Fine-tuning Text Completions Model for Product Recommendation Use Case

[uncut turret]

I am trying to train a Text Completions model to use a product feed, which includes product names and features, to generate product recommendations based on keywords provided by real users as prompts. According to the API Documentation, the Text Completions model can be fine-tuned using "prompts" and "completions". The product feed includes the keywords that real users may use as prompts, such as "Recommend me a slim fit shirt of black color for around $10". I am seeking guidance on how to use the fine-tuned model to generate a list of product recommendations in response to such prompts.

[indigo kiln]

Key words.

Black, $10, slim, slim-fit, slim fitting, t-shirt, tshirt

The closer your users prompt and the more parameters to check, the warmer the response.

[uncut turret]

So while trying to fine-tune the text completions model, I have to give the product features/attributes as the prompt and the completion would be the product name?

Am I right?

[uncut turret]

@smoky bone/ @tough silo /@nimble salmon Sorry to bother you, but could you please help me with this?

[indigo kiln]

You got it

[sick drum]

🫡