NVIDIA Updates TAO Toolkit
NVIDIA today released the latest version of TAO Toolkit, a low-code toolkit of the NVIDIA Train, Adapt, and Optimize (TAO) framework that simplifies and accelerates creation of artificial intelligence models to power enterprise speech and vision AI applications and services.
The NVIDIA TAO Toolkit is built on TensorFlow and PyTorch and accelerates model training by limiting AI/deep learning framework complexity. TAO allows developers to use transfer learning to create customized, production-ready models optimized for a broad range of industry specific use cases, including defect detection, language translation, custom voices, traffic management, and more.
The latest release of TAO Toolkit enables developers to do the following:
- Build new AI services or integrate TAO into existing ones with REST APIs and manage and orchestrate the TAO Toolkit service on Kubernetes.
- Fine-tune and optimize non-TAO models with TAO. Import pretrained weights from ONNX and take advantage of TAO features like pruning and quantization on models. This is supported for image classification and segmentation tasks.
- Understand model training performance by visualizing scalars such as training and validation loss, model weights, and predicted images in TensorBoard. Compare results between experiments by changing hyperparameters and choose the one that best fits needs.
- Use pretrained models to speed up the customization process to fine-tune through transfer learning with less data.
Some of the new pretrained models in this latest version can do the following:
- Apply data gathered from LIDAR sensors for robotics and automotive applications.
- Classify human actions based on human poses that can be used in public safety, retail, and worker safety use cases.
- Estimate keypoints on humans, animals, and objects to help portray actions or define object shapes.
- Create custom voices with just 30 minutes of recorded data to power smart devices, game characters, quick-service restaurant systems, and more.
Enterprise support for TAO Toolkit is available with NVIDIA AI Enterprise, an end-to-end software suite for AI development and deployment.