Skip to main content
Warning: You are using the test version of PyPI. This is a pre-production deployment of Warehouse. Changes made here affect the production instance of TestPyPI (
Help us improve Python packaging - Donate today!

Use any TensorFlow model in a single line of code

Project Description

TFI provides a simple Python interface to any TensorFlow model. It does this by automatically generating a Python class on the fly.

Here’s an example of using TFI with a SavedModel based on Inception v1. This particular SavedModel has a single predict method and a SignatureDef that looks something like: predict(images float <1,224,224,3>) -> (categories string <1001>, scores float <1,1001>)

TFI in Action

>>> import tfi
>>> InceptionV1 = tfi.saved_model.as_class("./inception-v1.saved_model")

Passing in data

TFI can automatically adapt any data you provide to the shape expected by the graph. Let’s take a random photo of a dog I found on the internet…

>>> model = InceptionV1()
>>> image ="./dog-medium-landing-hero.jpg")
>>> result = model.predict(images=[image])
>>> categories, scores = result.categories, result.scores[0]

If we print the top 5 probabilities, we see:

>>> [(scores[i], categories[i].decode('utf-8')) for i in scores.argsort()[:-5:-1]]
[(0.80796158, 'bloodhound, sleuthhound'),
 (0.10305813, 'English foxhound'),
 (0.064740285, 'redbone'),
 (0.009166114, 'beagle')]

Not bad!

Image data

The function uses mimetypes to discover the right data decoder to use. If an input to a graph is an "image/*", TFI will automatically decode and resize the image to the proper size. In the example above, the JPEG image of a dog is automatically decoded and resized to 224x224.


If you look closely at the example code above, you’ll see that the images argument is actually an array. The class generated by TFI is smart enough to convert an array of images to an appropriately sized batch of Tensors.

Graphs with variables

Each instance of the class has separate variables from other instances. If a graph’s variables are mutated during a session in a useful way, you can continue to use those mutations by calling methods again on that same instance.

If you’d like to have multiple instances that do not interfere with one another, you can create a second instance and call methods on each of them separately.

TFI and SavedModels

TFI uses the information in a SavedModel’s SignatureDefs to generate methods on the resulting class. The keyword argument names for each method are also pulled from info in the SignatureDef.

The SavedModel used in the example was created using the tf.estimator.Estimator#export_savedmodel function.

Getting Started

TFI is on PyPI, install it with pip install tfi.

Future work

Adapting functions to handle queues and datasets wouldn’t require much effort. If this is something you’d like me to do, please file an issue with your specific use case!

Extending to support more formats is also quite straightforward. File an issue with a specific format you’d like to see. For bonus points, include the expected tensor dtype and shape. For double bonus points, include a way for me to test it in a real model.

It’s not very easy to create well-formed SavedModels today. If this is something you’d like TFI to do in the future… file an issue. ;)


If you’re curious, the photo used above was from a random Google image search.

PyPI packaging was way easier because of this fantastic guide.

Release History

Release History

This version
History Node


History Node


History Node


History Node


Download Files

Download Files

Download the file for your platform. If you're not sure which to choose, learn more about installing packages.

File Name & Checksum SHA256 Checksum Help Version File Type Upload Date
tfi-0.4-py2.py3-none-any.whl (9.9 kB) Copy SHA256 Checksum SHA256 py2.py3 Wheel Sep 14, 2017
tfi-0.4.tar.gz (7.5 kB) Copy SHA256 Checksum SHA256 Source Sep 14, 2017

Supported By

WebFaction WebFaction Technical Writing Elastic Elastic Search Pingdom Pingdom Monitoring Dyn Dyn DNS Sentry Sentry Error Logging CloudAMQP CloudAMQP RabbitMQ Heroku Heroku PaaS Kabu Creative Kabu Creative UX & Design Fastly Fastly CDN DigiCert DigiCert EV Certificate Rackspace Rackspace Cloud Servers DreamHost DreamHost Log Hosting