Rugby Nutrition Planning Fine-Tuning Dataset (JSONL) — Mistral-7B Splits v1.0

This dataset accompanies an MSc dissertation at UCL on generative AI for real-time, personalised performance nutrition. It contains schema-standardised JSONL splits (train/validation/test) used to fine-tune and evaluate a Mistral-7B model for rugby-specific meal-plan generation. Records encode athlete/context inputs and structured plan outputs to support parameter-efficient fine-tuning (LoRA), schema-adherent generation, and practitioner-in-the-loop evaluation.