fix(pytorch): correct epoch-based learning rate decay behavior

rickstaa / stable-learning-control

A framework for training theoretically stable (and robust) Reinforcement Learning control algorithms.

https://rickstaa.dev/stable-learning-control

MIT License

6 stars 1 forks source link

fix(pytorch): correct epoch-based learning rate decay behavior #410

Closed rickstaa closed 6 months ago

rickstaa commented 6 months ago

This pull request addresses a bug in the epoch-based learning rate decay mechanism. Previously, the decay process did not correctly reach the specified final learning rate. This fix ensures that the learning rate accurately decays to the intended final value throughout the epochs.