<div align="center">
# Jackrong LLM Fine-Tuning Guide
An educational, end-to-end open-source knowledge base for LLM fine-tuning,
dataset distillation, reinforcement learning, and local deployment.
🌐 **Languages:** English | [中文](docs/README_zh.md) | [한국어](docs/README_ko.md) | [日本語](docs/README_ja.md)
🤗 **Hugging Face:** [Jackrong](https://huggingface.co/Jackrong)
🌐 **Product Website:** [Explore Qwopus3.6 models and fine-tuning guides →](https://r6410418.github.io/Jackrong-llm-finetuning-guide/)
<br>
[](https://github.com/unslothai/unsloth)
[](https://colab.research.google.com/)
[](https://pytorch.org/)
[](https://huggingface.co/)
[](#)
[](#)
</div>
---
This repository is a growing educational resource portal for beginners and developers who want reproducible training pipelines, SFT and RL workflows including GRPO and GSPO, data preparation and distillation recipes, 16-bit export and GGUF deployment workflows, and agent-ready Qwen MTP GGUF conversion tools.
## 📚 Table of Contents
- [🚀 Start Here](#-start-here)
- [🗺️ Repository Map](#️-repository-map)
- [🏋️ Training Recipes](#️-training-recipes)
- [✅ Supported Workflows](#-supported-workflows)
- [🛣️ Model Support Roadmap](#️-model-support-roadmap)
- [⚙️ Qwen MTP GGUF Conversion Skill](#️-qwen-mtp-gguf-conversion-skill)
- [📘 Guides and Reports](#-guides-and-reports)
- [🧠 High-Fidelity Dataset Catalog](#-high-fidelity-dataset-catalog)
- [🤝 Open-Source Commitment](#-open-source-commitment)
- [📚 Citation](#-citation)
## 🚀 Start Here
| I want to... | Recommended entry |
|---|---|
| Fine-tune my first model in a browser | [Open the training recipe catalog](train_code/) |
| Run the Qwopus3.6 27B GSPO tutorial | [Open the GSPO Python tutorial](train_code/Qwopus3.6-27B-GSPO/qwopus3_6_27b_gspo_training.py) |
| Prepare or distill training data | [Browse data-processing recipes](data_processing_code/) |
| Find curated reasoning, coding, and conversation datasets | [Open the dataset catalog](High-fidelity%20Dataset/) |
| Convert a Qwen model to MTP-enabled GGUF | [Open the Qwen MTP GGUF Skill](qwen-mtp-gguf/) |
| Read full beginner guides and reports | [Open the PDF guide library](guidePDF/) |
| Automate repeatable Codex workflows | [Open the Codex Goal templates](codex-goals/) |
## 🗺️ Repository Map
| Resource | What you will find | Entry |
|---|---|---|
| 🏋️ Training Recipes | SFT, GRPO, and GSPO notebooks and Python tutorials | [Open](train_code/) |
| 🧪 Data Processing | Distillation, preprocessing, filtering, and sampling workflows | [Open](data_processing_code/) |
| 🧠 Dataset Catalog | Curated high-fidelity datasets and download helpers | [Open](High-fidelity%20Dataset/) |
| ⚙️ Qwen MTP GGUF Skill | Agent-ready MTP extraction, injection, conversion, validation, quantization, and upload pipeline | [Open](qwen-mtp-gguf/) |
| 📘 Guides and Reports | Long-form PDF tutorials and technical reports | [Open](guidePDF/) |
| 🌐 Multilingual Docs | Chinese, Korean, and Japanese landing pages plus documentation indexes | [Open](docs/) |
| 🤖 Codex Goal Templates | Editable goal templates for RL training, MTP GGUF conversion, and repository maintenance | [Open](codex-goals/) |
## 🏋️ Training Recipes
| Model | Method | Environment | Quick setup |
|---|---|---|---|
| Qwopus3.5 27B | SFT | Google Colab | [](https://colab.research.google.com/github/R6410418/Jackrong-llm-finetuning-guide/blob/main/train_code/Qwopus3-5-27b-Colab.ipynb) |
| Qwopus3.6 27B | GSPO | Python script | [](train_code/Qwopus3.6-27B-GSPO/qwopus3_6_27b_gspo_training.py) |
| Qwen3.5 9B | SFT | Kaggle | [](https://kaggle.com/kernels/welcome?src=https://github.com/R6410418/Jackrong-llm-finetuning-guide/blob/main/train_code/Qwen3.5-9B-Neo-Kaggle.ipynb) |
| Qwopus3.5 35B | SFT | Kaggle | [](https://kaggle.com/kernels/welcome?src=https://github.com/R6410418/Jackrong-llm-finetuning-guide/blob/main/train_code/Qwopus-3.5-35B-A3B-Kaggle.ipynb) |
| Llama3.2-R1 3B | GRPO | Kaggle | [](https://kaggle.com/kernels/welcome?src=https://github.com/R6410418/Jackrong-llm-finetuning-guide/blob/main/train_code/Llama-3.2-3B-R1-Zero-GRPO.ipynb) |
Browse the full catalog in [train_code/README.md](train_code/README.md).
