3 ms·
Curie can roughly match davinci quality when fine tuned. Fine tuning's cost is mostly on initial training data. After that, it's half the price as davinci on p
by muzani 4y ago
Curie can roughly match davinci quality when fine tuned. Fine tuning's cost is mostly on initial training data.
After that, it's half the price as davinci on paper, but you need only a fraction of the tokens. They've said they plan on making it cheaper in the future too.
- moomoo11 4y agoCan you give me some pointers to any articles or discussions about how to do this so it works? Thanks!
- muzani 4y agoI did some samples for fine tuning in this repo: https://github.com/smuzani/openai-samples https://github.com/smuzani/openai-samples The sample is in /ft/ folder, instructions at bottom of Readme. Feel free to drop an issue if it's hard to understand. I copy-pasted the instructions in 3 min so it's missing a few steps.
- moomoo11 4y agoThanks!