grpo-rl-training

SKILL.md

Expert-level guidance for implementing Group Relative Policy Optimization (GRPO) using the Transformer Reinforcement Learning (TRL) library. This skill provides battle-tested patterns, critical insights, and production-ready workflows for fine-tuning language models with custom reward functions.

| Correctness | Verifiable tasks (math, code) | 2.0 (highest) | | Format | Strict structure enforcement | 0.5-1.0 | | Length | Encourage verbosity/conciseness | 0.1-0.5 | | Style | Penalize unwanted patterns | -0.5 to 0.5 |

Critical Insight: Combine 3-5 reward functions for robust training. Order matters less than diversity of signals.

Orientación experta para el ajuste de GRPO/RL con TRL para el razonamiento y la capacitación de modelos de tareas específicas Fuente: orchestra-research/ai-research-skills.

Ver original