pytorch_utils module

pytorch_utils.build_activation(type)

pytorch_utils.build_optimizer(model, type, lr, kerasDefaults, trainable_only=True)

pytorch_utils.get_function(name)

pytorch_utils.initialize(weights, type, kerasDefaults, seed=None, constant=0.0)

pytorch_utils.mse(y_true, y_pred)

pytorch_utils.set_parallelism_threads()

pytorch_utils.set_seed(seed)

Set the random number seed to the desired value

Parameters

seed (integer) – Random number seed.

pytorch_utils.xent(y_true, y_pred)