Pilot environment for evaluation. For production deployments, contact us.
Batch - non interactive command line application
Summary
A TensorFlow image classifier — example of containerized ML on OSP.
Description
What it is A batch application that runs a pre-trained TensorFlow model to classify images into categories. It serves as a working example of how machine learning workloads can be packaged and run on OneSciencePlace through Docker containers.
Use cases
- Demonstrating how containerized ML applications can be exposed through OneSciencePlace
- A reference for app contributors building their own ML or batch applications
- Educational settings where students can submit jobs and see results without setting up TensorFlow themselves
What's included A Docker container with TensorFlow and a pre-trained image classification model. Use a prefilled value or provide an image URL to generate classification results, which are written to a file in the job output directory.
Where it runs The OneSciencePlace server, in a Docker container. Runs in under a minute for typical inputs.
Getting started Click Launch and select an image to classify. The job will run for a brief time, then write the classification results to the output directory.