simaaisigmoid ============= The ``simaaisigmoid`` GStreamer plugin performs a sigmoid activation function on input tensor data. It takes a tensor as input and produces a tensor with the sigmoid applied to each element as output. The plugin is designed to work with SiMa.ai’s memory management system. Properties ---------- The following properties can be configured: =========== ======= ======= ==================================================================================================== Property Type Default Description =========== ======= ======= ==================================================================================================== silent boolean FALSE If TRUE, suppresses verbose output. transmit boolean FALSE If TRUE, transmits KPI data via GStreamer messages. dump-data boolean FALSE If TRUE, saves the output buffer data to ``/tmp/[plugin-name]-[frame-id].out``. num-buffers uint 5 Number of buffers allocated in the internal buffer pool. buffer-size uint 0 Size (in bytes) of each buffer in the internal buffer pool. Must be set appropriately for your data. =========== ======= ======= ==================================================================================================== Usage ----- The ``simaaisigmoid`` plugin can be integrated into a GStreamer pipeline as follows: .. code:: bash gst-launch-1.0 ...input... ! simaaisigmoid name='simaai_sigmoid' num-buffers=4 buffer-size=1600 ! fakesink Replace ``...input...`` with your input element. The ``buffer-size`` property *must* be set to a value appropriate for the size of your input tensors. The ``num-buffers`` property controls the number of buffers pre-allocated by the plugin. The plugin expects input caps of ``application/vnd.simaai.tensor, format=(string)DETESSDEQUANT`` and outputs caps of ``application/vnd.simaai.tensor, format=(string)SIGMOID``. Installation ------------ To install the ``simaaisigmoid`` plugin, copy the plugin library from ``/usr/local/simaai/plugin_zoo/gst-simaai-plugins-base/gst/`` into your project’s ``plugins/`` directory. Ensure the plugin is accessible to GStreamer’s plugin loading mechanism. Integration ----------- This plugin integrates into a GStreamer pipeline by receiving input buffers, processing them using the sigmoid function, and outputting the results to the next element in the pipeline. The plugin utilizes SiMa.ai’s memory management for efficient data handling. It also provides mechanisms for silent operation and KPI data transmission. The output buffers will contain metadata including a buffer ID and timestamp.