by qiqihezh · Agent Tool · ★ 84
Fixing GRPO training collapse in long-horizon multi-tool agents. A lightweight PRM-Lite + LATA joint approach achieves +37% over vanilla GRPO on τ-bench airline (50-task, multi-turn).
| Stars | 84 |
| Forks | 8 |
| Language | Python |
| Category | Agent Tool |
| Quality Score | 52.2894523196696/100 |
| Open Issues | 1 |
| Last Updated | 2026-06-27 |
| Created | 2026-04-29 |
| Platforms | python |
| Est. Tokens | ~13k |
Explore other popular agent tool tools:
agentic-grpo-longhorizon is Fixing GRPO training collapse in long-horizon multi-tool agents. A lightweight PRM-Lite + LATA joint approach achieves +37% over vanilla GRPO on τ-bench airline (50-task, multi-turn).. It is categorized as a Agent Tool with 84 GitHub stars.
agentic-grpo-longhorizon is primarily written in Python. It covers topics such as agentic-ai, grpo, long-horizon.
You can find installation instructions and usage details in the agentic-grpo-longhorizon GitHub repository at github.com/qiqihezh/agentic-grpo-longhorizon. The project has 84 stars and 8 forks, indicating an active community.