How do you perform fine-tuning in AiAssistWorks?

Fine-tuning an AI model using AiAssistWorks is straightforward. Simply select your input and output pairs, choose a model, and begin fine-tuning. Here's a breakdown:

1. Open the "Fine Tune AI a Model" menu from a sheet where you have an input and output column. 
2. The range is automatically detected. 

3. Construct an input message using any column value, utilizing colX (e.g., col1, col2) where X is position x position in input range, for example col1 is the first column of the input range, etc. For example, "Translate 'col1' to 'col2' language."

4. Select the output column, which should be a single column like B.
5. Optionally, adjust hyperparameters such as batch size, learning rate, or epoch. 
6. Choose an AI model. Currently, we only support Gemini since it's free and the fine-tuned model costs the same as a regular model in Gemini.
7. Name your fine-tuned model, then click "Preview and Start."
8. Preview the input and output pairs to ensure they meet your requirements. 
9. Scroll to the bottom and click "Fine Tune Model."

10. Your fine-tuned model will be in a "creating" status. Please wait for some time, as it can take up to an hour. Check the status in the "List Tuned Model" menu.

11. You can use your fine-tuned model anywhere in AiAssistWorks or with your Google API.