How to set constant learning rate in PPO1?

hill-a / stable-baselines

A fork of OpenAI Baselines, implementations of reinforcement learning algorithms

http://stable-baselines.readthedocs.io/

MIT License

4.14k stars 723 forks source link

Closed nicehzj closed 3 years ago

nicehzj commented 3 years ago

I checked the doc, but I still cannot find a way to set a constant learning rate in PPO1.

I tried a lot of ways, and the learning rate is always 1e-3.

Can anyone help me?

Thanks a lot

Miffyli commented 3 years ago

Did you mean setting a learning schedule? Because if learning rate is always 1e-3, it is being constant.

Edit: I recommend checking out stable-baselines3, where there is an example on learning rate schedule.

araffin commented 3 years ago

Closing as the answer is in the doc.