README.md

metadata

title: NewsReX
emoji: 📰
colorFrom: blue
colorTo: purple
sdk: static
pinned: false

NewsReX - News Recommendation Model Zoo

A modular, multi-framework benchmark for neural news recommendation research. All models are implemented in JAX/Flax, PyTorch, and Keras 3, trained and evaluated under identical conditions.

Implemented Models

Model	Paper	Venue	JAX	PyTorch	Keras
NRMS	Wu et al.	EMNLP 2019	✅	✅	✅
NAML	Wu et al.	IJCAI 2019	✅	✅	✅
LSTUR	An et al.	ACL 2019	✅	✅	✅
MINER	Li et al.	ACL Findings 2022	✅	✅	✅
PP-Rec	Qi et al.	ACL 2021	✅	✅	✅
CAUM	Qi et al.	SIGIR 2022	✅	✅	✅
CROWN	Ko et al.	WWW 2025	✅	✅	✅
DIGAT	Mao et al.	EMNLP Findings 2022	✅	✅	✅
GLORY	Yang et al.	RecSys 2023	✅	✅	✅
TCCM	Qi et al.	CIKM 2023	✅	✅	✅

Quick Start

# Install
git clone https://github.com/igor17400/NewsReX.git
cd NewsReX
uv sync --extra all-cuda

# Train NRMS on MIND-small
uv run python src/train.py experiment=mind/nrms framework=jax

# Load pretrained weights from HuggingFace
uv run python src/train.py experiment=mind/nrms framework=jax \
    weights=hf://newsrex/NRMS-JAX-MIND-small/model.safetensors

Repository

GitHub: github.com/igor17400/NewsReX

Citation

If you use NewsReX in your research, please cite:

@misc{azevedo2025newsrexefficientapproachnews,
      title={NewsReX: A More Efficient Approach to News Recommendation with Keras 3 and JAX},
      author={Igor L. R. Azevedo and Toyotaro Suzumura and Yuichiro Yasui},
      year={2025},
      eprint={2508.21572},
      archivePrefix={arXiv},
      primaryClass={cs.IR},
      url={https://arxiv.org/abs/2508.21572},
}