[Examples] RL example implementation
See original GitHub issueI’m trying to port actor-critic example to ignite. Here is ported code. The logic is not that straight forward but if user understands how ignite.engine
works it can be implemented without problems.
Some term correspondences:
episode <-> epoch
timestep <-> iteration
However, ignite lacks of an option that maybe typical for RL, that when episode is finished when agent is failed for example, we need to stop our “batch” iterations, execute handlers and go on the next epoch. On the other hand ignite.engine
provides should_terminate
flag that can stops the training and it is not what we would like to do.
Any thoughts?
Issue Analytics
- State:
- Created 5 years ago
- Comments:7 (4 by maintainers)
Top Results From Across the Web
Best Reinforcement Learning Tutorials, Examples, Projects ...
1. RL with Mario Bros · 2. Machine Learning for Humans: Reinforcement Learning · 3. An introduction to Reinforcement Learning · 4. Reinforcement...
Read more >Beginner's guide to Reinforcement Learning & its ...
Learn problem formulation, Q learning and RL examples. ... Reinforcement learning problems; An implementation of Reinforcement Learning ...
Read more >A Beginner's Guide to Reinforcement Learning and its Basic ...
A Reinforcement Learning (RL) task is about training an agent that interacts with its environment. The agent transitions between different scenarios of the ......
Read more >Reinforcement Learning Real-world examples - Data Analytics
In this blog post, we'll learn about some real-world / real-life examples of Reinforcement learning, one of the different approaches to ...
Read more >Reinforcement Learning: What is, Algorithms, Types & Examples
Example of Reinforcement Learning · Your cat is an agent that is exposed to the environment. · Our agent reacts by performing an...
Read more >Top Related Medium Post
No results found
Top Related StackOverflow Question
No results found
Troubleshoot Live Code
Lightrun enables developers to add logs, metrics and snapshots to live code - no restarts or redeploys required.
Start FreeTop Related Reddit Thread
No results found
Top Related Hackernoon Post
No results found
Top Related Tweet
No results found
Top Related Dev.to Post
No results found
Top Related Hashnode Post
No results found
Top GitHub Comments
@alykhantejani yes, exactly. Will send a PR today with actor-critic example
We should be able to do this with a normal handler attached to the
ITERATION_COMPLETE
event. we just need anotherterminate
flag that terminates the epoch but not the whole trainer.Now I understand your original message 😃