📖Technical Report
smoltalk-chinese is a Chinese fine-tuning dataset constructed with reference to the SmolTalk dataset. It aims to provide high-quality synthetic data support for training large language models (LLMs). The dataset consists entirely of synthetic data, comprising over 700,000 entries. It is specifically designed to enhance the performance of Chinese… See the full description on the dataset page:
https://huggingface.co/datasets/opencsg/smoltalk-chinese.
