search

eleurent / rl-agents

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

Some problems in the evaluation.py #117

Open Louise599 opened 6 months ago

Louise599 commented 6 months ago

In function def after_all_episodes(self, episode, rewards, duration), I tried to change the output of logger by changing “logger.info("Episode {} score: {:.1f}".format(episode, sum(rewards)))” to logger.info("Episode:{} score:{:.1f} length:{:.1f}".format(episode, sum(rewards), len(rewards))).

However, after running command python experiments.py evaluate configs/IntersectionEnv/env.json \ configs/IntersectionEnv/agents/DQNAgent/ego_attention_2h.json \ --train --episodes=4000 --name-from-config again,

the output didn't change, I tried to trace the function call and found that file evaluation.py doesn't seem to have been called, do you have any idea about what is the reason for this?

This has been bothering me for a long time and I am very much looking forward to your help in solving it.

eleurent commented 6 months ago

Maybe you have a version of rl-agents installed locally, and this is the one that's being used instead of your local version with your changes.

You can try to uninstall it with pip uninstall rl_agents