Inertia-1: An Open Exploration to a Unified Motion Foundation Model
Which summary reads better? Pick one — models revealed after.Both summaries are AI-generated.
Pretrained on over 18 million hours of unlabeled accelerometry data, Inertia-1 is a motion foundation model that generalizes zero-shot to unseen body placements, sampling rates down to 1 Hz, and novel sensor modalities like gyroscopes and magnetometers. For engineers shipping wearable or hardware-integrated software, this eliminates the high overhead of training and maintaining bespoke models for every distinct device form factor and sensor layout. You can now deploy a single, robust backbone that natively scales across diverse hardware configurations and multi-sensor streams without retraining.
A single accelerometry backbone pretrained self-supervised on 18M+ hours transfers zero-shot across body placements and even unseen sensor modalities (gyroscope, magnetometer), holding accuracy down to 1Hz sampling with 30–60s windows as the sweet spot. If you build wearable/IMU pipelines, this replaces per-placement, per-task bespoke models with one adaptable representation—cutting retraining and labeling costs—but note the practical constraints: keep full triaxial input rather than vector-magnitude, use time-domain modeling for gait/health signals, and bump sampling rate for fine-grained clinical tasks.