5fa1a76
1
BLOOM was released around the same time, and the largest model in the family has 176B parameters and is trained on 366B tokens in 46 languages and 13 programming languages.
