Reproduce “A lightweight C Code Generator for Heterogeneous TinyML Systems” on OpenResearch
Turn your favorite coding agents into research agents that review literature, develop hypotheses, and run experiments.
Universal app for Apple silicon and Intel. Requires macOS 11 or later.
About this paper
Chen Liu, Bernhard Vogginger, Christian Mayr
The authors present a pure-C generator for bare-metal heterogeneous devices that moves data reordering and quantized-value rescaling from inference into code generation, using backend-specific templates to emit C. On RadarNet running on one custom radar DSP, it records 630 μs and a 25.0 kB binary, versus 1,470 μs and 31.2 kB for accelerated microTVM plus UMA; the comparison does not isolate individual optimizations.
Summary by alphaXiv.
Tip: change arxiv.org to autoarxiv.org in any arXiv abstract link to open the paper here.