Xiaowen Chu

A new framework improves pruning methods for large language models without retraining.

A new system optimizes training for large AI models by reducing communication costs.

A new method enhances language models for better performance and lower resource use.