def test_sequence_offset() -> None:
seq = utils.make_xor_data_sequences()[0]
offset_seq = keras._SequenceWithOffset(seq, batch_offset=1)
assert len(seq) == len(offset_seq)
for i in range(len(offset_seq)):
a = offset_seq[i]
b = seq[(i + 1) % len(seq)]
assert len(a) == len(b)
for i in range(len(a)):
assert np.equal(a[i], b[i]).all()
def test_adapt_sequence() -> None:
seqs = utils.make_xor_data_sequences()
train = keras._adapt_keras_data(seqs[0], batch_size=1)
assert isinstance(train, keras.SequenceAdapter)
test = keras._adapt_keras_data(seqs[1], batch_size=1)
assert isinstance(test, keras.SequenceAdapter)
assert seqs[0] is train._sequence._sequence
assert seqs[1] is test._sequence._sequence
assert train is keras._adapt_keras_data(train, batch_size=1)
assert test is keras._adapt_keras_data(test, batch_size=1)
"global_batch_size": 4,
"trial_type": "default",
}
}
context = init(
config=config, local=args.local, test=args.test, context_dir=str(pathlib.Path.cwd())
)
model = Sequential()
model.add(Dense(context.get_hparam("hidden_size"), activation="sigmoid", input_shape=(2,)))
model.add(Dense(1))
if args.use_dataset:
data, labels = utils.xor_data()
train = context.wrap_dataset(tf.data.Dataset.from_tensor_slices((data, labels)))
train = train.batch(context.get_hparam("global_batch_size"))
valid = context.wrap_dataset(tf.data.Dataset.from_tensor_slices((data, labels)))
valid = valid.batch(context.get_hparam("global_batch_size"))
else:
train, valid = utils.make_xor_data_sequences(batch_size=4)
model = context.wrap_model(model)
model.compile(
SGD(lr=context.get_hparam("learning_rate")),
binary_crossentropy,
metrics=[categorical_error],
)
model.fit(x=train, steps_per_epoch=100, validation_data=valid, workers=0)
def build_validation_data_loader(self) -> keras.InputData:
_, test = make_xor_data_sequences(batch_size=4)
return keras.SequenceAdapter(test, workers=0)
def build_training_data_loader(self) -> keras.InputData:
train, _ = make_xor_data_sequences(batch_size=4)
return keras.SequenceAdapter(train, workers=0)
def test_adapt_invalid_data_type() -> None:
seqs = utils.make_xor_data_sequences()
test = keras._adapt_data_from_data_loader(seqs[1], batch_size=1)
with pytest.raises(det.errors.InvalidDataTypeException) as err:
keras._adapt_data_from_data_loader((None, test), batch_size=1)
assert err is not None
import determined as det
from determined import keras
from determined_common import check
from tests.experiment import utils # noqa: I100
class Empty(Sequence):
def __getitem__(self, index: int) -> list:
return []
def __len__(self) -> int:
return 0
SEQ = utils.make_xor_data_sequences()[0]
MULTITHREADING_MULTIPROCESS_SUITE = [
(0, False, SEQ),
(1, False, SEQ),
(1, True, SEQ),
(2, False, SEQ),
(2, True, SEQ),
]
def test_sequence_offset() -> None:
seq = utils.make_xor_data_sequences()[0]
offset_seq = keras._SequenceWithOffset(seq, batch_offset=1)
assert len(seq) == len(offset_seq)
for i in range(len(offset_seq)):
def build_validation_data_loader(self) -> keras.InputData:
_, test = make_xor_data_sequences(batch_size=4)
return test
def build_training_data_loader(self) -> keras.InputData:
train, _ = make_xor_data_sequences(batch_size=4)
return train