Class for managing a single dataset, includes data and target fields and has some utility functions. It allows also to convert the dataset i

Simple object for standard datasets. Has the field `train` `validation` and `test` and support indexing

Base object for building neural networks

mirror of the tf.train.Optimizer.

Optimizer that implements the Momentum algorithm. Computes (if `use_nesterov = False`): ``` accumulation = momentum * accumulation + gra

Optimizer that implements the gradient descent algorithm.

Wrapper for performing gradient-based parameter optimization

Implementation follows Pedregosa's algorithm HOAG

adds some meta_parameters and outer_gradients computation related keys