Step #1: Train YOLO-v4 model with NVIDIA TAO

This lab will use four important links from the left-hand navigation pane throughout the course of the lab.

  • TAO training Jupyter Notebook

  • System Console

  • Tensorboard Visualization

  • Deepstream Web Application


Open a Training Jupyter Notebook environment and follow the instructions in 01-Data-Preprocessing and 02-Object-Detection-with-NVIDIA-TAO.

As part of this lab you will learn:

  1. How does NVIDIA TAO work?

  2. Setting up the lab environment.

  3. Object Detection workflow with TAO

    1. Dataset conversion

    2. Exploring and downloading TAO pretrained models.

    3. Fine Tuning the pre trained model on the openimages dataset.

    4. Visualize the training on Tensorboard

    5. Prune the model to reduce the model size and accelerate inference time

    6. Retrain the pruned model to recover lost accuracy.

    7. Visualize inferences

    8. Export the pruned, retrained model to a .etlt file for deployment to DeepStream.

    9. Create a TensorRT engine

    10. Run and visualize inference on the exported .etlt model to verify deployment using TensorRT


To run a cell on the Jupyter Notebook, Click on the cell you want to run and press Shift + Enter. Linux bash commands can be run inside the Jupyter Notebook by adding a bang symbol (!) before the command inside the Jupyter Notebook cell.

© Copyright 2022-2023, NVIDIA. Last updated on Mar 23, 2023.