The fastest deep learning inference software for Arm Cortex-M microcontrollers

Follow the full discussion on Reddit.
we at Plumerai built the fastest and most memory-efficient deep learning inference software for Arm Cortex-M microcontrollers. It has 40% lower latency and requires 42% less RAM than TensorFlow Lite for Microcontrollers with Arm’s CMSIS-NN kernels while retaining the same accuracy. It also greatly outperforms any other deep learning inference software for Arm Cortex-M.