BigHugger
GH Repository · langfengQ

verl-agent

verl-agent is an extension of veRL, designed for training LLM/VLM agents via RL. verl-agent is also the official code for paper "Group-in-Group Policy Optimization for LLM Agent Training"

stars
2,313
30-day movement
starts with the next reading
Related entries
60
Connections
1
Pythonagent-frameworkgrpopythondeepseek-r1large-language-modelsgigporeinforcement-learningllm-trainingllm-agents

verl-agent is an extension of veRL for training LLM and VLM agents with reinforcement learning. It also serves as the official code release for the paper 'Group-in-Group Policy Optimization for LLM Agent Training' (GiGPO).

You want to apply RL training to language-model agents on top of the veRL stack, using the GiGPO method from the paper.

Use it to

  • Train LLM agents with reinforcement learning
  • Train VLM agents with RL
  • Reproduce GiGPO paper experiments
  • Extend veRL-based training pipelines to agentic settings

For Researchers and engineers training LLM/VLM agents with RL

Role
agent-framework
Language
Python
Licence
Apache-2.0
Forks
222
Open issues
59
Last push
2026-06-09
Latest release
v0.1.0 · 2025-12-11
topicsllm-agentsreinforcement-learningllm-traininggrpoagent-frameworkgigpo