def test_copyparams(self):
l1 = chainer.Link()
with l1.init_scope():
l1.x = chainer.Parameter(shape=(2, 3))
l2 = chainer.Link()
with l2.init_scope():
l2.x = chainer.Parameter(shape=2)
l3 = chainer.Link()
with l3.init_scope():
l3.x = chainer.Parameter(shape=3)
c1 = chainer.Chain()
with c1.init_scope():
c1.l1 = l1
c1.l2 = l2
c2 = chainer.Chain()
with c2.init_scope():
c2.c1 = c1
c2.l3 = l3
l1.x.data.fill(0)
l2.x.data.fill(1)
l3.x.data.fill(2)
self.c2.copyparams(c2)
numpy.testing.assert_array_equal(self.l1.x.data, l1.x.data)
numpy.testing.assert_array_equal(self.l2.x.data, l2.x.data)
numpy.testing.assert_array_equal(self.l3.x.data, l3.x.data)
def test_addgrads(self):
l1 = chainer.Link()
with l1.init_scope():
l1.x = chainer.Parameter(shape=(2, 3))
l1.y = chainer.Parameter(shape=(2, 3))
l2 = chainer.Link()
with l2.init_scope():
l2.x = chainer.Parameter(shape=2)
l3 = chainer.Link()
with l3.init_scope():
l3.x = chainer.Parameter(shape=3)
c1 = chainer.ChainList(l1, l2)
c2 = chainer.ChainList(c1, l3)
l1.x.grad.fill(1)
l2.x.grad.fill(2)
l3.x.grad.fill(3)
l1.y.grad.fill(4)
self.l1.x.grad.fill(-1)
self.l1.y.cleargrad()
self.l2.x.grad.fill(-2)
self.l3.x.grad.fill(-3)
self.c2.addgrads(c2)
numpy.testing.assert_array_equal(self.l1.x.grad, numpy.zeros((2, 3)))
numpy.testing.assert_array_equal(self.l1.y.grad, l1.y.grad)
numpy.testing.assert_array_equal(self.l2.x.grad, numpy.zeros(2))
numpy.testing.assert_array_equal(self.l3.x.grad, numpy.zeros(3))
def test_addgrads(self):
l1 = chainer.Link()
with l1.init_scope():
l1.x = chainer.Parameter(shape=(2, 3))
l2 = chainer.Link()
with l2.init_scope():
l2.x = chainer.Parameter(shape=2)
l3 = chainer.Link()
with l3.init_scope():
l3.x = chainer.Parameter(shape=3)
c1 = chainer.Chain()
with c1.init_scope():
c1.l1 = l1
c1.l2 = l2
c2 = chainer.Chain()
with c2.init_scope():
c2.c1 = c1
c2.l3 = l3
l1.x.grad.fill(1)
l2.x.grad.fill(2)
l3.x.grad.fill(3)
self.l1.x.grad.fill(-1)
self.l2.x.grad.fill(-2)
self.l3.cleargrads()
self.c2.addgrads(c2)
numpy.testing.assert_array_equal(self.l1.x.grad, numpy.zeros((2, 3)))
numpy.testing.assert_array_equal(self.l2.x.grad, numpy.zeros(2))
numpy.testing.assert_array_equal(self.l3.x.grad, numpy.full(3, 3.))
