NVIDIA-AI-IOT/nanoowl
A TensorRT optimization toolkit for OWL-ViT that enables real-time zero-shot object detection on NVIDIA Jetson Orin platforms.
Not currently ranked — collecting fresh signals.
star history
NanoOWL optimizes the OWL-ViT vision-language model for deployment on NVIDIA Jetson edge devices using TensorRT. It converts the model to optimized TensorRT engines to achieve real-time performance. The project also introduces a tree detection pipeline that combines OWL-ViT and CLIP models for hierarchical open-vocabulary object detection and classification.
Frequently asked
- What is NVIDIA-AI-IOT/nanoowl?
- A TensorRT optimization toolkit for OWL-ViT that enables real-time zero-shot object detection on NVIDIA Jetson Orin platforms.
- Is nanoowl open source?
- Yes — NVIDIA-AI-IOT/nanoowl is open source, released under the Apache-2.0 license.
- What language is nanoowl written in?
- NVIDIA-AI-IOT/nanoowl is primarily written in Python.
- How popular is nanoowl?
- NVIDIA-AI-IOT/nanoowl has 438 stars on GitHub.
- Where can I find nanoowl?
- NVIDIA-AI-IOT/nanoowl is on GitHub at https://github.com/NVIDIA-AI-IOT/nanoowl.