Machine Learning API<\/a> and delegates them to various hardware acceleration devices.\nThis HAL is not used directly by application developers; instead, it serves as a backend for NNStreamer.<\/p>\n<p>The process is as follows:<\/p>\n<ul>\n<li>An application uses the Tizen ML API.<\/li>\n<li>The underlying framework (NNStreamer's tizen-hal tensor_filter subplugin) calls the ML HAL API.<\/li>\n<li>The ML HAL API routes the request to a specific HAL ML Accelerator Backend.<\/li>\n<li>The backend executes the task on the target device and returns the result.<\/li>\n<\/ul>\n<img src=https://samsungtizenos.com/"media//tizen-hal-ml.png/" width=600\/>\n<p>You may check the HAL usage in the <code>tensor_filter<\/code> implementation <a href=https://samsungtizenos.com/"https:////github.com//nnstreamer//nnstreamer//blob//main//ext//nnstreamer//tensor_filter//tensor_filter_tizen_hal.cc/">here./n","table_of_content":"