## ✅ Supported Workflows
| Workflow | Status | Documentation |
|---|---|---|
| SFT with LoRA / QLoRA | ✅ Released | [Training recipes](train_code/) |
| GRPO reinforcement learning | ✅ Released | [Training recipes](train_code/) |
| GSPO reinforcement learning | ✅ Released | [Qwopus3.6 27B GSPO tutorial](train_code/Qwopus3.6-27B-GSPO/qwopus3_6_27b_gspo_training.py) |
| Dataset distillation and preprocessing | ✅ Released | [Data-processing recipes](data_processing_code/) |
| LoRA adapter save and merged 16-bit export | ✅ Released | [Training recipes](train_code/) |
| GGUF quantization | ✅ Released | [Training recipes](train_code/) |
| Qwen MTP GGUF conversion | ✅ Released | [MTP conversion skill](qwen-mtp-gguf/) |
## 🛣️ Model Support Roadmap
Released RL recipes may use GRPO or GSPO depending on the model and training objective.
| Model Family | SFT Support | RL Support |
|---|---:|---:|
| Qwen 3.5 | ✅ Released | Scheduled |
| Qwen 3.6 | ✅ Released | ✅ Released |
| Qwen 3 | Scheduled | Scheduled |
| Llama3.2-R1 3B | ✅ Included | ✅ Released |
| Llama 3.1 / 3.3 | Scheduled | Scheduled |
## ⚙️ Qwen MTP GGUF Conversion Skill
The [`qwen-mtp-gguf`](qwen-mtp-gguf/) subproject supports Qwen-family MTP / nextn GGUF release workflows. It performs disk, RAM, tooling, token-access, and compatibility preflight checks, extracts compatible MTP tensors, injects them into the target model, converts with llama.cpp, smoke-tests outputs, quantizes releases, and supports safer upload/resume workflows.
[🚀 Open the MTP Skill](qwen-mtp-gguf/) · [📖 Read the Pipeline Guide](qwen-mtp-gguf/docs/Qwen-MTP-GGUF-Pipeline-Guide.md) · [🤖 Read the Agent Usage Guide](qwen-mtp-gguf/docs/Qwen-MTP-GGUF-Agent-Usage.md)
## 📘 Guides and Reports
Long-form PDFs live in the [guide and technical report library](guidePDF/README.md).
| Guide | Topic | File |
|---|---|---|
| Qwopus3.5 27B Colab complete guide | Beginner-friendly end-to-end fine-tuning walkthrough | [PDF](guidePDF/Qwopus3-5-27b-Colab_complete_guide_to_llm_finetuning.pdf) |
| Qwopus GLM 18B technical report | Model design and training notes | [PDF](guidePDF/Qwopus-GLM-18B-Technical-Report.pdf) |
## 🧠 High-Fidelity Dataset Catalog
The repository includes 24 curated high-fidelity datasets for reasoning, mathematics, coding, instruction following, conversation, and domain-specific distillation. Browse the full [dataset catalog](High-fidelity%20Dataset/README.md), or use [`download_datasets.py`](download_datasets.py) to batch download the suite for local training.
## 🤝 Open-Source Commitment
This project keeps the training source code and documentation for released fine-tuned models available so learners can reproduce, inspect, and adapt the workflows. The longer project philosophy and original message to builders are preserved in [docs/PROJECT_PHILOSOPHY.md](docs/PROJECT_PHILOSOPHY.md).
## 📚 Citation
If you find this repository helpful in your learning or research, please consider citing it:
```bibtex
@misc{jackrong-llm-finetuning,
author = {Jackrong},
title = {Jackrong LLM Fine-Tuning Guide: An Educational LLM Fine-Tuning Knowledge Base},
year = {2026},
publisher = {GitHub},
journal = {GitHub repository},
howpublished = {\url{https://github.com/R6410418/Jackrong-llm-finetuning-guide}}
}
```