deltakitsune/properly-v1.04

properly-E4-v1.04是一个用于文本生成任务的混合数据集，包含30,000行数据。该数据集由两个来源的数据混合而成：HF startc/synthetic-spelling（占比50.2%，目标15,060行）和HF grammarly/coedit（占比50.2%，目标15,060行，保留15,060行）。数据集语言为英语，标签包括kitsune-training-suite和curated，规模在10K到100K之间。properly-E4-v1.04 is a mixture dataset for text-generation tasks, containing 30,000 rows. The dataset is composed of data from two sources: HF startc/synthetic-spelling (50.2% mix, target 15,060 rows) and HF grammarly/coedit (50.2% mix, target 15,060 rows, kept 15,060 rows). The dataset is in English, with tags including kitsune-training-suite and curated, and its size category is between 10K and 100K.