对于关注国产 AI 基础设施和低资源部署的开发者，BitCPM-CANN 展示了在昇腾 NPU 上训练三元模型的完整路径，可以直接参考其开源训练脚本做本地化部署或模型优化。
由 ModelBest、清华大学和 OpenBMB 社区联合开发的 BitCPM-CANN 成为全球首个完全基于中文 AI 基础设施（华为昇腾 910B NPU）训练的开源 1.58-bit 三元大语言模型。该模型采用三元权重（仅三种状态），大幅降低部署内存需求，适合手机、PC、汽车等本地设备。其训练系统在昇腾 910B 上实现了量化感知训练（QAT）、直通估计器（STE）等全套流程，并开源了可复现的训练脚本。这一成果展示了在硬件成本上升背景下，模型能否在真实约束下训练、复现、部署和优化比单纯追求榜单分数更重要。
BitCPM-CANN just became the world’s first open-sou…
BitCPM-CANN just became the world’s first open-sourced 1.58-bit ternary LLM trained entirely on Chinese-developed AI infrastructure.
Developed by ModelBest, Tsinghua Univ, and OpenBMB community, the entire training pipeline, from quantization operators and algorithms to the full-stack framework, was natively executed on Huawei Ascend 910B NPUs.
1.58-bit ternary weights use only 3 weight states, so the model needs far less memory when deployed on phones, PCs, cars, and local industrial devices.
The harder achievement is the training system behind it: QAT, STE, low-bit operators, algorithms, framework work, and reproducible training scripts all had to hold together on Ascend 910B.
When hardware costs rise, the winning model is not merely the one that scores higher in a chart, but the one that can be trained, reproduced, deployed, and improved under real constraints.