Use a small vision model with Delphi
Budget: $250 – $750 USD
Write an application with Delphi 12 (or compatible) that uses an open source vision model such as SmolVLM locally. By locally, I mean the application does not use any server or any remote API, instead the vision model exists only within the same folder in the user's computer as the application itself.
The application allows user to select an image e.g. via TOpenDialog and a prompt entry field, e.g. TEdit and a button TButton. When the button is clicked, the vision model is prompted with the user provided prompt and the result text is printed to the screen to a TMemo.
For example, this application could be asked to describe what is in the given picture, and the vision model responses what it sees there.
The application allows user to select an image e.g. via TOpenDialog and a prompt entry field, e.g. TEdit and a button TButton. When the button is clicked, the vision model is prompted with the user provided prompt and the result text is printed to the screen to a TMemo.
For example, this application could be asked to describe what is in the given picture, and the vision model responses what it sees there.