polaris-audio/polaris-asr-tiny
Speech recognition40M paramsonnxLicense: apache-2.0en#asr#on-device
Updated Sep 3, 2026
Part of the Open Pace seed catalog. Weights are listed for the preview and become downloadable at launch.
Overview
A tiny recogniser for voice commands and short dictation on phones.
polaris-audio/polaris-asr-tiny is a 40M-parameter speech recognition model, published in onnx under the apache-2.0 license.
Intended use
- Speech recognition in en.
- Research, prototypes and products that keep a person in the loop.
- Fine-tuning as a starting point for a narrower task.
How to use
# Command-line client (planned; the shape may change)
pace pull polaris-audio/polaris-asr-tiny
# Python (planned)
from openpace import load
model = load("polaris-audio/polaris-asr-tiny")Limitations
- Heavy accents, overlapping speakers and very noisy audio lower accuracy.
- Never use a synthetic voice to impersonate a real person.
License
Released under apache-2.0. Read the license file before using the weights commercially.