Search papers, labs, and topics across Lattice.
University of Science and Technology
2
0
4
Dynamic MoE load balancing can run at every single microbatch with zero CPU synchronization by having GPUs deterministically solve bitwise-identical expert replication plans locally, boosting cluster throughput by up to 11.4%.
Orthonormal initialization can transform low-rank adaptation in RLVR, outperforming established methods and stabilizing training dynamics.
