Intel Software258 тыс
Опубликовано 8 июня 2022, 3:31
Learn how Intel and Google have collaborated to deliver TensorFlow optimizations such as quantization and op fusions. Penporn Koanantakook of Google joins Anavai Ramesh and Andres Rodriguez of Intel to discuss how to use TensorFlow with Intel Neural Compressor to automatically convert to int8 or Bfloat16 data types for improved performance with minimal accuracy loss. They also discuss the new PluggableDevice mechanism co-architected by Intel and Google to deliver a scalable way to add device support to TensorFlow.
Intel and Google have collaborated closely for over four years to deliver optimizations such as quantization and op fusions. Now these optimizations are available in standard TensorFlow packages.
Learn how to access these optimizations and how to use them from Anavai Ramesh and Andres Rodriguez of Intel, and Penporn Koanantakook of Google. This includes a live demo of Intel Neural Compressor automatically quantizing a network to improve performance by 4x with a 0.06% accuracy loss. Topics include:
• Low-precision (8-bit) inference optimization
• Bfloat16 overview and acceleration on 3rd Gen Intel® Xeon®
• TensorFlow Auto Mixed Precision
This session also provides an overview of the PluggableDevice mechanism in TensorFlow 2.5, co-architected by Intel and Google to deliver a scalable way for developers to add new device support as plug-in packages.
Finally, get a preview of performance gain results from pre-production next-generation Intel Xeon Scalable processors (code-name Sapphire Rapids).
TensorFlow: bit.ly/3tm6vSj
Intel Neural Compressor: intel.ly/3xtQ1dd
Getting Started with Bfloat16 mixed precision: intel.ly/3aCeF2d
Model Zoo for Intel Architecture: bit.ly/3mu3FqB
The Intel® Developer Zone encourages and supports software developers developing applications for Intel hardware and software products. The Intel Software YouTube channel is a place to learn tips and tricks, get the latest news, and watch product demos from Intel and our many partners across multiple fields. You'll find videos covering the topics listed below, and to learn more, you can follow the links provided!
Connect with Intel Software:
Visit INTEL SOFTWARE WEBSITE: intel.ly/2KeP1hD
Like INTEL SOFTWARE on FACEBOOK: bit.ly/2z8MPFF
Follow INTEL SOFTWARE on TWITTER: bit.ly/2zahGSn
INTEL SOFTWARE GITHUB: bit.ly/2zaih6z
INTEL DEVELOPER ZONE LINKEDIN: bit.ly/2z979qs
INTEL DEVELOPER ZONE INSTAGRAM: bit.ly/2z9Xsby
INTEL GAME DEV TWITCH: bit.ly/2BkNshu
#intel #intelsoftware #ai #intelon
Accelerate Deep Learning with Intel-Optimized TensorFlow | Intel® On | Intel Software
Intel and Google have collaborated closely for over four years to deliver optimizations such as quantization and op fusions. Now these optimizations are available in standard TensorFlow packages.
Learn how to access these optimizations and how to use them from Anavai Ramesh and Andres Rodriguez of Intel, and Penporn Koanantakook of Google. This includes a live demo of Intel Neural Compressor automatically quantizing a network to improve performance by 4x with a 0.06% accuracy loss. Topics include:
• Low-precision (8-bit) inference optimization
• Bfloat16 overview and acceleration on 3rd Gen Intel® Xeon®
• TensorFlow Auto Mixed Precision
This session also provides an overview of the PluggableDevice mechanism in TensorFlow 2.5, co-architected by Intel and Google to deliver a scalable way for developers to add new device support as plug-in packages.
Finally, get a preview of performance gain results from pre-production next-generation Intel Xeon Scalable processors (code-name Sapphire Rapids).
TensorFlow: bit.ly/3tm6vSj
Intel Neural Compressor: intel.ly/3xtQ1dd
Getting Started with Bfloat16 mixed precision: intel.ly/3aCeF2d
Model Zoo for Intel Architecture: bit.ly/3mu3FqB
The Intel® Developer Zone encourages and supports software developers developing applications for Intel hardware and software products. The Intel Software YouTube channel is a place to learn tips and tricks, get the latest news, and watch product demos from Intel and our many partners across multiple fields. You'll find videos covering the topics listed below, and to learn more, you can follow the links provided!
Connect with Intel Software:
Visit INTEL SOFTWARE WEBSITE: intel.ly/2KeP1hD
Like INTEL SOFTWARE on FACEBOOK: bit.ly/2z8MPFF
Follow INTEL SOFTWARE on TWITTER: bit.ly/2zahGSn
INTEL SOFTWARE GITHUB: bit.ly/2zaih6z
INTEL DEVELOPER ZONE LINKEDIN: bit.ly/2z979qs
INTEL DEVELOPER ZONE INSTAGRAM: bit.ly/2z9Xsby
INTEL GAME DEV TWITCH: bit.ly/2BkNshu
#intel #intelsoftware #ai #intelon
Accelerate Deep Learning with Intel-Optimized TensorFlow | Intel® On | Intel Software
Свежие видео