uber-research / poet

Paired Open-Ended Trailblazer (POET) and Enhanced POET

POET and Enhanced POET

This repo contains implementation of the POET and Enhanced POET algorithms described in:

Paired Open-Ended Trailblazer (POET): Endlessly Generating Increasingly Complex and Diverse Learning Environments and Their Solutions

Enhanced POET: Open-Ended Reinforcement Learning through Unbounded Invention of Learning Challenges and their Solutions

An article on Uber Engineering Blog describing POET can be found here.

Requirements

• Fiber
• NEAT-Python
• OpenAI Gym

Run Enhanced POET locally

To run locally on a multicore machine

./run_poet_local.sh final_test

Run Enhanced POET on a Kubernetes cluster

Follow instructions here to create a persistent volume.

Then run the following command:

./run_poet_remote.sh final_test

To get the training logs:

fiber cp nfs:/persistent/logs/final_test . fiber cp nfs:/persistent/logs/poet_final_test poet_final_test

Run Enhanced POET on a computer cluster

To containerize and run the code on a computer cluster (e.g., Google Kubernetes Engine on Google Cloud), please refer to Fiber Documentation.

Run original POET

Use this legacy branch