"Model with 180 Billion Parameters is Trained on 3.5 Trillion Tokens, with 4 times the Compute Resources of Meta’s LLaMA 2"だって。やはりChatGPT一強よりも、Hugging Faceみたいな場で強豪群雄割拠の方が今はイメージが湧く www.businesswire.com M…