The latest version of the NVIDIA TAO Toolkit 4.0 boosts developer productivity with all-new AutoML capability, integration with third-party MLOPs services, and…
The latest version of the NVIDIA TAO Toolkit 4.0 boosts developer productivity with all-new AutoML capability, integration with third-party MLOPs services, and new pretrained vision AI models. The enterprise version now includes access to the full source code and model weights for pretrained models.
The toolkit enables efficient model training for vision and conversational AI. By simplifying complex AI models and deep learning frameworks, even developers without AI expertise can use the toolkit to produce AI models. Using transfer learning to fine-tune NVIDIA pretrained models with your own data, it is now possible to optimize model inference throughput without AI expertise or large training datasets.
Developers can create custom production-ready models optimized for specific environments and scenarios with TAO. A notable new feature helps developers build object detection models without massive amounts of data. The use cases include detecting assembly line defects, translating particular phrases across languages, or managing city traffic.
Release highlights
- Access to TAO source code and model weights for pretrained models.
- Train high-quality models with AutoML without the hassle of manually fine-tuning hundreds of parameters.
- Deploy on VMs from leading cloud providers and Kubernetes services like Amazon EKS or Azure AKS.
- Simplify infrastructure management and scaling on cloud machine learning services such as Google Colab, Google Vertex AI, and Microsoft Azure Machine Learning.
- New cloud integrations and third-party MLOps services, such as W&B and ClearML, provide developers and enterprises with an optimized AI workflow.
- Integrate with REST APIs. Quickly build a new AI service or integrate into an existing one with REST APIs.
- Use new transformer-based pretrained models (CitySemSegformer, Peoplenet Transformer) and retail-specific pretrained models (RetailObjectDetection, RetailObjectRecognition, and ReIdentificationNet.)
Resources
Get started
- Visit the TAO Toolkit getting started page for instructional videos and quick start guides.
- Download TAO Toolkit and pretrained models from NGC.
New Tutorials
- Read Train like an AI pro using the new AutoML feature in NVIDIA TAO.
- Read Creating Custom AI Models Using NVIDIA TAO Toolkit with Azure Machine Learning learn how to train and deploy custom AI models using NVIDIA TAO Toolkit with Azure Machine Learning.
- Watch a video to experiment with the NVIDIA TAO Toolkit and pretrained models on Google Colab.
Solutions using TAO Toolkit
- Next-generation IT and business service provider, Trifork uses TAO Toolkit to accelerate the development of the AI-based baggage tracking solution for airports.
- Fingermark, a company developing data-driven solutions for decision-making, uses TAO Toolkit in developing industrial vision AI solutions to improve worker safety.