FaceTron FAQs

Question 1

What ONNX models can I use with FaceTron?

Accepted Answer

FaceTron supports various ONNX models for face embedding, including ArcFace, SCRFD, and Glint360K. It comes with lightweight ONNX models under the `/models` directory for testing. You can replace these with your high-resolution models.

Question 2

What is FaceTron?

Accepted Answer

FaceTron is a high-performance face embedding server built with ONNX Runtime. It allows you to easily integrate face recognition capabilities into your applications with dynamic multi-model support.

Question 3

What are the key features of FaceTron?

Accepted Answer

FaceTron offers dynamic multi-model support, OpenAPI 3.1 & Swagger UI, built-in OpenTelemetry tracing, face detection with bounding boxes, and MCP metadata support. It's designed for scalable face recognition applications.

Question 4

Does FaceTron support tracing?

Accepted Answer

Yes, FaceTron includes built-in support for OpenTelemetry tracing. You can enable tracing by calling `setup_telemetry()` in `main.py`. Tracing can be disabled via the `DISABLE_OTEL=true` environment variable.

Question 5

How do I set up and run FaceTron?

Accepted Answer

You can set up FaceTron using Docker Compose or pip. Clone the repository, add your ONNX models, and then run the server using either `docker-compose up -d` or `pip install -r requirements.txt` followed by `python main.py`.

FaceTron

About

Key Features

Use Cases