Recently, China Telecom's Artificial Intelligence Research Institute (TeleAI) and Beijing Smart Source Artificial Intelligence Research Institute jointly released the world's first single dense trillion parameter semantic model Tele-FLM-1T, becoming the first institution in China to release a dense trillion-parameter large model. According to reports, in response to the high consumption of computing power required for large model training, TeleAI and Smart Source jointly developed the Tele-FLM series of models based on key technologies such as model growth and loss prediction, using only 9% of the computing resources of the industry's standard training plan, and completed the training of three models totaling 2.3T tokens based on 112 A800 servers in 4 months.
中国电信发布单体稠密万亿参数语义模型 用了普通训练方案9%的算力资源
China Telecom releases a monolithic dense trillion parameter semantic model that utilizes 9% of the computing resources of ordinary training schemes.
The translation is provided by third-party software.
The above content is for informational or educational purposes only and does not constitute any investment advice related to Futu. Although we strive to ensure the truthfulness, accuracy, and originality of all such content, we cannot guarantee it.