Geti Instant Learn

GitHub Documentation

A flexible and modular framework for exploring, developing, and evaluating visual prompting algorithms.

Visual prompting offers a powerful alternative to the typical learning process. Instead of training a dedicated model based on huge datasets or manually segmenting and labeling thousands of images, you can feed the model with just a few examples of what you’re looking for. The model can find and segment similar objects in additional images.

Geti Instant Learn offers:

• Rapid prototyping of vision AI

• Scenarios with limited labeled data

• Edge deployment (CPU, GPU, NPU)

• Real‑time inference without retraining