def setUp(self):
self.l1 = chainer.Link(x=(2, 3))
self.l2 = chainer.Link(x=2)
self.l3 = chainer.Link(x=3)
self.c1 = chainer.ChainList(self.l1)
self.c1.add_link(self.l2)
self.c2 = chainer.ChainList(self.c1, self.l3)
def test_namedpersistent():
# This test case is adopted from
# https://github.com/chainer/chainer/pull/6788
l1 = chainer.Link()
with l1.init_scope():
l1.x = chainer.Parameter(shape=(2, 3))
l2 = chainer.Link()
with l2.init_scope():
l2.x = chainer.Parameter(shape=2)
l2.add_persistent('l2_a', numpy.array([1, 2, 3], dtype=numpy.float32))
l3 = chainer.Link()
with l3.init_scope():
l3.x = chainer.Parameter()
l3.add_persistent('l3_a', numpy.array([1, 2, 3], dtype=numpy.float32))
c1 = chainer.Chain()
with c1.init_scope():
c1.l1 = l1
c1.add_link('l2', l2)
c1.add_persistent('c1_a', numpy.array([1, 2, 3], dtype=numpy.float32))
c2 = chainer.Chain()
with c2.init_scope():
c2.c1 = c1
c2.l3 = l3
c2.add_persistent('c2_a', numpy.array([1, 2, 3], dtype=numpy.float32))
namedpersistent = list(chainerrl.misc.namedpersistent(c2))
assert ([(name, id(p)) for name, p in namedpersistent
] == [('/c2_a', id(c2.c2_a)), ('/c1/c1_a', id(c2.c1.c1_a)),
('/c1/l2/l2_a', id(c2.c1.l2.l2_a)),
('/l3/l3_a', id(c2.l3.l3_a))])
def setUp(self):
self.l1 = chainer.Link(x=(2, 3))
self.l1.add_uninitialized_param('y')
self.l2 = chainer.Link(x=2)
self.l3 = chainer.Link(x=3)
self.c1 = chainer.ChainList(self.l1)
self.c1.add_link(self.l2)
self.c2 = chainer.ChainList(self.c1, self.l3)
def setUp(self):
self.l1 = chainer.Link(x=(2, 3), y=None)
self.l2 = chainer.Link(x=2)
self.l3 = chainer.Link(x=3)
self.c1 = chainer.ChainList(self.l1)
self.c1.add_link(self.l2)
self.c2 = chainer.ChainList(self.c1)
self.c2.append(self.l3)
def setUp(self):
self.l1 = chainer.Link(x=(2, 3))
self.l2 = chainer.Link(x=2)
self.l3 = chainer.Link(x=3)
self.c1 = chainer.Chain(l1=self.l1)
self.c1.add_link('l2', self.l2)
self.c2 = chainer.Chain(c1=self.c1, l3=self.l3)
def setUp(self):
if self.accfun is None:
self.link = SklearnWrapperClassifier(chainer.Link())
else:
self.link = SklearnWrapperClassifier(chainer.Link(),
accfun=self.accfun)
self.link.compute_accuracy = self.compute_accuracy
self.x = numpy.random.uniform(-1, 1, (5, 10)).astype(numpy.float32)
self.t = numpy.random.randint(3, size=5).astype(numpy.int32)
def setUp(self):
self.l1 = chainer.Link(x=(2, 3))
self.l2 = chainer.Link(x=2)
self.l3 = chainer.Link(x=None)
self.c1 = chainer.Chain(l1=self.l1)
self.c1.add_link('l2', self.l2)
self.c2 = chainer.Chain(c1=self.c1)
with self.c2.init_scope():
self.c2.l3 = self.l3
def setUp(self):
self.l1 = chainer.Link()
with self.l1.init_scope():
self.l1.x = chainer.Parameter(shape=(2, 3))
self.l1.y = chainer.Parameter()
self.l2 = chainer.Link()
with self.l2.init_scope():
self.l2.x = chainer.Parameter(shape=2)
self.l3 = chainer.Link()
with self.l3.init_scope():
self.l3.x = chainer.Parameter(shape=3)
self.c1 = chainer.ChainList(self.l1)
self.c1.add_link(self.l2)
self.c2 = chainer.ChainList(self.c1)
self.c2.append(self.l3)
def test_copyparams(self):
l1 = chainer.Link(x=(2, 3))
l2 = chainer.Link(x=2)
l3 = chainer.Link(x=3)
c1 = chainer.ChainList(l1, l2)
c2 = chainer.ChainList(c1, l3)
l1.x.data.fill(0)
l2.x.data.fill(1)
l3.x.data.fill(2)
self.c2.copyparams(c2)
numpy.testing.assert_array_equal(self.l1.x.data, l1.x.data)
numpy.testing.assert_array_equal(self.l2.x.data, l2.x.data)
numpy.testing.assert_array_equal(self.l3.x.data, l3.x.data)
