By / AMD Deep Learning Software Director Mayank Daga
We are excited to announce the launch of TensorFlow v1.8 for ROCm GPUs, which includes the Radeon Instinct MI25. This is a significant milestone for AMD’s ongoing deep learning acceleration efforts. ROCm, which stands for Radeon Open Ecosystem, is our open-source software foundation for GPU computing on Linux. Our TensorFlow implementation utilizes MIOpen, a highly optimized GPU routine library for deep learning.
AMD provides a pre-built whl package, and the installation process is straightforward, similar to installing the general TensorFlow on Linux. We have released installation instructions and a pre-built Docker image.
In addition to supporting TensorFlow v1.8, we are also working to include all ROCm enhancements into the TensorFlow main repository. Some of these patches have already been merged upstream, and several others are under active review. While striving to comprehensively integrate enhancements, we will also release and maintain future TensorFlow versions that support ROCm, such as v1.10.
We believe the future of deep learning optimization, portability, and scalability is rooted in domain-specific compilers. Inspired by the early results of XLA, we are also working to enable and optimize XLA for AMD GPUs.
For more information on AMD’s work in this area, please visit
www.amd.com/deeplearning
More AI Reading:
· Code and Run Graphs with Eager Execution: Example of Optimizing Code with RevNet
· Alexa Can Now Respond to Sign Language! Using a Webcam and TensorFlow.js
· Want to Become an AI Artist? Use tf.keras and Eager Execution
Click “Read Original” to learn more about AMD!