Jilong Xue

A new compiler enhances AI chip communication and task management.

Explore how floating-point summation affects accuracy in computations.

How Mixture-of-Experts architecture boosts performance in language models.