isaac-sim / OmniIsaacGymEnvs

Reinforcement Learning Environments for Omniverse Isaac Gym
Other
762 stars 203 forks source link

【Open Question】About Deep Reinforcement Learning (DRL) #125

Open SengHongLEE opened 6 months ago

SengHongLEE commented 6 months ago

Hi there, I have some questions during the using experience in Isaac Gym. I am confused about why my testing can achieve an amazing reward value beyond the training result while I got a low-reward training model weight file (.pth/.pt) ?

Welcome to share any further understanding and insight about the DRL in this issue. Thanks, in advanced.

Regards.

image image