大家预训练或精调这个模型时，一般使用什么数据集呢？能分享下直接用于sft精调的如下格式的数据集吗？感谢！What datasets are generally used when pre-training or fine-tuning this model? Could you share a dataset in the following format that can be directly used for SFT fine-tuning? Thank you!

[ykk624]

{"messages": [sample1...], "format": "chatml"}
{"messages": [sample2...], "format": "chatml"}
{"messages": [sample3...], "format": "chatml"}