NVIDIA pre-trained deep learning models and the Transfer Learning Toolkit (TLT) give you a rapid path to building your next AI project. Whether you’re a DIY enthusiast or building a next-gen product with AI, you can use these models out of the box or fine-tune with your own dataset. The purpose-built, pre-trained models are trained on the large datasets collected and curated by NVIDIA and can be applied to a wide range of use cases. TLT is a simple AI toolkit, shipped with Jupyter notebooks, that requires little to no coding for taking pre-trained models and customizing them with your own data.

In this webinar, we’ll show you how to train your own gesture recognition deep learning pipeline. We’ll start with a pre-trained detection model, repurpose it for hand detection, and use it together with the purpose-built gesture recognition model. Once trained, we’ll deploy this model on NVIDIA® Jetson™. The gesture-recognition application can be deployed on a robot to understand human gestures and interact with humans. In addition, TLT supports other pre-trained models such as facial landmark estimation, gaze estimation, people detection, and many more. Using pre-trained models and TLT lets you accelerate your development time—from concept to production.

• Highly optimized pre-trained models for various industry use cases

• How to fine-tune with your own data on new pre-trained models and use them to reduce your total development time

• Developing an end-to-end training pipeline and deploying the trained model on NVIDIA SDKs

Content

content goes here