NXP eIQ Model Zoo

The NXP eIQ Model Zoo is a collection of optimized deep learning models for embedded vision applications. Developed during my time at NXP Semiconductors, this project provides 30+ pre-trained and optimized models for various computer vision tasks including:

• Object detection
• Image classification
• Semantic segmentation
• Pose estimation
• Super resolution

These models are optimized for deployment on NXP’s i.MX applications processors and microcontrollers, featuring quantization, pruning, and other optimization techniques for efficient edge AI deployment.

The model zoo has been widely adopted across NXP product lines and has seen 6 major releases, becoming a key resource for developers working with NXP’s eIQ machine learning software stack.