LAION-AI
Incorrect url in SFT prompt_dialogue dataset
As mentioned in #1368 the url to the prompt_dialogue dataset is broken. AFAICT the new url is here so we just need to change the link and check we can load the data using the new link(s).
Hi! I am new here and would like to contribute, can I be assigned this issue?
Hi, the links for the datasets were not only wrong but the datasets had been deleted. I have retrieved them through the commit history and I would like to upload them as datasets to the Open-Assistant organisation on HuggingFace to avoid this issue in future. Please can someone approve my request to join on HuggingFace?
Hi, the links for the datasets were not only wrong but the datasets had been deleted. I have retrieved them through the commit history and I would like to upload them as datasets to the Open-Assistant organisation on HuggingFace to avoid this issue in future. Please can someone approve my request to join on HuggingFace?
I am not sure who is in charge of the HF org. Maybe one of the ML leads @sanagno @theblackcat102?
The incorrect url datasets will be removed in the PR#1793
So we don't need these datasets at all any more and we can close this issue?
Rallio moved his data here: https://github.com/LAION-AI/Open-Instruction-Generalist/tree/main/small_instruction_set
Now it's an official LAION-AI dataset. We can pull from here as you wish and have his blessing as well :) @theblackcat102
The incorrect url datasets will be removed in the PR#1793 We can close this issue because the code containing the urls is removed in the above PR.