Together Fine-Tuning Platform Adds Preference Optimization and Continued Training

AI Impact Summary

Together AI has significantly enhanced its Fine-Tuning Platform with Preference Optimization and Continued Training capabilities. This allows businesses to dynamically adapt their language models based on user preferences and evolving data, moving beyond static model deployments. The introduction of Direct Preference Optimization (DPO) and the ability to resume training from checkpoints dramatically simplifies the fine-tuning process and enables continuous model improvement, a key factor for applications like Protege AI’s personalized compliance solutions.