I built willitrun, a small CLI that tries to answer a question I kept running into with local/edge ML: will this model actually fit and run on my hardware?
It uses benchmark data when available and falls back to lightweight estimation otherwise.
One thing I wanted from the start was support for Hugging Face model IDs directly, so you can point the tool at a model from the Hub instead of manually entering all metadata yourself.
The goal right now is not to be perfect, but to be useful enough to filter out obviously bad choices before spending time downloading or testing models manually.
GitHub: GitHub - smoothyy3/willitrun: CLI to tell you if an ML model will fit and run on your device, using real benchmarks + lightweight estimation. · GitHub
PyPI: willitrun · PyPI
It is still early, and I’d be especially interested in feedback on:
-
HF model ID support
-
missing model metadata / parsing edge cases
-
benchmark coverage
-
cases where the estimates are off