def __init__(self, n_hidden=512, ch=512,
channel_evolution=(512, 512, 512, 512, 256, 128, 64, 32, 16), conditional=False):
super(ProgressiveGenerator, self).__init__()
self.n_hidden = n_hidden
self.max_stage = (len(channel_evolution) - 1) * 2
with self.init_scope():
self.c0 = EqualizedConv2d(n_hidden, ch, 4, 1, 3)
if conditional:
self.c1 = EqualizedConv2d(ch, ch, 3, 1, 1)
else:
self.c1 = EqualizedConv2d(ch, ch, 3, 1, 1)
bs = [
chainer.Link() # dummy
]
outs = [
]
if conditional:
outs.append(EqualizedConv2d(channel_evolution[0], 3, 1, 1, 0))
else:
outs.append(EqualizedConv2d(channel_evolution[0], 3, 1, 1, 0))
for i in range(1, len(channel_evolution)):
if conditional:
bs.append(GeneratorBlock(channel_evolution[i - 1] * 2, channel_evolution[i]))
outs.append(EqualizedConv2d(channel_evolution[i], 3, 1, 1, 0))
else:
bs.append(GeneratorBlock(channel_evolution[i - 1], channel_evolution[i]))
outs.append(EqualizedConv2d(channel_evolution[i], 3, 1, 1, 0))
self.bs = chainer.ChainList(*bs)
self.outs = chainer.ChainList(*outs)
def test_has_uninitialized_params(self):
l = chainer.Link(y=2)
self.assertFalse(l.has_uninitialized_params)
l.add_uninitialized_param('x')
self.assertTrue(l.has_uninitialized_params)
l.add_param('x', (2, 3))
self.assertFalse(l.has_uninitialized_params)
def setUp(self):
self.target = chainer.Link()
self.iterator = DummyIterator([(numpy.array(1), numpy.array(2))])
self.optimizer = DummyOptimizer()
self.optimizer.setup(self.target)
self.updater = training.updaters.StandardUpdater(
self.iterator, self.optimizer)
def test():
model = chainer.Link()
dataset = [((numpy.ones(
(2, 5, 5)) * i).astype(numpy.float32), numpy.int32(0))
for i in range(100)]
batch_size = 5
devices = (0, )
iters = [
chainer.iterators.SerialIterator(i, batch_size)
for i in chainer.datasets.split_dataset_n_random(
dataset, len(devices))
]
optimizer = chainer.optimizers.SGD(lr=1.0)
optimizer.setup(model)
# Initialize CUDA context.
cuda.cupy.cuda.runtime.runtimeGetVersion()
try:
mpu.MultiprocessParallelUpdater(iters, optimizer, devices=devices)
except RuntimeError as e:
assert 'CUDA context' in str(e)
return
assert False
def check_call(self, gpu, label_key, args, kwargs, model_args,
model_kwargs):
init_kwargs = {'label_key': label_key}
if self.accfun is not None:
init_kwargs['accfun'] = self.accfun
link = links.Classifier(chainer.Link(), **init_kwargs)
if gpu:
xp = cuda.cupy
link.to_gpu()
else:
xp = numpy
link.compute_accuracy = self.compute_accuracy
y = chainer.Variable(self.y)
link.predictor = mock.MagicMock(return_value=y)
loss = link(*args, **kwargs)
link.predictor.assert_called_with(*model_args, **model_kwargs)
self.assertTrue(hasattr(link, 'y'))
self.assertIsNotNone(link.y)
self.assertTrue(hasattr(link, 'loss'))
xp.testing.assert_allclose(link.loss.data, loss.data)
self.assertTrue(hasattr(link, 'accuracy'))
if self.compute_accuracy:
self.assertIsNotNone(link.accuracy)
else:
self.assertIsNone(link.accuracy)
def test_device(self, model_initial_backend_config, model_backend_config,
input_backend_config):
model_initial_device = model_initial_backend_config.device
device = model_backend_config.device
input_device = input_backend_config.device
model = chainer.Link()
model.to_device(model_initial_device)
optimizer = DummyOptimizer()
optimizer.setup(model)
iterator = DummyIterator([numpy.array(1), numpy.array(2)])
updater = training.updaters.StandardUpdater(iterator,
optimizer,
device=device,
input_device=input_device)
