QWEN-3: EASIEST WAY TO FINE-TUNE WITH REASONING 🙌

5 May 2025 · 20 min
19:32

German-language tutorial showing how to fine-tune QUIN family models (Qwen/Quin) using LoRA adapters and a mix of reasoning (chain-of-thought) and non-reasoning instruction data. Covers the rationale for using adapters (avoid catastrophic forgetting), dataset preparation and prompt templates (special think/chain-of-thought tokens), quantization and memory-saving tools/notebooks, SFT trainer setup and example hyperparameters, and example inference settings to toggle hybrid 'think' vs 'non-think' behavior. Demonstrates a small-scale Colab workflow (4-bit loading, modest VRAM ~12 GB, short run ~30 steps) and discusses saving/loading only LoRA adapters for edge deployment.

Questions this video answers

248 tokens

4-bit

~12 GB

Maximum ~30 steps

~36%

Chapters

More from LARRUZ

8:49
youtube.com

OpenAI’s New AI Is Doing Things We Barely Understand

3d ago
5:38
youtube.com

DeepSeek Solved AI’s Memory Problem

5d ago
33:17
youtube.com

DeepMind's New AI Just Cracked The Code Of Life

1w ago
3:43
youtube.com

The Billion Dollar AI Advantage Is Disappearing

1w ago
5:14
youtube.com

Claude Opus 5.5 AI: An Incredible Leap Forward

3w ago
5:16
youtube.com

Yes, Jev Is Insane, But There's A Catch

3w ago
5:11
youtube.com

DeepSeek’s Insane New Architecture

4w ago
4:16
youtube.com

Claude Is Now Leaving Invisible Fingerprints In Its Text

1mo ago
9:18
youtube.com

I Never Thought I’d See This Happen

1mo ago
5:22
youtube.com

GPT-6 Astra Changes Everything

1mo ago
4:55
youtube.com

Claude Fable AI Is Much Stranger Than The Headlines Suggest

1mo ago
See all content