def test_integration_with_keras():
TestTransition = collections.namedtuple('TestTransition', ['observation'])
# Just a smoke test, that nothing errors out.
n_transitions = 10
obs_shape = (4, )
network_sig = data.NetworkSignature(
input=data.TensorSignature(shape=obs_shape),
output=data.TensorSignature(shape=(1, )),
)
trainer = supervised.SupervisedTrainer(
network_signature=network_sig,
target=supervised.target_solved,
batch_size=2,
n_steps_per_epoch=3,
replay_buffer_capacity=n_transitions,
)
trainer.add_episode(
data.Episode(
transition_batch=TestTransition(
observation=np.zeros((n_transitions, ) + obs_shape), ),
return_=123,
solved=False,
))
network = keras.KerasNetwork(network_signature=network_sig)
trainer.train_epoch(network)
def test_multiple_targets():
TestTransition = collections.namedtuple('TestTransition',
['observation', 'agent_info'])
network_sig = data.NetworkSignature(
input=data.TensorSignature(shape=(1, )),
# Two outputs.
output=(
data.TensorSignature(shape=(1, )),
data.TensorSignature(shape=(2, )),
),
)
trainer = supervised.SupervisedTrainer(
network_signature=network_sig,
# Two targets.
target=(supervised.target_solved, supervised.target_qualities),
batch_size=1,
n_steps_per_epoch=1,
replay_buffer_capacity=1,
)
trainer.add_episode(
data.Episode(
transition_batch=TestTransition(
observation=np.zeros((1, 1)),
agent_info={'qualities': np.zeros((1, 2))},
),
return_=123,
solved=False,
))
class TestNetwork(core.DummyNetwork):
"""Mock class."""
def train(self,
data_stream,
n_steps,
epoch,
validation_data_stream=None):
np.testing.assert_equal(
list(data_stream()),
[
testing.zero_pytree(
(network_sig.input, network_sig.output),
shape_prefix=(1, ))
],
)
return {}
trainer.train_epoch(TestNetwork(network_sig))
def test_integration_with_keras():
# Just a smoke test, that nothing errors out.
n_transitions = 10
obs_shape = (4, )
trainer = supervised.SupervisedTrainer(
input_shape=obs_shape,
target_fn=supervised.target_solved,
batch_size=2,
n_steps_per_epoch=3,
replay_buffer_capacity=n_transitions,
)
trainer.add_episode(
data.Episode(
transition_batch=_TestTransition(
observation=np.zeros((n_transitions, ) + obs_shape), ),
return_=123,
solved=False,
))
network = keras.KerasNetwork(input_shape=obs_shape)
trainer.train_epoch(network)