Fine-tuned decisions for everyone

Bespoke intelligence, owned by you.

Upload a spreadsheet of examples with the expected outputs filled in. clef-tune fine-tunes a A decision model is a large language model, the kind of AI behind chatbots, trained to choose instead of chat: it reads a case and picks an answer from your list, with how sure it is. clef-tune tunes one on your examples. Learn more to predict them for new cases, and shows you how accurate it is before you use it.

Start with your data

Or try a sample: · ·

No sign-in needed to look around. When you're ready to train, you can sign in with Hugging Face.

How it works

01Your data

Upload examples from past work, or label some just for this.

02What to predict

Say what each output column means, as a plain question.

03Settings

Pick how much training. See the time and price first.

04Train

It runs on Hugging Face. Leave, and come back to the results.

Pay only for what runsOften under a dollar for a few hundred examples, billed by Hugging Face. You see the price before anything starts.

Your data stays yoursTraining data and the finished model are stored privately in your own Hugging Face account.

See before you trustEvery model is tested on examples it never saw, before and after training.

Fine-tunes Cloudflare's open Clef decision models · Open source on GitHub · MIT licence