⚙️ Models

​

Overview

​

Bounding Box Detection Models

Overeasy supports a variety of models that specialize in bounding box object detection. In Overeasy’s case these Bounding Box models are Open Vocabulary Detection Models, which means they draw bounding boxes based on their semantic understanding of the labels. This allows you to create very powerful zero shot detection workflows.

​

Embedding Models

Overeasy supports embeddings models such as CLIP, which can be used for tasks like Classification and Retrieval.

​

Multimodal LLMs

Overeasy supports multimodal LLM capable of handling both text and visual inputs, which are useful for tasks that require understanding and generating content based on combined input types.

​

LLMs

Overeasy supports standard LLMs that can handle text inputs only, which are still useful in many Overeasy workflows.