TESMR

The code repository for the paper, "From Raw Features to Effective Embeddings: A Three-Stage Approach for Multimodal Recipe Recommendation" (https://arxiv.org/abs/2511.19176)

Datasets

We use two widely used datasets for recipe recommendation, Allrecipes and Food.com. For both dataset, please follow preprocessing steps in the 'preprocessing_and_analysis' directory.

---

How to run TESMR

• Installation

pip install -r requirements.txt

• Allrecipes

python main.py -m TESMR -d allrecipes

• Food.com

python main.py -m TESMR -d foodcom

---

Acknowledgement

This code is implemented based on two open source codes

• https://github.com/enoche/MMRec
• https://github.com/sdu-zyx/Multi-modal-Food-Recommendation