KhoomeiK/LlamaGym
A Python library that abstracts the complexity of training LLM agents with reinforcement learning on Gym environments.
Not currently ranked — collecting fresh signals.
star history
LlamaGym provides an Agent abstract class that handles the machinery required for online RL fine-tuning of LLM-based agents. It manages LLM conversation context, episode batching, reward assignment, and PPO setup, letting developers quickly experiment with agent prompting and hyperparameters across any Gymnasium environment. The library targets researchers and developers building autonomous agents that learn through interaction and reward signals.
Frequently asked
- What is KhoomeiK/LlamaGym?
- A Python library that abstracts the complexity of training LLM agents with reinforcement learning on Gym environments.
- Is LlamaGym open source?
- Yes — KhoomeiK/LlamaGym is open source, released under the MIT license.
- What language is LlamaGym written in?
- KhoomeiK/LlamaGym is primarily written in Python.
- How popular is LlamaGym?
- KhoomeiK/LlamaGym has 1.3k stars on GitHub.
- Where can I find LlamaGym?
- KhoomeiK/LlamaGym is on GitHub at https://github.com/KhoomeiK/LlamaGym.