Robotics & hardware ·
A 13-million-parameter speech recognizer on a $10 microcontroller
A distilled, quantized version of Nvidia's small Conformer model, squeezed into 14 MB of flash on an ESP32-S3 so it can transcribe speech on the chip.
Made by u/wunschpunsch3D. This write-up follows the original post; it is not an independent product test.
Keep up with what gets built.
Read the daily AI briefing, browse the tools, or sign up for briefing email updates.