def test_network_step_no_mix_forward(self):
spec = blur.NetworkSpec(use_forward_activations_for_synapse_update=False)
genome = genome_util.convert_genome_to_tf_variables(
genome_util.create_backprop_genome(num_species=1))
initializer = lambda params: 2 * tf.ones(params.shape, dtype=tf.float32)
data = random_dataset()
state = blur_meta.init_first_state(
genome,
synapse_initializer=initializer,
data=data,
hidden_layers=[256, 128])
input_fn = data.make_one_shot_iterator().get_next
data_support_fn, _ = blur_meta.episode_data_fn_split(input_fn)
blur.network_step(
state,
genome,
data_support_fn,
data.make_one_shot_iterator().get_next,
network_spec=spec,
env=blur_env.tf_env)
g = tf.get_default_graph()
synapse_pre = g.get_operation_by_name(
'step/backward/synapse_update/hebbian_pre').inputs[0]
synapse_post = g.get_operation_by_name(
'step/backward/synapse_update/hebbian_post').inputs[0]
self.assertIn('backward', synapse_pre.name)
self.assertIn('backward', synapse_post.name)
self.assertNotIn('forward', synapse_pre.name)
self.assertNotIn('forward', synapse_post.name)
def get_blur_state(env, inp, out, ow):
pre = np.concatenate([inp, np.zeros_like(inp)],
axis=-1).astype(blur_env.NP_FLOATING_TYPE)
post = np.concatenate([np.zeros_like(out), out],
axis=-1).astype(blur_env.NP_FLOATING_TYPE)
ww = ow.astype(blur_env.NP_FLOATING_TYPE)
ww = ww[Ellipsis, None]
synapse = synapse_util.combine_in_out_synapses(
ww, synapse_util.transpose_synapse(ww, env), env=env)
synapse = synapse_util.sync_states_synapse(synapse, env, num_states=2)
genome = genome_util.convert_genome_to_tf_variables(
genome_util.create_backprop_genome(num_species=1))
network_spec = blur.NetworkSpec()
network_spec.symmetric_synapses = True
network_spec.batch_average = False
network_spec.backward_update = 'multiplicative_second_state'
return pre, post, synapse, genome, network_spec
def init_first_state(genome,
data,
hidden_layers,
synapse_initializer,
create_synapses_fn=synapse_util.create_synapses,
network_spec=blur.NetworkSpec(),
env=blur_env.tf_env):
"""Initialize the very first state of the graph."""
if create_synapses_fn is None:
create_synapses_fn = blur.create_synapses
if callable(genome):
num_neuron_states = genome(0).synapse.transform.pre.shape[-1]
else:
num_neuron_states = genome.synapse.transform.pre.shape[-1]
output_shape = data.element_spec[1].shape
num_outputs = output_shape[-1]
num_inputs = data.element_spec[0].shape[-1]
batch_dims = output_shape[:-1]
layer_sizes = (num_inputs, *hidden_layers, num_outputs)
layers = [
tf.zeros((*batch_dims, h, num_neuron_states)) for h in layer_sizes
]
synapses = create_synapses_fn(layers, synapse_initializer)
if network_spec.symmetric_in_out_synapses:
for i in range(len(synapses)):
synapses[i] = blur.sync_in_and_out_synapse(synapses[i],
layers[i].shape[-2],
env)
if network_spec.symmetric_states_synapses:
for i in range(len(synapses)):
synapses[i] = blur.sync_states_synapse(synapses[i], env)
num_updatable_units = len(hidden_layers) + 1
ground_truth = tf.zeros((*batch_dims, num_outputs))
return blur.NetworkState(layers=layers,
synapses=synapses,
ground_truth=ground_truth,
updatable=[False] + [True] * num_updatable_units)