How to Ensure Consistent Output with Gemini API for Handwritten Text Recognition?

I’m using the Gemini API to extract medicine names from handwritten prescriptions, but I’m encountering inconsistent results each time I run my program. The output varies despite using the same input image, which makes it challenging to rely on for downstream processing. I’ve tried adjusting the temperature parameter to a lower value (e.g., 0.1), but the issue persists.

Hey @Syeed_Talha , You can use the seed parameter to reproduce the same output. Please refer the gist file below for details.

Code

Thanks..

Thanks for you kind information…