Mingkai Zheng
1 indexed paper
Recent (6 mo)
1With code
0Influential cites
0Benchmarked
0Publications per year
126
Top categories
Distributed×1ML×1
Frequent co-authors
Research Timeline
2026
GIFT: Geometry-Informed Low-precision Gradient Communication for LLM Pretraining
The paper presents GIFT, a method for reducing communication volume in large language model pretraining by transforming gradients into a near-isotropic space before quantization.
Highlighted terms show continued research focus across papers