Could you please tell me what dataset was used for quantification?

#2
by LiuTTong - opened

The default dataset used seems to be "mit-han-lab/pile-val-backup", but I'm not sure if it is suitable for Chinese.

Sign up or log in to comment