this is not an officially released repo (I didn't spend much time on maintaining this), I will not provide thorough evaluation of algorithms. Please feel free to do this by yourself (it should be quite straightforward with a bash script passing random seeds into some arguments for running, since each algorithm script is independent), and I welcome such contribution in this repo if you have done that! Thanks.