OWL-ViT (Open-World Localization Vision Transformer), developed by Google Research, is an open-vocabulary object detection model that integrates CLIP's vision-language pretraining with detection frameworks to detect objects described by arbitrary text without fine-tuning. It extends CLIP's image and text encoders into a detection architecture, aligning image regions with text descriptions via contrastive learning to generate bounding boxes and match scores. Built on Vision Transformers (ViT) for global feature extraction and lightweight detection heads, it supports zero-shot transfer to unseen categories (e.g., "purple unicorn toy" or "logo-covered backpack"), demonstrating strong generalization on open-world datasets like LVIS.

Source model

• Input shape: [[1,3,768,768]], [[1,16],[1,16]],[[1,24,24,768],[1,512],[1,16]]
• Number of parameters: 84.92M, 60.46M, --
• Model size: 339.91M, 242.06M, 1.51M
• Output shape: [[1,24,24,768],[1,576,4]], [[1,512]], [[1,576,1]]

The source model can be found here

When the user has fine-tuned the source model, the model conversion process must be performed again.

Users can refer to either of the following two methods to complete the model conversion:

• Using AIMO for model conversion: Click Model Conversion Reference in the Performance Reference section on the right to view the conversion steps.

• Using Qualcomm QNN for model conversion: Please refer to the Qualcomm QNN Documentation.

The model performance benchmarks and example code provided by Model Farm are all implemented based on the APLUX AidLite SDK.

For models in .bin format, you can use either of the following two inference engines to run inference on Qualcomm chips:

Inference using APLUX AidLite: please refer to the APLUX AidLite Developer Documentation

Inference using Qualcomm QNN: Please refer to the Qualcomm QNN Documentation

Inference Example Code

The inference example code is implemented using the AidLite SDK.

Click Model & Code to download the model files and the inference code package. The file structure is as follows:

/model_farm_{model_name}_aidlite
    
    |__ models # folder where model files are stored

    |__ python # aidlite python model inference example

    |__ cpp # aidlite cpp model inference example

    |__ README.md