A team from the StatNLP group at the Singapore University of Technology and Design (SUTD) published a paper on arXiv on the 5th of May, proposing a 1.1 billion-parameter small-scale language model, TinyLlama, which has been pre-trained in three rounds on about 1 trillion tokens.
SUTD team proposes 1.1 billion parameter miniatures TinyLlama
Previous: 国泰君安：GPT商店即将上线 AI应用迎来爆发
Next: 小度推出小度学习机K16 具备20多项AI能力