Minimal Python client to communicate with TensorFlow Serving.
Python client to communicate with the TensorFlow RPC API. Similar to the examples in the documentation of TensorFlow Serving, we use gRPC as the RPC client.
- Link TensorFlow Serving as a sub-module here. To compile protobufs properly during build.
- When distributing, get rid of tf.contrib functions to remove TensorFlow as a dependency.
Check out the source e.g. via git clone https://github.com/sebastian-schlecht/tensorflow-serving-python.git and run
python setup.py install.
This package depends on the following pip modules, so you might want to install those first:
- Cython
- grpcio
- grpcio-tools
- tensorflow
Alternatively, we offer an unofficial python wheel to be installed with:
pip install http://www.mealomi.com/storage/tensorflow_serving_python/tensorflow_serving_python-0.1-py2-none-any.whl
Note
When running on Linux, make sure pip is upgraded via pip install --upgrade pip such that TensorFlow is installed correctly.
Examples can be found in the examples folder.