Huawei deploys a 30-billion-parameter MoE model on the Mate XT 2.

iconKuCoinFlash
Share
AI summary iconSummary
Huawei has deployed a 30B-parameter, 2B-activation MoE model on the Mate XT 2, built on MetaEra. This model, the first full-modal MoE on-device, runs on the Kirin 9050 Pro chip with Da Vinci NPU support. On-chain analysis reveals increasing interest in on-device AI performance. Announced in June, the model was optimized for Kirin chips using expert prediction techniques to reduce memory usage and enhance throughput. Integration into future Kirin chips and smartphones is planned for fall.
ME AI News, Dongcha Beating AI Brief: Huawei has officially deployed a 30B total parameter, 2B activated parameter edge-side MoE large model on its new triple-fold Mate XT 2 smartphone. Huawei describes it as the "industry's first edge-side MoE multimodal large model." The model runs on the newly released Kirin 9050 Pro, powered by Da Vinci architecture NPU for edge-side computing. This 30B/2B edge-side model was first previewed in June. At that time, Huawei stated it was specifically optimized for Kirin chips and reduces memory usage while improving inference throughput through techniques such as expert prediction, with plans to roll out on Kirin chips and smartphones this fall. (Source: BlockBeats)
Source:Show original
Disclaimer: The information on this page may have been obtained from third parties and does not necessarily reflect the views or opinions of KuCoin. This content is provided for general informational purposes only, without any representation or warranty of any kind, nor shall it be construed as financial or investment advice. KuCoin shall not be liable for any errors or omissions, or for any outcomes resulting from the use of this information. Investments in digital assets can be risky. Please carefully evaluate the risks of a product and your risk tolerance based on your own financial circumstances. For more information, please refer to our Terms of Use and Risk Disclosure.