most citedMinerU2.5: A Decoupled Vision-Language Model for Efficient High-Resolution Document Parsing

2 citations · 2 across the 3 of their papers we have counts for

collaborators
Showing cs.CLShow all

1 paper · 1 filter

cs.CL2025

TWEO: Transformers Without Extreme Outliers Enables FP8 Training And Quantization For Dummies

Guang Liang, Jie Shao, Ningyuan Tang +2

Native FP8 support in modern hardware is essential for training large Transformers, but is severely hindered by extreme activation outliers. Existing solutions either rely on compl…