assert updater.device is device
assert updater.input_device is input_device
# Check the model device.
assert model.device == device
updater.update_core()
assert optimizer.update.call_count == 1
args, kwargs = optimizer.update.call_args
assert len(args) == 2
assert len(kwargs) == 0
loss, v1 = args
# Check the input device.
assert backend.get_device_from_array(v1) == input_device
def test_link_as_torch_model():
# initialized parameter
a_arr = numpy.ones((3, 2), 'float32')
a_chainer_param = chainer.Parameter(a_arr)
# 0-size parameter
b_arr = numpy.ones((2, 0, 1), 'float32')
b_chainer_param = chainer.Parameter(b_arr)
link = chainer.Link()
with link.init_scope():
link.a = a_chainer_param
link.b = b_chainer_param
# Conversion
torched = cpm.LinkAsTorchModel(link)
params = list(torched.parameters())
assert len(params) == 2
assert isinstance(params[0], torch.nn.Parameter)
assert isinstance(params[1], torch.nn.Parameter)
assert params[0].shape == (3, 2)
assert params[1].shape == (2, 0, 1)
assert (params[0].data.numpy() == numpy.ones((3, 2))).all()
# Test memory sharing
params[0].data[...] = torch.tensor(numpy.arange(6).reshape((3, 2)))
assert (a_chainer_param.array == numpy.arange(6).reshape((3, 2))).all()
def setUp(self):
x_shape_0 = 2
x_shape_1 = numpy.int64(3)
self.link = chainer.Link(x=((x_shape_0, x_shape_1), 'd'), y=2)
self.p = numpy.array([1, 2, 3], dtype='f')
self.link.add_persistent('p', self.p)
self.link.name = 'a'
def __init__(self,
ch=512,
pooling_comp=1.0,
channel_evolution=(512, 512, 512, 512, 256, 128, 64, 32, 16),
first_channel=3,
use_both_conditional_and_latent=False):
super().__init__()
self.use_both_conditional_and_latent = use_both_conditional_and_latent
self.max_stage = (len(channel_evolution) - 1) * 2
self.pooling_comp = pooling_comp # compensation of ave_pool is 0.5-Lipshitz
with self.init_scope():
ins = [
EqualizedConv2d(first_channel, channel_evolution[0], 1, 1, 0)
]
bs = [
chainer.Link() # dummy
]
for i in range(1, len(channel_evolution)):
ins.append(
EqualizedConv2d(first_channel, channel_evolution[i], 1, 1,
0))
bs.append(
DiscriminatorBlock(channel_evolution[i],
channel_evolution[i - 1], pooling_comp))
self.ins = chainer.ChainList(*ins)
self.bs = chainer.ChainList(*bs)
self.out0 = EqualizedConv2d(ch + 1, ch, 3, 1, 1)
self.out1 = EqualizedConv2d(ch, ch, 4, 1, 0)
self.out2 = EqualizedLinear(ch, 1)
def check_call(self, gpu, label_key, args, kwargs, model_args,
model_kwargs, metrics_fun, compute_metrics):
init_kwargs = {'label_key': label_key}
if metrics_fun is not None:
init_kwargs['metrics_fun'] = metrics_fun
link = Classifier(chainer.Link(), **init_kwargs)
if gpu:
xp = cuda.cupy
link.to_gpu()
else:
xp = numpy
link.compute_metrics = compute_metrics
y = chainer.Variable(self.y)
link.predictor = mock.MagicMock(return_value=y)
loss = link(*args, **kwargs)
link.predictor.assert_called_with(*model_args, **model_kwargs)
assert hasattr(link, 'y')
assert link.y is not None
