LLaMA(Open and Efficient Foundation Language Models )论文解读(二)
此篇博客主题:LLAMA模型数据、训练时长、功耗及碳排放量LLaMA:OpenandEfficientFoundationLanguageModelspaperhttps://arxiv.org/pdf/2302.13971v1.pdf1训练样本Overall,ourentiretrainingdatasetcontainsroughly1.4Ttokensaftertokenization.Fo