摩尔线程MTT S5000完成对智谱最新大模型的适配

Core Viewpoint - Moore Threads has successfully completed the full-process adaptation and verification of the new generation large model GLM-5 on its flagship AI training and inference GPU MTT S5000, which is designed for large model training, inference, and high-performance computing [1] Group 1 - The MTT S5000 is a fully functional AI computing card based on the fourth-generation MUSA architecture "Pinghu" [1] - The MTT S5000 offers a maximum AI computing power of 1000 TFLOPS per card [1]