1 comment

[ 5.8 ms ] story [ 12.4 ms ] thread
Introducing NVFP4 for Efficient and Accurate Low-Precision Inference