sanjaypantdsd/

fine-tuning-socratic-dataset

A dataset of 100 conversation pairs teaching fine-tuning concepts through Socratic questioning.

Size: 100 conversations Format: Chat format (system, user, assistant) Method: Socratic questioning - guides learning through questions rather than direct answers Topics: Fine-tuning, PEFT methods (LoRA, QLoRA), data quality, troubleshooting

from datasets import load_dataset

dataset =… See the full description on the dataset page: https://huggingface.co/datasets/sanjaypantdsd/fine-tuning-socratic-dataset.