The Khronos Group announces a significant expansion in the ecosystem for the NNEF (Neural Network Exchange Format) open, royalty-free standard that enables hardware manufacturers to reliably exchange trained neural networks between training frameworks and inference engines. New and improved NNEF open source convertors, including for TensorFlow Lite and ONNX, enables NNEF to be used to carry trained frameworks from a wider range of training frameworks. A set of extensions to the NNEF 1.0 specification enables NNEF files to contain richer networks of operations and topologies. Finally, an openly available NNEF Model Zoo enables inferencing engines to test their reliable import of NNEF models. More information on NNEF can be found at the Home Page.
Another blog explaining Vulkan synchronization has just been posted by @Themaister. Synchronization in Vulkan is a large hurdle to overcome when learning the API, and rather than mechanically explaining how it works, the goal in this blog is to instil a mental model in the reader. Despite its reputation for maddening complexity, it is actually understandable and quite logical once you get over the initial hurdles.
The Khronos Group had several session on BOF Day at SIGGRAPH 2019. Most of the slides and video are now online:
Be sure to visit the event page for a detailed list of links to all the presentations and videos.
Vulkan 1.1.119 was released right after 1.1.118. In total, three new extensions have been added: VK_AMD_shader_core_properties2, VK_AMD_pipeline_compiler_control and VK_KHR_pipeline_executable_properties. The first two are AMD vendor extensions, while the most recent extension was worked on jointly by engineers from Intel, Valve, Google, NVIDIA, AMD, Arm, and Samsung. See the Change log on GitHub for additional the details.
The latest version of the Oculus Unreal Engine Integration adds support for Vulkan on Oculus Quest and Oculus Go. Unity told us they will add the same “later this year.”
This year at Hot Chips, the Khronos Group will have a demo table at which Khronos member AMD will demonstrate using OpenVX for inference using an NNEF model from the model zoo.