diff --git a/README.md b/README.md index 75cba1c..34eb829 100755 --- a/README.md +++ b/README.md @@ -330,6 +330,7 @@ | 多模态数据集MMC4 | 5.8亿图片,1亿文档,400亿token | [github](https://github.com/allenai/mmc4) | | EleutherAI 数据 | 800g的文本语料给你整合好了免费下载,不知道trian出来的model质量如何,打算试试: | [pile data](huggingface.co/datasets/EleutherAI/the_pile) [paper](http://t.cn/A6NqJ2Zl) | |UltraChat|大规模、信息丰富、多样化的多轮对话数据|[github](https://github.com/thunlp/UltraChat)| +|AI/ML API|AI/ML API提供300+AI模型,包括Deepseek,Gemini和ChatGPT。 这些型号以企业级速率限制和正常运行时间运行|[link](https://aimlapi.com/app/?utm_source=funNLP&utm_medium=github&utm_campaign=integration) | |ConvFinQA金融数据问答||[github](https://robustfin.github.io/2023/shared_task)| | The botbots dataset | 一个包含对话内容的数据集,对话内容来自于两个ChatGPT实例(gpt-3.5-turbo),CLT命令和对话提示来自GPT-4,覆盖多种情境和任务,生成成本约为35美元,可用于研究和训练更小的对话模型(如Alpaca) | [github](https://github.com/radi-cho/botbots) | | alpaca_chinese_dataset - 人工精调的中文对话数据集 | | [github](https://github.com/hikariming/alpaca_chinese_dataset) |