On-device TinyML on the Silicon Labs Thunderboard Sense 2 (EFR32MG12, no NPU): an 85% IMU gesture recognizer running fully on-chip at ~87.5ms, plus a deep-audited hardware reference set.
iot embedded-systems bluetooth-low-energy cortex-m4 gesture-recognition silicon-labs efr32 embedded-machine-learning tinyml edge-impulse edge-ml efr32mg12 on-device-inference thunderboard-sense-2
-
Updated
Jun 7, 2026 - JavaScript