EPEleventh Press International Publishing™An Enterprise Research Ecosystem

Search

Search articles, authors and pages, or jump to a dashboard section

Computer Science

Sparse Mixture-of-Experts Routing for Resource-Constrained Edge Inference

By Liu Chen, Priya Raghunathan, Sven Larsson

DOI: 10.52011/epip.2024.8096

Large AI language models are often too big and slow to run on small, low-power devices. This paper presents a new method for organizing how different parts of these models handle tasks, making them run more efficiently on affordable hardware like a Raspberry Pi while keeping accuracy nearly the same as full-sized models.

Download PDF

Chen, L., Raghunathan, P., Larsson, S. (2024). Sparse Mixture-of-Experts Routing for Resource-Constrained Edge Inference. Eleventh Press International Journal of Multidisciplinary Research, 4(2). https://doi.org/10.52011/epip.2024.8096