hace 1 d
Let's build a reasoning LLM using GRPO, from scratch (100% local):
Today, we're going to learn how to turn any model into a reasoning powerhouse.
We'll do so without any labeled data or human intervention, using Reinforcement Finetuning (GRPO)!
Tech stack:
- @UnslothAI for efficient fine-tuning
- @HuggingFace TRL to apply GRPO
Let's go! 🚀
What is GRPO?
Group Relative Policy Optimization is a reinforcement learning method that fine-tunes LLMs for math and reasoning tasks using deterministic reward functions, eliminating the need for labeled data.
Here's a brief overview of GRPO before we jump into code:
1️⃣ Load the model
We start by loading Qwen3-4B-Base and its tokenizer using Unsloth.
You can use any other open-weight LLM here.
Check this 👇
2️⃣ Define LoRA config
We'll use LoRA to avoid fine-tuning the entire model weights. In this code, we use Unsloth's PEFT by specifying:
- The model
- LoRA low-rank (r)
- Modules for fine-tuning, etc.
Check this 👇
3️⃣ Create the dataset
We load Open R1 Math dataset (a math problem dataset) and format it for reasoning.
Each sample includes:
- A system prompt enforcing structured reasoning
- A question from the dataset
- The answer in the required format
Check this code 👇
4️⃣ Define reward functions
In GRPO we use deterministic functions to validate the response and assign a reward.
No manual labelling required!
The reward functions:
- Match format exactly
- Match format approximately
- Check the answer
- Check numbers
Check this out 👇
5️⃣ Use GRPO and start training
Now that we have the dataset and reward functions ready, it's time to apply GRPO.
HuggingFace TRL provides everything we described in the GRPO diagram, out of the box, in the form of the GRPOConfig and GRPOTrainer.
Check this out👇
6️⃣ Comparison
Again, we can see how GRPO turned a base model into a reasoning powerhouse.
Check this out👇
Before we conclude, let me address an important question:
When should you use reinforcement fine-tuning (RFT) versus supervised fine-tuning (SFT)?
I created this diagram to provide an answer:
Finally, I'll leave you with an overview of the GRPO process.
Let me know what other techniques you have used in the comments!
You can find all the code and everything you need on the @LightningAI⚡️Studio here:
13,8 mil
124
El contenido de esta página lo proporcionan terceros. A menos que se indique lo contrario, OKX no es el autor de los artículos citados y no reclama ningún derecho de autor sobre los materiales. El contenido se proporciona únicamente con fines informativos y no representa las opiniones de OKX. No pretende ser un respaldo de ningún tipo y no debe ser considerado como un consejo de inversión o una solicitud para comprar o vender activos digitales. En la medida en que la IA generativa se utiliza para proporcionar resúmenes u otra información, dicho contenido generado por IA puede ser inexacto o incoherente. Lee el artículo vinculado para obtener más detalles e información. OKX no es responsable del contenido alojado en sitios de terceros. El holding de activos digitales, incluyendo stablecoins y NFT, implican un alto grado de riesgo y pueden fluctuar en gran medida. Debes considerar cuidadosamente si el trading o holding de activos digitales es adecuado para ti a la luz de tu situación financiera.