GLM-5.3-Flash-REAP50-FP8

Papers using or citing this model 1

• REAP the Experts: Why Pruning Prevails for One-Shot MoE compression 56 citations · 2025 · arXiv (Cornell University) · arXiv

  Sparsely-activated Mixture-of-Experts (SMoE) models offer efficient pre-training and low latency but their large parameter counts create significant memory overhead, motivating research into expert compression. Contrary to recent...