haroldng/aya-collection-subset
收藏官方服务：
资源简介：
Aya-collection-subset数据集是从Aya Collection中选取的prompt-completion对子集，用于Aya-101模型的训练，采用Min-K% Prob方法识别。数据以Parquet文件格式存储，并按配置分目录存放。
This dataset is a subset of prompt-completion pairs from the Aya Collection, identified as likely seen during the training of the Aya-101 model using the Min-K% Prob method. The data is stored in Parquet files and organized in subdirectories according to configurations.
提供机构：
haroldng
