-
Install package from source code (not published to PyPi)
$ git clone https://github.com/IrinaTerlizhenko/ml-neural-net-constructor.git $ cd ml-neural-net-constructor/src $ pip install .
-
Use it in your code
from netconstructor.network import NeuralNetwork net = NeuralNetwork()\ .with_dense_layer(...)\ .with_logistic_activation()\ .with_square_error() net.train(x, y, num_iterations) error = net.test(x_test)
The sample can be found here. numpy
and tensorflow
are required in
order to run this sample (tensorflow
is used only to load MNIST data). They can be installed separately
or with pip install -r requirements/requirements_dev.txt
(note: this also installs a bunch of other dev
dependencies).
Available on this page.
-
Install all required packages
pip install -r requirements/requirements.txt
-
Install all required packages
pip install -r requirements/requirements_dev.txt
-
Activate pytest if you're working in a JetBrains IDE
File > Settings > Tools > Python Integrated Tools > Testing > Default test runner