Transformers FastAPI Kit

A small, production-oriented wrapper around Hugging Face Transformers plus a FastAPI REST service for sentiment analysis.

Version 0.1.1 now includes a complete tutorial and a ZIP starter bundle inside the installed package.

Install

pip install "transformers-fastapi-kit[api]"

The first prediction downloads the default model (distilbert-base-uncased-finetuned-sst-2-english) from Hugging Face and caches it locally.

Python usage

from transformers_fastapi_kit import SentimentModel

model = SentimentModel()
print(model.predict("This package is excellent!"))
# {'label': 'POSITIVE', 'score': 0.99...}

Use a different model with SentimentModel(model_name="..."); pass device=0 for a CUDA GPU or leave the default -1 for CPU.

Run the API

From the project root:

pip install -e ".[api]"
uvicorn app.main:app --host 0.0.0.0 --port 8000

The interactive Swagger UI is then available at the API server's /docs route. For a public deployment, replace the local address with the public domain assigned to your API service.

curl -X POST http://127.0.0.1:8000/predict \\
  -H 'Content-Type: application/json' \\
  -d '{"text":"I love this API"}'

Endpoints: GET /health, POST /predict, and the generated OpenAPI schema at /openapi.json.

Included tutorial and ZIP bundle

The package includes transformers_fastapi_kit/resources/TUTORIAL.md and transformers_fastapi_kit/resources/transformers-fastapi-kit-starter.zip.

To locate them after installation:

from importlib.resources import files

resources = files("transformers_fastapi_kit") / "resources"
print(resources / "TUTORIAL.md")
print(resources / "transformers-fastapi-kit-starter.zip")

The full public documentation is available at https://transformers-ynoszgzv.manus.space.

Development

pip install -e ".[api,dev]"
pytest
python -m build
python -m twine check dist/*

License

MIT