maxidl / Places365-server

A lightweight server for Places365 scene classification

Places365-server

A lightweight server for the Places365 scene classification models, using pytorch 0.4 and starlette. See http://places2.csail.mit.edu/ and https://github.com/CSAILVision/places365 for the models.

Setup

1. Install dependencies via conda: conda env create -f environment.yml
2. Activate environment: source activate places365
3. Specify model architecture, host and port in serve_places365.py and run it.
4. Run example_query.py to classify an example image (change host and port if changed before).

Example output: {'patio': 0.621, 'restaurant_patio': 0.296, 'porch': 0.021, 'beer_garden': 0.018, 'courtyard': 0.012}