sparrow-decider/sparrow-small-general
Text Classification • Updated
None defined yet.
Decisions that fit on a chip. Decision models from 47 MB on a laptop down to ~2 KB on a microcontroller. They answer yes/no, pick-one and score questions about a state, with calibrated probabilities.
| Model | Size | Runs on | Speed |
|---|---|---|---|
| sparrow-pico | ~2 KB per question | any microcontroller | 9.3 µs on ESP32-S3 |
| sparrow-nano-robot | 1.11 MB | ESP32-S3, Pi, laptop | 3.5 ms on ESP32-S3 |
| sparrow-nano-general | 2.2 MB | ESP32-S3, Pi, laptop | 3.8 ms on ESP32-S3 |
| sparrow-micro | 7.6 MB | Pi, laptop | 0.22 ms on a laptop |
| sparrow-tiny | 17.9 MB | laptop, Pi | 23 ms on a laptop |
| sparrow-small | 47 MB | laptop, Pi, Jetson | 64 ms on a laptop |
pip install sparrow-decider
- Website: https://sparrowdecider.com
- Code: https://github.com/sparrow-decider/sparrow-decider
- PyPI: https://pypi.org/project/sparrow-decider/
- Contact: vivek@sparrowdecider.com
Python (numpy only) and C99 (no malloc, ESP-IDF component). Code: Apache-2.0.
Unofficial community project. Not an AWS or Strands product.
Edit this `README.md` markdown file to author your organization card.