[net] subdivisions=1 inputs=256 batch = 1 momentum=0.9 decay=0.001 time_steps=1 learning_rate=0.5 policy=poly power=4 max_batches=2000 [gru] batch_normalize=1 output = 1024 [gru] batch_normalize=1 output = 1024 [gru] batch_normalize=1 output = 1024 [connected] output=256 activation=linear [softmax] [cost] type=sse