train_rnn

mlpractice.rnn_torch.train_rnn(encoded_lines, model, optimizer, iterations=1000)

Trains RNN on a given text.

Parameters

encoded_linesnp.ndarray, shape(n_samples, MAX_LENGTH)

Lines of input text converted to a matrix.

modeltorch.nn.Module

A model to train.

optimizertorch.optim.Optimizer

Optimizer that will be used to train a model.

iterationsint, optional

Number of optimization steps that the model will make.

Returns

training_historylist of float

Training history consisting of mean-loss-per-iteration records.