#60 Faísca: The modern LLM stack in a single script
Update: 2025-10-14
Description

- Why do this?
- A small dataset of news headlines
- GPT2 in PyTorch
- Pre-training headlines in Portuguese
- Supervised fine-tuning (SFT) on Portuguese from Portugal
- Reinforcement Learning (GRPO) for sports news
- Final thoughts & Acknowledgements
Why do this?
ML and AI are moving at an incredible pace. The amount of research coming out …
Comments
In Channel



