Các nguồn dữ liệu (GPT4, ShareGPT, Dolly 2.0 ...)

[tiendung]

Kho dữ liệu huấn luyện chỉ dẫn, QnA, hội thoại ...

Dữ liệu nổi bật

• SharedGPT
• https://github.com/Instruction-Tuning-with-GPT-4/GPT-4-LLM
• https://github.com/teknium1/GPTeacher
• https://github.com/project-baize/baize-chatbot/tree/main/data dữ liệu multiple dialog (đã dịch quora)
• https://github.com/lightaime/camel#data-hosted-on-hugging-face sinh bởi gpt-4
• https://github.com/anthropics/hh-rlhf
• https://huggingface.co/datasets/balisujohn/RWKV-oasst1

Khác

• https://huggingface.co/datasets/samsum
• https://github.com/XueFuzhao/InstructionWild
• https://github.com/google-research-datasets/presto
• https://huggingface.co/datasets/JosephusCheung/GuanacoDataset/raw/main/guanaco_chat_all-utf8.json
• https://huggingface.co/datasets/blended_skill_talk
• https://huggingface.co/datasets/laion/community-chat-contributions
• https://github.com/radi-cho/botbots
• https://github.com/curai/curai-research/tree/main/DERA/medqa_open
• https://github.com/BenjiKCF/awesome-language-dataset

[tiendung]

lm-sys/FastChat#90 (comment)

ShareGPT Dataset:

Zipped jsons with 90 000 conversations from sharegpt. Split in two files with 45k each:
part 1: https://files.catbox.moe/bhtp9i.zip
part 2: https://files.catbox.moe/ahoivx.zip

Format should work as is for training. Use clean tool to remove html markup: https://github.com/lm-sys/FastChat/blob/main/docs/commands/data_cleaning.md

(Note: I'm just relaying this info from someone who sent it my way. So I don't know anything more than anyone else.)

The entire pre-cleaned 90k conversation dataset is also available here: https://huggingface.co/datasets/anon8231489123/ShareGPT_Vicuna_unfiltered/tree/main/HTML_cleaned_raw_dataset

A pre-cleaned, English only, "unfiltered," and 2048 token split version of the ShareGPT dataset ready for finetuning is available here: https://huggingface.co/datasets/anon8231489123/ShareGPT_Vicuna_unfiltered

[trinhdoduyhungss]

ViQuAD + Alpaca (Vietnamese) + daily_conversation (Vietnamese) + GPT4ALL in Alpaca format (400K+):
v1: https://drive.google.com/file/d/1F121M9f2LNy6RgXWlFBSSm626LWzfAvR/view?usp=sharing
v2: https://drive.google.com/file/d/1bE0B3Q86uG26A540acrysF-La07E29f4/view?usp=share_link

[tiendung]

ViQuAD + Alpaca (Vietnamese) + daily_conversation (Vietnamese) + GPT4ALL in Alpaca format (400K+): v1: https://drive.google.com/file/d/1F121M9f2LNy6RgXWlFBSSm626LWzfAvR/view?usp=sharing v2: https://drive.google.com/file/d/1bE0B3Q86uG26A540acrysF-La07E29f4/view?usp=share_link

Nice, em có thể cung cấp dữ liệu dưới dạng nguồn riêng lẻ để mix & match cho dễ đc ko?

[trinhdoduyhungss]

ViQuAD + Alpaca (Vietnamese) + daily_conversation (Vietnamese) + GPT4ALL in Alpaca format (400K+): v1: https://drive.google.com/file/d/1F121M9f2LNy6RgXWlFBSSm626LWzfAvR/view?usp=sharing v2: https://drive.google.com/file/d/1bE0B3Q86uG26A540acrysF-La07E29f4/view?usp=share_link

Nice, em có thể cung cấp dữ liệu dưới dạng nguồn riêng lẻ để mix & match cho dễ đc ko?

Dạ tại đây ạ: https://drive.google.com/drive/folders/156yLw2lZHMu6W4rnEIMNhXx5BLqRz_Yu?usp=sharing

[tiendung]

https://www.chatorg.ai/blog/chat-language-models-tracker