Below is fine-tuning Llama3.1 ran locally using Ollama on text2sql dataset from HuggingFace. Using Unsloth library and PEFT technique to fine-tune. Can run on Colab.

https://www.youtube.com/watch?v=pxhkDaKzBaY

Below is fine-tuning Florence-2 multimodal model to do object detection

https://www.youtube.com/watch?v=i3KjYgxNH6w

Below is fine-tuning GPT-4o multimodal model to do object detection

https://www.youtube.com/watch?v=6Q6TieCBA4E

Quick Unsloth + Ollama

https://www.youtube.com/watch?v=UWF6dxQYcbU