Jackrong-llm-finetuning-guide

Why we included this project

Beginners adapting open-weight models will find this less like a tutorial and more like a course you can work through in a browser. It bundles runnable Colab and Kaggle notebooks covering supervised fine-tuning and reinforcement-learning methods such as GRPO and GSPO, along with data prep and distillation recipes. The dataset catalog and long-form guides walk you from raw data to a deployed GGUF checkpoint, so you are not stitching together scattered blog posts. There is also a dedicated pipeline for turning Qwen models into MTP-enabled GGUF files for local or agent use. Because it is organized by task, it gives developers a single self-paced path through the whole fine-tuning workflow.