The Digital Sous Chef -- A Comparative Study on Fine-Tuning Language Models for Recipe Generation

TL;DR

This paper establishes a benchmark for recipe generation, demonstrating that fine-tuning large language models with specialized tokenization significantly improves performance over smaller models and traditional methods.

Contribution

It introduces a novel tokenization strategy tailored for recipe text, enabling better preservation of domain-specific structures and quantities in large transformer models.

Findings

Large GPT-2 model outperforms smaller models and RNN baselines in recipe generation.

Specialized tokenization improves model understanding of recipe structures and numerical data.

Model achieves over 20% relative improvement in semantic relevance metrics.

Abstract

We established a rigorous benchmark for text-based recipe generation, a fundamental task in natural language generation. We present a comprehensive comparative study contrasting a fine-tuned GPT-2 large (774M) model against the GPT-2 small (124M) model and traditional LSTM/RNN baselines on the 5-cuisine corpus from RecipeDB. Our key contribution is a targeted tokenization strategy that augments the vocabulary with 23 common fraction tokens and custom structural markers. This approach addresses a critical limitation of generic tokenizers by preserving essential recipe structures and precise numerical quantities, thereby enhancing domain specificity. Performance is evaluated using a comprehensive suite of seven automatic metrics spanning fluency (BLEU-4, METEOR), coherence (ROUGE-L), semantic relevance (BERTScore), and diversity. Our experiments show that the large transformer-based approach yields a >20% relative improvement in BERTScore (F1) (0.92 vs 0.72) over the best recurrent baseline, while reducing perplexity by 69.8%. We conclude with a discussion of remaining challenges, particularly regarding factual accuracy, and outline how this foundational study paves the way for integrating real-world constraints and multi-modal inputs in advanced recipe generation research.