3 papers
cs.CV2026
Lens: Rethinking Training Efficiency for Foundational Text-to-Image Models
Dong Chen, Fangyun Wei, Ziyu Wan +18
We introduce Lens, a 3.8B-parameter T2I model that achieves performance competitive with, and in several cases surpassing, state-of-the-art models with more than 6B parameters acro…
cs.AI2024
Minor DPO reject penalty to increase training robustness
Shiming Xie, Hong Chen, Fred Yu +3
Learning from human preference is a paradigm used in large-scale language model (LLM) fine-tuning step to better align pretrained LLM to human preference for downstream task. In th…
cs.AI2024
Minor SFT loss for LLM fine-tune to increase performance and reduce model deviation
Shiming Xie, Hong Chen, Fred Yu +2
Instruct LLM provide a paradigm used in large scale language model to align LLM to human preference. The paradigm contains supervised fine tuning and reinforce learning from human…