diff --git a/data/dataset_info.json b/data/dataset_info.json index 1d226b3a..e321196a 100644 --- a/data/dataset_info.json +++ b/data/dataset_info.json @@ -524,13 +524,13 @@ "prompt": "text" } }, - "fileweb": { + "fineweb": { "hf_hub_url": "HuggingFaceFW/fineweb", "columns": { "prompt": "text" } }, - "fileweb_edu": { + "fineweb_edu": { "hf_hub_url": "HuggingFaceFW/fineweb-edu", "columns": { "prompt": "text" @@ -550,5 +550,25 @@ "prompt": "content" }, "folder": "python" + }, + "Magpie-Pro-300K-Filtered": { + "hf_hub_url": "Magpie-Align/Magpie-Pro-300K-Filtered", + "columns": { + "messages": "conversations" + }, + "tags": { + "role_tag": "from", + "content_tag": "value", + "user_tag": "human", + "assistant_tag": "gpt" + }, + "formatting": "sharegpt" + }, + "WebInstructSub": { + "hf_hub_url": "TIGER-Lab/WebInstructSub", + "columns": { + "prompt": "question", + "response": "answer" + } } -} \ No newline at end of file +}