In 2023, domestic large models will develop rapidly and make significant progress in computing power, data and algorithms. Although there are limitations in computing power, the ban has promoted the rise of domestic chips, and public cloud manufacturers have also concentrated computing resources to meet market demand. However, the scarcity of high-quality Chinese data sets has become a new challenge, and future development may focus on the mining and utilization of high-quality data.
In 2023, domestic large models have made great progress in the troika of computing power, data, and algorithms. Although computing power is restricted and the ban has promoted the rise of domestic chips, public cloud manufacturers have solved the demand by concentrating computing power. However, data quality may become a new problem. Chinese data sets are scarce, and mining high-quality data may become the focus of future development.
All in all, the development momentum of domestic large models is strong, but the problem of data quality cannot be ignored. In the future, improving data quality and building high-quality Chinese data sets will be the key to the development of domestic large models.