Validation of QT-Opt - Githubissues

DeepX-inc / machina

Control section: Deep Reinforcement Learning framework

MIT License

278 stars 45 forks source link

Validation of QT-Opt #264

Open leinxx opened 4 years ago

leinxx commented 4 years ago

Hi Machina,

Great job in implementing the QT-Opt algorithm. This repo is the only full implementation of QT-Opt I found. Do you have validation results on the current implementation?

Thanks,

Lei

rarilurelo commented 4 years ago

We tried QT-Opt on Pendulum and reacher (ReacherBulletEnv-v0), and we confirmed to learn as these graphs show (compared with TRPO). Input information is not image but default input information of the envs in these experiments.