spatialml

Convert ONNX, TFLite, PB, and PyTorch models to QNN context binaries.

2|Updated Mar 25, 2025

Search Tags:#debugging #onnx #tflite #ml model conversion #qnn #pico device #securemr

Ready to run

npx skills add https://github.com/Pico-Developer/pySecureMR --skill spatialml

💬 Prompt for ChatGPT / Claude Chat▼

Please help me install this Agent Skill.
Skill: spatialml
Source: https://github.com/Pico-Developer/pySecureMR/tree/main/skills/spatialml
Command: npx skills add https://github.com/Pico-Developer/pySecureMR --skill spatialml

SYSTEM DOCUMENTATION & REQUIREMENTS

💡 This Skill includes scripts (resource) and references (resource) components.

What problem does it solve?

This Skill streamlines the process of converting machine learning models into a format compatible with Pico devices and provides tools for debugging these models on the device.

Core Features & Use Cases

Model Conversion: Convert various ML model formats (ONNX, TFLite, PB, PyTorch) into QNN context binaries.
On-Device Debugging: Inspect and debug ML models and SecureMR pipelines directly on an Android (Pico) device using Python tools.
Use Case: You have a PyTorch model trained for object detection. Use this Skill to convert it to a .serialized.bin format for your Pico headset and then use the model inspector to verify its performance on the device.

Quick Start

Use the spatialml skill to convert the attached model file 'my_model.onnx' to a QNN context binary.

Frequently Asked Questions about spatialml

High-intent search queries and answers about installing and using this skill.

FAQPage Schema

How do I convert an ONNX or PyTorch model to a QNN context binary for Pico devices?▼

To convert an ONNX or PyTorch model to a QNN context binary for Pico devices, you use a conversion process that requires Docker. This transforms your model into a .serialized.bin format for headset deployment.

Can I debug a machine learning model directly on a Pico headset?▼

Yes, you can debug a machine learning model directly on a Pico headset using Python-based inspection tools. This requires ADB for on-device interaction to evaluate model and SecureMR pipeline performance.

What formats are supported for ML model conversion to Pico devices?▼

Supported formats for ML model conversion to Pico devices include ONNX, TFLite, PB, and PyTorch. These models are converted into QNN context binaries to run on the target hardware.

Do I need Docker and ADB to convert and debug ML models for Pico devices?▼

Yes, you need Docker to perform the ML model conversion to QNN context binaries and ADB to facilitate on-device debugging and interaction with the Pico device.

How do I inspect a SecureMR pipeline on an Android device?▼

To inspect a SecureMR pipeline on an Android device, you use Python-based inspection tools connected via ADB. This allows you to debug and verify the pipeline directly on the hardware.