evgenii-nikishin/rl_with_resets
JAX implementation of deep RL agents with resets from the paper "The Primacy Bias in Deep Reinforcement Learning"
This project provides a method to improve the performance of deep reinforcement learning agents by addressing a common issue called 'primacy bias.' It takes existing deep RL algorithms (like SAC, DrQ, or SPR) and enhances them with a 'resetting' mechanism. The output is a more robust and better-performing reinforcement learning agent, which would be used by researchers and practitioners developing and training autonomous agents in simulated environments.
105 stars. No commits in the last 6 months.
Use this if you are training deep reinforcement learning agents and find that they struggle to learn effectively after initial experiences or exhibit unstable performance.
Not ideal if you are not working with deep reinforcement learning, or if you require an out-of-the-box solution without modifying existing agent code.
Stars
105
Forks
7
Language
Python
License
MIT
Category
Last pushed
May 17, 2022
Commits (30d)
0
Get this data via API
curl "https://pt-edge.onrender.com/api/v1/quality/ml-frameworks/evgenii-nikishin/rl_with_resets"
Open to everyone — 100 requests/day, no key needed. Get a free key for 1,000/day.
Higher-rated alternatives
explosion/thinc
🔮 A refreshing functional take on deep learning, compatible with your favorite libraries
google-deepmind/optax
Optax is a gradient processing and optimization library for JAX.
patrick-kidger/diffrax
Numerical differential equation solvers in JAX. Autodifferentiable and GPU-capable....
google/grain
Library for reading and processing ML training data.
patrick-kidger/equinox
Elegant easy-to-use neural networks + scientific computing in JAX. https://docs.kidger.site/equinox/