ServeIt lets you serve model predictions and supplementary information from a RESTful API using your favorite Python ML library in as little as one line of code:

from serveit.server import ModelServer
from sklearn.linear_model import LogisticRegression
from sklearn.datasets import load_iris

# fit logistic regression on Iris data
clf = LogisticRegression()
data = load_iris()
clf.fit(data.data, data.target)

# initialize server with a model and start serving predictions
ModelServer(clf, clf.predict).serve()

Your new API is now accepting POST requests at localhost:5000/predictions! Please see the examples directory for detailed examples across domains (e.g., regression, image classification), including live examples.

Current ServeIt features include:

1. Model inference serving via RESTful API endpoint
2. Extensible library for inference-time data loading, preprocessing, input validation, and postprocessing
3. Supplementary information endpoint creation
4. Automatic JSON serialization of responses
5. Configurable request and response logging (work in progress)

The following libraries are currently supported:

• Scikit-Learn
• Keras
• PyTorch

Installation is easy with pip: pip install serveit

You can build locally with: python setup.py

MIT

Please consider buying me a coffee if you like my work: