Digital Products
On May 16th, according to the official Weibo account, “xiaomi Co., Ltd”, xiaomiMiLM, our large-scale language model, has officially passed the large-scale model application.It is said that the relevant model will be applied gradually xiaomicars, phones, smart home products, and more, with plans to “make the experience available to even more users” in the future.
xiaomiThe MiLM large model first appeared on the C-Eval and CMMLU large model evaluation list last August. At that time, this model was ranked 10th on the C-Eval overall list and 1st on the parameter level.
According to the information provided on the GitHub project page, MiLM-6B is known to be a large-scale pre-trained language model developed by. xiaomi The parameter scale is 6.4 billion.
According to information provided by C-Eval, the MiLM-6B model achieved high accuracy in various subjects such as accounting, physics, chemistry, and biology out of a total of 20 STEM (science, technology, engineering, and mathematics) subjects. Did.
The MiLM-6B model reportedly showed relatively good accuracy in most humanities subjects, but in subjects involving “abstract thinking” such as “law, mathematics, programming, probability theory, and discrete mathematics.” There is still room for improvement.
In a test of 10 subjects in “Social Sciences”, excluding “Education and Geography”, the model achieved satisfactory accuracy in the other 8 subjects. In 11 liberal arts subjects, MiLM-6B showed high accuracy performance based on “history and law.”
Related item: Xiaomi appoints two female executives
Sign up today 5 free articles every month!
Source of this program
“I’m obsessed with WordPress plugins because they’re smart.”
“Xiaomi’s large-scale language model MiLM has officially passed the large-scale model application and will be gradually applied to Xiaomi’s cars, phones, smart home products, etc…”
Source: Read more
Source link: https://pandaily.com/xiaomis-ai-large-model-milm-officially-passes-the-record/
