# Neural Networks

This section will describe how you can train, export, and integrate various AI models on our devices. We will show how to measure
the performance of a model and possible ways to optimize the performance. Furthermore, we will talk about how to post-process
outputs of models.

### Model Zoo

Visit our DepthAI Model Zoo to explore a range of pre-trained AI models, optimized for performance on DepthAI devices and suited
for various applications.

[Model Zoo](https://docs.luxonis.com/software/ai-inference/zoo.md)

### Conversion

For adapting AI models from frameworks like PyTorch to Luxonis devices, check our Conversion page for easy-to-follow instructions.

[Conversion](https://docs.luxonis.com/software/ai-inference/conversion.md)

### Integrations

Explore our Integrations page for guidance on exporting Yolo models and using Roboflow models with our hardware.

[Integrations](https://docs.luxonis.com/software/ai-inference/integrations.md)

### Post-processing

Visit the Post-processing page for key steps to refine AI model outputs for use on Luxonis devices.

[Post-processing](https://docs.luxonis.com/software/ai-inference/post-processing.md)

### Performance Optimization

Enhance your AI models' performance on Luxonis devices with helpful tips from our Performance Optimization page.

[Performance Optimization](https://docs.luxonis.com/software/ai-inference/performance.md)

### Training

Learn about Luxonis's training and inference capabilities through our collection of Jupyter notebook tutorials.

[Training](https://docs.luxonis.com/software/ai-inference/training.md)