assert hasattr(link, 'loss')
xp.testing.assert_allclose(link.loss.data, loss.data)
assert hasattr(link, 'metrics')
if compute_metrics:
assert link.metrics is not None
else:
assert link.metrics is None
def test_copyparams(self):
self.link.x.grad.fill(0)
self.link.y.grad.fill(1)
self.link.u.initialize((2, 3))
self.link.u.data.fill(0)
self.link.u.grad.fill(1)
self.link.v.zerograd()
gx = self.link.x.grad.copy()
gy = self.link.y.grad.copy()
gu = self.link.u.grad.copy()
l = chainer.Link(x=(2, 3), y=2, u=(2, 3), v=(3, 2))
l.x.data.fill(2)
l.x.grad.fill(3)
l.y.data.fill(4)
l.y.grad.fill(5)
l.u.data.fill(6)
l.u.grad.fill(7)
l.v.data.fill(8)
l.v.grad.fill(9)
self.link.copyparams(l)
numpy.testing.assert_array_equal(self.link.x.data, l.x.data)
numpy.testing.assert_array_equal(self.link.x.grad, gx)
numpy.testing.assert_array_equal(self.link.y.data, l.y.data)
numpy.testing.assert_array_equal(self.link.y.grad, gy)
numpy.testing.assert_array_equal(self.link.u.data, l.u.data)
numpy.testing.assert_array_equal(self.link.u.grad, gu)
numpy.testing.assert_array_equal(self.link.v.data, l.v.data)
numpy.testing.assert_array_equal(self.link.v.grad, 0)
def test_addgrads(self):
l = chainer.Link()
with l.init_scope():
l.x = chainer.Parameter(shape=(2, 3),
initializer=initializers.NaN('d'))
l.y = chainer.Parameter(shape=2)
l.u = chainer.Parameter(shape=(2, 3))
l.v = chainer.Parameter()
l.x.grad.fill(1)
l.y.grad.fill(2)
l.u.grad.fill(3)
self.link.x.grad.fill(-1)
self.link.y.grad.fill(-2)
self.link.u.cleargrad()
self.link.addgrads(l)
gx_expect = numpy.zeros_like(l.x.grad)
gy_expect = numpy.zeros_like(l.y.grad)
gu_expect = l.u.grad
numpy.testing.assert_array_equal(self.link.x.grad, gx_expect)
numpy.testing.assert_array_equal(self.link.y.grad, gy_expect)
numpy.testing.assert_array_equal(self.link.u.grad, gu_expect)
self.assertIsNone(self.link.v.grad, None)
def test_copyparams(self):
self.link.x.grad.fill(0)
self.link.y.grad.fill(1)
self.link.u.initialize((2, 3))
self.link.u.data.fill(0)
self.link.u.grad.fill(1)
self.link.v.cleargrad()
gx = self.link.x.grad.copy()
gy = self.link.y.grad.copy()
gu = self.link.u.grad.copy()
l = chainer.Link()
with l.init_scope():
l.x = chainer.Parameter(shape=(2, 3))
l.y = chainer.Parameter(shape=2)
l.u = chainer.Parameter(shape=(2, 3))
l.v = chainer.Parameter(shape=(3, 2))
l.x.data.fill(2)
l.x.grad.fill(3)
l.y.data.fill(4)
l.y.grad.fill(5)
l.u.data.fill(6)
l.u.grad.fill(7)
l.v.data.fill(8)
l.v.grad.fill(9)
self.link.copyparams(l)
numpy.testing.assert_array_equal(self.link.x.data, l.x.data)
numpy.testing.assert_array_equal(self.link.x.grad, gx)
numpy.testing.assert_array_equal(self.link.y.data, l.y.data)
numpy.testing.assert_array_equal(self.link.y.grad, gy)
numpy.testing.assert_array_equal(self.link.u.data, l.u.data)
numpy.testing.assert_array_equal(self.link.u.grad, gu)
numpy.testing.assert_array_equal(self.link.v.data, l.v.data)
numpy.testing.assert_array_equal(self.link.v.grad, None)
