One of the first large-scale Generative Chinese Pre-trained Language Models. 2.6B parameters; demonstrated strong performance across Chinese NLP tasks and laid the groundwork for the CPM series.

By Zhengyan Zhang, Xu Han, Hao Zhou et al. of the Department of Computer Science and Technology, Tsinghua University & BAAI. The CPM lineage is later stewarded externally by OpenBMB, which inherits Tsinghua THUNLP's open-source releases — including the descendant MiniCPM family.

Model Details

Parameters 2.6B

Paper

Citations 22
nlptrainingresearch

Related