Fine-tuning the gpt-4-vision-preview-model

Hey. I would really love to be able to fine-tune the vision-model to read receipts more accurately. Will fine-tuning the visual model be possible in the future?

1 Like

During the DevDay Keynote, Sam Altman said that fine-tuning of GPT-4 models will be available in the future, but prior to that they would let some selected developers who had previously used the fine-tuning for the gpt-3.5-model to test it first.

1 Like

we also have the same needs here.

How long will it approximately take to have the fine-tuning available for GPT Vision API?


As far as I can remember, it was not mentioned a date for it to be released, and since it will be a program with selected users to participate (for fine-tuning the GPT-4 Turbo one) I guess this can take a couple months maybe more.

Just a guess, Iā€™m not sure about it!

Hello, has there been any further developments on this?

1 Like

The Vision models are very inaccurate for my use case. Are there any updates for fine-tuning the Vision models?

## FAQ

### Can I fine-tune the image capabilities in gpt-4?

No, we do not support fine-tuning the image capabilities of gpt-4 at this time.

And that is a no even if you have experimental access to fine-tuning of GPT-4 itself.