Tool maintenance plays a crucial role in maximizing the tools’ life and efficiency in the manufacturing industry. Performing maintenance manually is time-consuming and prone to human error. This gives rise to the need to automate and optimize the process for desired qualitative and quantitative output from the machinery.
Yuri Winche Achermann, in his article, highlights the Tool Detective project that checks the wear and tear of metal cutting tools in every machining cycle. The product uses computer vision to segment the tool's wear and shows its analysis on a dashboard. It increases operational efficiency and helps Enterprise Resource Planning (ERP) software in efficient resource management.
The Tool Detective uses the following Intel® oneAPI components:
The multi-architecture, cross-vendor support of the Intel oneAPI tools makes it possible for the Tool Detective to operate across diverse hardware available at the client’s location. Instead of being restricted to proprietary vendor-locked architectures, it helps achieve greater portability, compatibility, easier upgrades, and reduced scalability costs.
Explore Intel® oneAPI Deep Neural Network (oneDNN) Library
oneDNN library provides highly optimized implementations of key building blocks of deep learning models such as convolution, pooling, RNN cells, LSTM cells, etc. It improves the performance of deep learning frameworks such as AI Tools from Intel, PyTorch, TensorFlow*, and OpenVINO™ toolkit. It enables accelerated development and deployment of platform-independent deep learning workloads using a common API for CPUs, GPUs, or both.
Learn About Intel® Extension for PyTorch*
Intel Extension for PyTorch enables leveraging Intel’s latest hardware and software optimizations for PyTorch. It accelerates PyTorch training and inference with oneDNN features and minimal code changes. It lets you leverage optimized instruction set features such as Intel® Advanced Vector Extensions (Intel® AVX-512) and Intel® Advanced Matrix Extensions (Intel® AMX) on Intel’s CPUs and GPUs. It also allows distributed training of deep learning models with Intel® oneAPI Collective Communications Library (oneCCL).
What’s Next?
Get started with oneDNN and Intel Extension for PyTorch today – speed up and improve the performance of deep learning applications on Intel’s latest hardware! We encourage you to check out other AI, HPC, and Rendering tools in Intel’s oneAPI-powered software portfolio.
Useful Resources
Get the Software
Install oneDNN as a part of the Intel® oneAPI Base Toolkit (Base Kit) or download its standalone version. Intel Extension for PyTorch is available as a part of the AI Tools Selector, or you can get its standalone version.
You can also experiment with Intel’s AI software optimizations on the latest hardware on Intel® Developer Cloud.
About the Author
Yuri Winche Achermann is a data engineer at the Fraunhofer Institute for Production Technology (Germany). He was the winner of the Intel Student Ambassador 2023 hackathon for oneAPI.
Technical Software Product Marketing Engineer, Intel