What is the Deep Learning Training Tool Beta?

If you’re a data scientist or researcher, this Beta 2 release of the Deep Learning Training Tool from Intel (formerly part of the Intel® Deep Learning SDK) helps you easily develop and train deep learning solutions using your own hardware.

• Plug and Train: Simplify installing popular deep learning frameworks, including Caffe* and TensorFlow*.
• Maximize Performance: Optimize creation of deep learning models on Intel® architecture.
• Increase Productivity: Easily prepare training data, design models, and train models with automated experiments and advanced visualizations.

Are you interested in unleashing inference performance for deep learning vision-related deployments? Get the Deep Learning Deployment Toolkit, which is part of Intel® Computer Vision SDK.

What's New

• Support for additional image classification topologies using Intel® Distribution for Caffe*
• TensorFlow 1.1 and Keras support through interactive notebooks and a model creation wizard that uses TensorFlow for document classification
• Support for TensorBoard*
• Easy-to-use APIs for multinode training with Caffe and TensorFlow using interactive notebooks
• Sample interactive notebooks for image segmentation and document classification
• Model compression integrated with the user interface

Release Notes

Technical Specifications

For more complete information about compiler optimizations, see our Optimization Notice.