Jackrong-llm-finetuning-guide
Jackrong LLM Fine-Tuning Guide is an educational, open-source knowledge base designed for beginners and developers seeking reproducible workflows in large language model development. It covers end-to-end processes including supervised fine-tuning, reinforcement learning with GRPO and GSPO methods, dataset distillation, and local model deployment. The resource supports accessing training recipes via Google Colab or Python scripts, with a focus on PyTorch and Unsloth frameworks. Key features include curated high-fidelity datasets for reasoning, coding, and conversation tasks, alongside detailed data preparation and filtering recipes. It provides specialized tools for converting Qwen models into agent-ready MTP-enabled GGUF formats, enabling efficient 16-bit exports and quantization. The repository offers multilingual documentation in English, Chinese, Korean, and Japanese, along with downloadable PDF guides and technical reports. Users can find quick-start tutorials for first-time fine-tuning in a browser, auto