nerva_tensorflow documentation

A tiny, educational set of neural network components built on TensorFlow.

Install and build

# from repository root
python -m pip install -U sphinx sphinx-rtd-theme
# build HTML docs into docs_sphinx/_build/html
sphinx-build -b html docs_sphinx docs_sphinx/_build/html

API reference

`nerva_tensorflow`	nerva_tensorflow: Minimal neural network components built on top of TensorFlow.
`nerva_tensorflow.activation_functions`	Activation functions and utilities used by the MLP implementation.
`nerva_tensorflow.datasets`	In-memory data loader helpers and one-hot conversions.
`nerva_tensorflow.layers`	Neural network layers used by the MultilayerPerceptron class.
`nerva_tensorflow.learning_rate`	Learning-rate schedulers.
`nerva_tensorflow.loss_functions`	Analytic loss functions and their gradients used during training.
`nerva_tensorflow.matrix_operations`	Matrix operations built on top of TensorFlow to support the math in the library.
`nerva_tensorflow.multilayer_perceptron`	A simple multilayer perceptron (MLP) class.
`nerva_tensorflow.optimizers`	Optimizers used to adjusts the model's parameters based on the gradients.
`nerva_tensorflow.softmax_functions`	Softmax and log-softmax functions together with stable variants.
`nerva_tensorflow.training`	Training helpers for the MLP, including a basic SGD loop and CLI glue.
`nerva_tensorflow.utilities`	Miscellaneous utilities (formatting, timing, parsing, I/O).
`nerva_tensorflow.weight_initializers`	Weight and bias initialization helpers for linear layers.