3 papers
cs.LG2025
Improving Model Representation and Reducing KV Cache via Skip Connections with First Value Heads
Zhoutong Wu, Yuan Zhang, Yiming Dong +4
Transformer models have driven breakthroughs across various language tasks by their strong capability to learn rich contextual representations. Scaling them to improve representati…
cs.LG2025
Conda: Column-Normalized Adam for Training Large Language Models Faster
Junjie Wang, Pan Zhou, Yiming Dong +6
Large language models (LLMs) have demonstrated impressive generalization and emergent capabilities, yet their pre-training remains computationally expensive and sensitive to optimi…
eess.IV2025
Exploring the Design Space of 3D MLLMs for CT Report Generation
Mohammed Baharoon, Jun Ma, Congyu Fang +2
Multimodal Large Language Models (MLLMs) have emerged as a promising way to automate Radiology Report Generation (RRG). In this work, we systematically investigate the design space…