Skip to content
This repository has been archived by the owner on Dec 29, 2022. It is now read-only.

Custom Training Loops for RL/GAN #2

Open
dennybritz opened this issue Mar 3, 2017 · 1 comment
Open

Custom Training Loops for RL/GAN #2

dennybritz opened this issue Mar 3, 2017 · 1 comment
Labels

Comments

@dennybritz
Copy link
Contributor

Add support for custom training loops to easily train GANs and RL algorithms. Two ideas on how to to implement this:

  • Make it part of the model and use SessionRunHooks to execute different train ops
  • Subclass Estimator with a custom MonitoredSession

Need to look into these options in more details.

@dennybritz
Copy link
Contributor Author

Idea: Have a TrainingRegime (or something like that) abstraction that defines how the loss or losses are being handled. Each regime is associated with a MonitoredSession loop.

Sign up for free to subscribe to this conversation on GitHub. Already have an account? Sign in.
Labels
Projects
None yet
Development

No branches or pull requests

1 participant