Seeking Access to Training Data for Research Purposes

Hey everyone,

I am a Master’s student from Concordia University, and I am currently conducting research using some open-source projects, including guildai/guildai.

I am writing to inquire if there might be a way for me to gain access to the data used to train your models across the different versions (let’s say, for a given commit or DVC hash). My understanding is that this might require access to the remote source from DVC, correct?

Furthermore, I am curious about the data sets that were used prior to the introduction of DVC. Would there be a chance to access this earlier data as well?

Any assistance you can offer would be incredibly valuable for my research, and I would be extremely grateful. Thanks for considering my request.

Hello and welcome!

We don’t use data sets that are not generally available. Any data sets that you run into associated with this project will be for samples. There are no trained models involved otherwise.

Hi @garrett . Thank you very much for the quick reply and the clarification!