def __init__(self,
n_layers,
in_size,
out_size,
dropout_rate,
name="",
use_cudnn=True):
weights = []
direction = 1 # ここでは，からなず一方向ずつ構築するので1にする
t_name = name
if name is not "":
t_name = '%s_' % name
for i in range(n_layers):
for di in range(direction):
weight = chainer.Link()
for j in range(8):
if i == 0 and j < 4:
w_in = in_size
elif i > 0 and j < 4:
w_in = out_size * direction
else:
w_in = out_size
weight.add_param('%sw%d' % (t_name, j), (out_size, w_in))
weight.add_param('%sb%d' % (t_name, j), (out_size, ))
getattr(weight, '%sw%d' %
(t_name, j)).data[...] = np.random.normal(
0, np.sqrt(1. / w_in), (out_size, w_in))
getattr(weight, '%sb%d' % (t_name, j)).data[...] = 0
weights.append(weight)
super(NStepLSTMpp, self).__init__(*weights)
self.n_layers = n_layers
self.dropout_rate = dropout_rate
self.use_cudnn = use_cudnn
self.out_size = out_size
self.direction = direction
self.ws = [[
getattr(w, '%sw0' % t_name),
getattr(w, '%sw1' % t_name),
getattr(w, '%sw2' % t_name),
getattr(w, '%sw3' % t_name),
getattr(w, '%sw4' % t_name),
getattr(w, '%sw5' % t_name),
getattr(w, '%sw6' % t_name),
getattr(w, '%sw7' % t_name)
] for w in self]
self.bs = [[
getattr(w, '%sb0' % t_name),
getattr(w, '%sb1' % t_name),
getattr(w, '%sb2' % t_name),
getattr(w, '%sb3' % t_name),
getattr(w, '%sb4' % t_name),
getattr(w, '%sb5' % t_name),
getattr(w, '%sb6' % t_name),
getattr(w, '%sb7' % t_name)
] for w in self]
def setUp(self):
x_shape_0 = 2
x_shape_1 = numpy.int64(3)
self.link = chainer.Link(x=((x_shape_0, x_shape_1), 'd'), y=2)
self.p = numpy.array([1, 2, 3], dtype='f')
self.link.add_persistent('p', self.p)
self.link.name = 'a'
if cuda.available:
self.current_device_id = cuda.cupy.cuda.get_device_id()
def test_copyparams_uninitialized(self):
l = chainer.Link(x=(2, 3))
l.add_uninitialized_param('y')
self.link.x.data.fill(2)
self.link.y.data.fill(4)
l.copyparams(self.link)
numpy.testing.assert_array_equal(l.x.data, self.link.x.data)
self.assertTrue(hasattr(l, 'y'))
numpy.testing.assert_array_equal(l.y.data, self.link.y.data)
def test_addgrads(self):
l1 = chainer.Link(x=(2, 3))
l2 = chainer.Link(x=2)
l3 = chainer.Link(x=3)
c1 = chainer.ChainList(l1, l2)
c2 = chainer.ChainList(c1, l3)
l1.x.grad.fill(1)
l2.x.grad.fill(2)
l3.x.grad.fill(3)
self.l1.x.grad.fill(-1)
self.l2.x.grad.fill(-2)
self.l3.x.grad.fill(-3)
self.c2.addgrads(c2)
numpy.testing.assert_array_equal(self.l1.x.grad, numpy.zeros((2, 3)))
numpy.testing.assert_array_equal(self.l2.x.grad, numpy.zeros(2))
numpy.testing.assert_array_equal(self.l3.x.grad, numpy.zeros(3))
def test_serialize_deserialize_to_uninitialized_param(self):
ret = numpy.random.rand(2, 3).astype('f')
serializer = mock.MagicMock(return_value=ret)
l = chainer.Link(x=None)
l.serialize(serializer)
self.assertEqual(serializer.call_count, 1)
serializer.assert_any_call('x', None)
self.assertIsInstance(l.x.data, numpy.ndarray)
numpy.testing.assert_array_equal(l.x.data, ret)