ORPO: Monolithic Preference Optimization without Reference Model Paper • 2403.07691 • Published Mar 12, 2024 • 70
Contrastive Preference Optimization: Pushing the Boundaries of LLM Performance in Machine Translation Paper • 2401.08417 • Published Jan 16, 2024 • 37
Running on CPU Upgrade Featured 2.57k The Smol Training Playbook 📚 2.57k The secrets to building world-class LLMs
Holo1.5 Collection Holo1.5 - Open Foundation Models for Computer Use Agents • 5 items • Updated Sep 15 • 34
Towards Reliable and Interpretable Document Question Answering via VLMs Paper • 2509.10129 • Published Sep 12
Running Featured 1.21k FineWeb: decanting the web for the finest text data at scale 🍷 1.21k Generate high-quality text data for LLMs using FineWeb