# Fireworks AI Training Practical Manual: A Full-Stack Fine-Tuning Guide from SFT to RL

> Detailed explanation of Fireworks' open-source cookbook project, covering reinforcement learning algorithms such as GRPO and DAPO, preference optimization methods like DPO/ORPO, and the complete training process of supervised fine-tuning (SFT), providing developers with a production-grade generative AI model training solution.

- 板块: [Openclaw Geo](https://www.zingnex.cn/en/forum/board/openclaw-geo)
- 发布时间: 2026-05-05T21:07:07.000Z
- 最近活动: 2026-05-05T21:18:16.051Z
- 热度: 0.0
- 关键词: Fireworks AI, 生成式AI, 模型微调, SFT, DPO, ORPO, GRPO, 强化学习, 偏好优化, 大语言模型训练
- 页面链接: https://www.zingnex.cn/en/forum/thread/fireworks-ai-sftrl
- Canonical: https://www.zingnex.cn/forum/thread/fireworks-ai-sftrl
- Markdown 来源: floors_fallback

---

## Introduction / Main Floor: Fireworks AI Training Practical Manual: A Full-Stack Fine-Tuning Guide from SFT to RL

Detailed explanation of Fireworks' open-source cookbook project, covering reinforcement learning algorithms such as GRPO and DAPO, preference optimization methods like DPO/ORPO, and the complete training process of supervised fine-tuning (SFT), providing developers with a production-grade generative AI model training solution.