Skip to main content

A neural network library built on top of TensorFlow for quickly building deep learning models.

Project description

A neural network library built on top of TensorFlow for quickly building deep learning models.

Build Status


nn.Tensor is the core data structure which is a wrapper for tf.Tensor and provides additional functionality. It can be created using the nn.tensor() function:

import nn

a = nn.tensor([1, 2, 3])
assert isinstance(a, nn.Tensor)
assert a.shape == (3, )

It supports method chaining:

c = a.square().sum()
assert c.numpy() == 14

and can be used with tf.Tensor objects:

import tensorflow as tf

b = tf.constant(2)
c = (a - b).square().sum()
assert c.numpy() == 2

It can also be used with high level APIs such as tf.keras:

model = nn.Sequential([
  nn.Dense(128, activation='relu'),

y = model(x)
assert isinstance(y, nn.Tensor)

and to perform automatic differentiation and optimization:

optimizer = nn.Adam()
with nn.GradientTape() as tape:
    outputs = model(inputs)
    loss = (targets - outputs).square().mean()
grads = tape.gradient(loss, model.trainable_variables)
optimizer.apply_gradients(zip(grads, model.trainable_variables))

To use it with ops that expect tf.Tensor objects as inputs, wrap the ops using nn.op():

mean = nn.op(tf.reduce_mean)
c = mean(a)
assert isinstance(c, nn.Tensor)

maximum = nn.op(tf.maximum, binary=True)
c = maximum(a, b)
assert isinstance(c, nn.Tensor)

or convert it to a tf.Tensor object using the tf() method or function:

b =
assert isinstance(b, tf.Tensor)

b =
assert isinstance(b, tf.Tensor)

See more examples here.



  • TensorFlow >= 2.0
  • Python >= 3.6

Install from PyPI (recommended):

pip install nn

Alternatively, install from source:

git clone
cd nn
pip install -e .

TensorFlow should be installed separately.


To run tests, install dependencies:

pip install -e .[tests]

and run:

pytest tests

Project details

Download files

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

Files for nn, version 0.1.1
Filename, size File type Python version Upload date Hashes
Filename, size nn-0.1.1.tar.gz (5.0 kB) File type Source Python version None Upload date Hashes View

Supported by

AWS AWS Cloud computing Datadog Datadog Monitoring DigiCert DigiCert EV certificate Facebook / Instagram Facebook / Instagram PSF Sponsor Fastly Fastly CDN Google Google Object Storage and Download Analytics Microsoft Microsoft PSF Sponsor Pingdom Pingdom Monitoring Salesforce Salesforce PSF Sponsor Sentry Sentry Error logging StatusPage StatusPage Status page