MInD
Menu
Menu
MInD
BrainService serves two primary functions within the MInD platform.
Firstly, it empowers users to train the neural network integrated into the measurement processing analysis document (Brain) effortlessly. Leveraging the user-friendly interface of the MInD editor, individuals can construct and fine-tune their Brains and Neural Networks without the need for programming skills. By connecting various nodes, users can navigate through a range of operations to build and optimize their Brains. The learning process of the Neural Network can be visually monitored, allowing users to track the progress and observe the development of the training on selected measurements.
TensorFlow framework is at the core of BrainService’s learning process, requiring a robust Nvidia video card for optimal performance.
Secondly, BrainService executes deployed Brains. By establishing a separate API connection, users can input measurements and obtain desired results. Both TensorFlow and ONNX frameworks can be used to run the deployed Brains.
While running inference tasks can be performed on lightweight hardware like the Raspberry Pi 4 (included in our compact MIB device), more powerful Nvidia hardware is recommended for enhanced performance in demanding scenarios.
BrainService offers flexibility in deployment options, allowing it to run locally or on a remote machine, including cloud environments. The operating system compatibility spans Windows and Linux.
Copyright © 2023 Machine Intelligence Zrt.
