Accelerating AI Development with NVIDIA Train-Adapt-Optimize (TAO) and Weights & Biases
Learn how combining NVIDIA TAO (Train-Adapt-Optimize) and Weights & Biases can help kickstart an organization’s journey to leverage AI
Created on December 16|Last edited on January 31
Comment
Introduction
Leveraging AI, such as image classification, object detection, automatic speech recognition, etc., can fuel massive transformation within companies and business sectors. However, building AI & Deep Learning models from scratch is a daunting task. Common prerequisites—and challenges—to building these models include having a large amount of high quality training data and the right expertise to prepare the data, build the neural network, and continuously fine-tune the models to optimize performance. The high barrier to entry often becomes a blocker for enterprises of all sizes to adopt AI, despite the clear and measurable benefits AI has provided.
In this blog, we will discuss how the combination of NVIDIA TAO (Train-Adapt-Optimize) and Weights and Biases can help kickstart an organization’s journey to leverage AI—providing an accelerated starting point for common AI tasks. Developers can now visualize and compare multiple training runs using NVIDIA TAO Toolkit and Weights and Biases. We will walk through an example workflow for building an object detection model.
But first, a little bit of background.
NVIDIA TAO Toolkit
NVIDIA TAO Toolkit is a low-code solution enabling developers and enterprises to accelerate their model training and optimization processes. TAO Toolkit reduces the barrier to entry for anyone starting with AI by abstracting away the complexity of AI models and deep learning frameworks. With TAO Toolkit, you can use the power of transfer learning to fine-tune NVIDIA pre-trained models with your own data and optimize the model for inference to fit the needs of your business.
TAO Toolkit supports a wide range of computer vision tasks such as classification, object detection, segmentation, key point estimation, OCR, and more. TAO Toolkit provides several turnkey inference optimization that reduces the model complexity and size and increases inference throughput.
Weights & Biases
Weights & Biases helps ML teams build better models faster. With just a few lines of code in your notebook, you can instantly debug, compare, and reproduce your models—architecture, hyperparameters, git commits, model weights, GPU usage, datasets, and predictions—all while collaborating with your teammates.
W&B is trusted by more than 200,000 ML practitioners from some of the most innovative companies and research organizations in the world. To try it for free, sign up at Weights & Biases.
The NVIDIA TAO Toolkit + W&B Integration
Users of NVIDIA TAO Toolkit can now visualize all of their experimentation data within Weights and Biases. Users will be able to visualize, compare, and contrast multiple training runs to determine which model candidates best suit the needs of the project and which hyperparameters have the largest impact on model performance. The integration will also show how each model training consumes the underlying hardware to ensure resources are being fully utilized. Here is a guide to learn how to configure NVIDIA TAO to log experiments to W&B.
How to Leverage W&B with NVIDIA TAO - a Computer Vision Example
1) Get access to NVIDIA TAO if you don’t have it already.
2) Create an account for W&B if you don’t have one already:
- For enterprise users, check with your W&B admin
3) To try out an example:
- Download the getting started resource for TAO Toolkit on NGC by running the command mentioned below
ngc registry resource download-version nvidia/tao/tao-getting-started:4.0.0
- Instantiate the detectnet_v2.ipynb notebook present at
notebooks/tao_launcher_starter_kit/detectnet_v2/detectnet_v2.ipynb
in the downloaded samples directory.
- Uncomment the cells to install wandb and fill out your wandb API key.
- W&B logging will be enabled by default once you pass in your API key.
Here’s an example of the dashboard you’ll get automatically from running the notebook:
Conclusion
The NVIDIA TAO Toolkit helps organizations of all sizes jump-start their ability to leverage AI to improve their business and operations. With the TAO Toolkit 4.0 release, users can now leverage the W&B platform to track and understand how each incremental tuning job improves the performance of the AI models. If you’re interested in learning more or want to try NVIDIA TAO, visit the TAO overview page. If you’d like to try out Weights & Biases, check out the introductory Google Collab or sign up for a W&B account. Finally, if you’d like instructions on how to log ML experiments from TAO into W&B, take a look at the documentation here.
Add a comment
Iterate on AI agents and models faster. Try Weights & Biases today.