Zero-Shot Learning

We pride ourselves in requiring no training data for our models. Our proprietery models are already extensively trained on a wide variety of data (30 large datasets) that are able to leverage the recent advancements in LLMs to infer what an object is in an image.

Benefits of Zero-Shot

Instead of needing to fine-tune a model on your own data, you can utilize zero-shot and use it to detect custom objects without explicit training.

All you need to do is send an array of objects to detect and let ezML do the rest. Eg: ["Road", "Mountain", "Sign"]