ENVI Deep Learning 1.0 Release Notes

This is the initial release of ENVI Deep Learning.

You must have ENVI 5.5.2 installed and licensed on your system.

Refer to the Deep Learning Help for instructions on using the tools and API. Access the help as follows:

• Windows: Go to the Start menu and select Programs > ENVI 5.5 > Help > ENVI Deep Learning 1.0 Help.
• Linux: Type envideeplearninghelp in a terminal window.

Supported Platforms

The following table lists the platforms that the current version of ENVI Deep Learning supports:

Platform	Hardware	Operating System	Supported Versions
Windows	Intel/AMD 64-bit	Windows	10, 2016 Server
UNIX	Intel/AMD 64-bit	Linux	Kernel 3.10 or higher, glibc 2.17 or higher

Supported versions indicate that ENVI Deep Learning was either built on or tested on that version.

ENVI Deep Learning uses TensorFlow™ technology. TensorFlow has a minimum set of hardware and software requirements. The following NVIDIA® software must be installed on your system. You will need to install the software yourself; the links to download it are provided here:

• NVIDIA GPU card with CUDA® Compute Capability 3.5 or higher. See the list of CUDA-enabled GPU cards.
• NVIDIA GPU drivers (https://www.nvidia.com/Download/index.aspx?lang=en-us): CUDA 9.0 requires version 384.x or higher.

To determine if your system meets the requirements for ENVI Deep Learning, start the Deep Learning Guide Map in the ENVI Toolbox. From the Deep Learning Guide Map menu bar, select Tools > Test Installation and Configuration.

A minimum of 8 GB of GPU memory is recommended, particularly when training deep learning models.

You can run ENVI Deep Learning using a CPU as long as it meets the NVIDIA driver version requirements (384.x or higher); however, this is not recommended because training and other processes will be significantly slower compared to using a GPU.