Python Embedded AI

Powerful frameworks for integrating AI into embedded systems using Python. Optimize, train, and deploy edge models efficiently.

Getting Started

Install SDK

Use pip to install the Python SDK and its dependencies.

pip install emb-python-sdk

Version 2.0.0 now supports PyTorch Mobile and ONNX optimization.

Import and Initialize


from emb.sdk import PythonEdge

client = PythonEdge(
    api_key="YOUR_API_KEY",
    project="edge-computing"
)

Key Features

Tensor Conversion

Convert PyTorch, Keras, or ONNX models for edge deployment.

onnx2emb converter available

Model Optimization

Quantization, pruning, and compression for resource-limited devices.

Prune models by 50%+ W/O loss

Edge Inference

Run inference models on devices with < 0.1s latency.

Real-time inference

Usage Examples

Quantize Neural Network


import emb.quantize as qt

model = qt.Quantizer(model="my-model.onnx")
result = model.quant8()
print(f"Saved: {result.size} MB")

Run Prediction


from emb.models import ObjectDetector

detector = ObjectDetector("resnet34-edge")
frames = detector.stream_capture("/dev/video0")
results = [detector.predict(frame) for frame in frames]

Ready to Build with Python?

Join 100,000+ developers using the EMB SDKs to power their embedded and AI projects.