Run on-device ML inference for Spectacles with SnapML on camera frames.
npx skills add https://github.com/specs-devs/agentic-tools --skill snap-ml
Please help me install this Agent Skill. Skill: snap-ml Source: https://github.com/specs-devs/agentic-tools/tree/main/skills/snap-ml Command: npx skills add https://github.com/specs-devs/agentic-tools --skill snap-ml
Integrate machine learning models directly into Spectacles for on-device, real-time vision tasks, enabling developers to deploy ML-powered features without cloud latency.
Run a starter SnapML workflow to load a trained model and bind it to the camera feed for real-time inference.
High-intent search queries and answers about installing and using this skill.
On-device ML inference runs directly on Spectacles using SnapML to execute object detection, classification, and segmentation on live camera frames. You need a trained model file, tensor bindings, and an on-device runtime in Lens Studio.
Yes, an ONNX model can be used for real-time computer vision in Lens Studio. You must load the trained model file and bind its input and output tensors to the camera feed for SnapML to process the frames.
SnapML is used for real-time scene understanding and interactive AI-assisted AR experiences on Spectacles. It processes camera frames on-device to enable live filtering and object recognition without cloud latency.
On-device machine learning on Spectacles does not require an internet connection for inference. SnapML executes the trained model locally on the device to provide real-time vision tasks without cloud latency.
To bind a trained ML model for live filtering, you establish proper input and output tensor bindings between the model file and the camera frames. This allows the SnapML runtime to process real-time visual data on Spectacles.
anton-abyzovBuild end-to-end computer vision pipelines for classification, detection, segmentation, and generation.
adamhjkBuild interactive 2D graphics and animations with PixiJS.
specs-devsIntegrate AR/mobile keyboards for text input in Spectacles applications.
ldzhhxxDeploy machine learning models as scalable APIs on serverless GPU cloud.
