Gemini Pro CLI

Use the Google Vertex API to send request to the Gemini Pro and Gemini Pro Vision LLMs

Pre-requisites

https://cloud.google.com/sdk/docs/install

https://docs.conda.io/projects/miniconda/en/latest/miniconda-install.html

Create a Conda Environment and install dependencies

.\install.bat

Run Gemini Pro CLI

conda activate gemini

python -m gemini_pro

Gemini Pro Vision

Add files to the /workspace folder and ask Gemini Pro to process them by filename.

Ask your question (or type 'exit' to quit):

What is the Year, Make, and, Model of the automobile in vehicle_1.jpg?

Extracted filenames: ['vehicle_1.jpg']
Calling ask_gemini_pro_vision

The year, make, and model of the automobile in vehicle_1.jpg is a 2021 Cadillac Escalade.

Ask your question (or type 'exit' to quit):

What is the subject of new_photo.jpg?

Extracted filenames: ['new_photo.jpg']
Calling ask_gemini_pro_vision

The subject of new_photo.jpg is Santa Claus.