Is it possible to fine-tune without programming knowledge?

API
1

Hello,

I have a few questions about using fine-tuning to train models for specific tasks. First, if I don’t have programming skills (I know a bit Processing language but my knowledge is very limited), can I still use fine-tuning to train a model for a very specific image-related task, such as creating images in a specific niche?

Can I also use fine-tuning to train a language model for a specific task, such as only creating stories for children? Is it possible to do this without programming skills? If so, where can I find information on how to do this for a less technical person? Could anyone direct me a bit, please?

Also, I saw a video where someone used fine-tuning to train a model that specialized in a specific type of text similar to someone’s style using GPT-3. I he had to pay $58 on OpenAI for this experiment. Is this a standard pricing or is there a cheaper option available (I know it is hard to answer because there is no details of his project)? Where is the cheapest place to train my own models? Are there any options other than OpenAI for training language models not for business purpose?

I apologize for the chaos in my questions, but I am new to this and trying to find my footing. Being a humanities person, it is difficult for me to navigate this technical landscape.

2

Hey! I suggest checking out the fine tuning guide, it walks you through everything you need: OpenAI API

As far as the costs, it entirely depends on how much data you use to fine tune and how long the model trains. It is entirely possible that it would be less than $58 but note that it is not possible to fine tune DALL-E through our API (image model). You can only fine tune our text models.

3

Thanks for this, it’s an important link for me. It does not look that complex as I thought. And it looks like a good point to start.