AI Mole

Training

  • Preparing training datasets for large language models
    Published on
    Machine learning large language models on custom datasets makes the models much smarter. There is a success story of the Alpaca dataset. It works wonders with models that at first, if they could do something, they did it very badly. We set out to understand how it iss done, and more importantly, what the challenges are along the way and whether newbies like us can figure it out.