Intel engineers recently launched the Intel NPU Acceleration Library v1.1, a Python library designed to work with Intel’s Neural Processing Unit (NPU) in the new Core Ultra (Meteor Lake) processors. This library simplifies interfacing with the Intel VPU/NPU kernel driver and allows for accelerated AI operations.
The latest version of the Intel NPU Acceleration Library, version 1.1, includes updates such as compilation improvements, support for Conv2D, a feature to clear the model cache to prevent out-of-memory issues, driver support for true quantization in eager mode, and other changes. This release marks the first update since the v1.0 milestone back in February.
For more information and to download the Intel NPU Acceleration Library v1.1, visit the official GitHub page.
Article Source
https://www.phoronix.com/news/Intel-NPU-Accel-Lib-1.1