Develop and Tune Computer Vision Models using NVIDIA TAO AutoML
Develop and Tune Computer Vision Models using NVIDIA TAO AutoML (Latest Version)

Next Steps

Congratulations!

You have successfully completed the object detection pipeline with NVIDIA TAO using AutoML.

The NVIDIA AI Enterprise software suite includes the same frameworks and tools used in this lab that is needed for best-in-class AI. And since NVIDIA has already done the integration and performance optimizations, you can get started with AI easier and faster.

What’s next?

  1. Want to do more exercises with TAO Toolkit?

  • Try training/optimization and AutoML on your own dataset

  • Try different networks. Notebooks for other networks are available in the client folder in the provide Jupyter environment. Users will find notebooks for Training/optimization in end2end, AutoML in automl directory. Minor changes will be required in the notebooks. Use the provided YOLOv4 notebook as reference for setup.

  • Download TAO and run on your own infrastructure. TAO can be downloaded from NGC catalog

  • Read blog: Training like an AI pro using NVIDIA TAO AutoML

  1. There is also another LaunchPad lab, that uses TAO Toolkit & Deepstream SDK to train, customize and deploy an object detection model available here: Develop a Computer Vision Custom Object Detection Model

  2. Need more information? Here are some resources to learn more about TAO Toolkit & NVIDIA AI Enterprise:

  1. Ready to start your AI journey? Contact your NVIDIA Account rep or NVIDIA Partner.

Note

After you have completed the lab, you can use the environment for additional proofs of concepts. A Jupyter lab instance and an SSH console into the instance are also accessible via the left navigation pane. The SSH Console can be used to download new datasets to use with the Jupyter notebook.

© Copyright 2022-2023, NVIDIA. Last updated on Mar 20, 2023.