search

eleurent / rl-agents

Implementations of Reinforcement Learning and Planning algorithms
MIT License
591 stars 153 forks source link

the instance command failed #77

Closed yuzhouxianzhi closed 2 years ago

yuzhouxianzhi commented 2 years ago

I run the instance command "python3 experiments.py evaluate configs/CartPoleEnv/env.json configs/CartPoleEnv/DQNAgent.json --train --episodes=200". It can work, but after 2 episodes, do not continue run. I wait for long time. It still not continue run. I check out directory. There is also 2 mp4 files. What's wrong with environment or configs?

eleurent commented 2 years ago

This is because by default, only a few episodes are rendered (so that training is faster). After training, you can visualise the trained policy by replacing --train by --test --restore.