Net Runner is an on device environment for running computer vision machine learning models and collecting labeled data.
With Net Runner you can:
- Visualize model inputs and debug model outputs in real time.
- Run models on individual photos or entire collections of photos.
- Measure the latency and accuracy of your models.
- Collect and label data for existing models or models you would like to train.
Net Runner includes a number of MobileNet image classification models to help you get started, which are also a lot of fun to run!
Net Runner is open source, and the applications complete source code may be found at https://github.com/doc-ai/net-runner-ios.
Net Runner relies on the open source framework TensorIO, developed in tandem and specifically designed to support the rapid prototyping and testing of computer vision machine learning models. Learn more about TensorIO at https://github.com/doc-ai/TensorIO.