def test_save(self):
obj = mock.MagicMock()
npz.save_npz(self.temp_file_path, obj, self.compress)
self.assertEqual(obj.serialize.call_count, 1)
(serializer,), _ = obj.serialize.call_args
self.assertIsInstance(serializer, npz.DictionarySerializer)
def setUp(self):
if self.file_type == 'filename':
fd, path = tempfile.mkstemp()
os.close(fd)
self.file = path
elif self.file_type == 'bytesio':
self.file = six.BytesIO()
else:
assert False
# Create and save a link
child = link.Chain()
with child.init_scope():
child.linear = links.Linear(2, 3)
parent = link.Chain()
with parent.init_scope():
parent.linear = links.Linear(3, 2)
parent.child = child
npz.save_npz(self.file, parent)
self.source = parent
if self.file_type == 'bytesio':
self.file.seek(0)
self.npzfile = numpy.load(self.file)
self.deserializer = npz.NpzDeserializer(self.npzfile, strict=False)
def test_load_optimizer(self):
for param in self.parent.params():
param.data.fill(1)
npz.save_npz(self.temp_file_path, self.parent, self.compress)
for param in self.parent.params():
param.data.fill(0)
npz.load_npz(self.temp_file_path, self.parent)
for param in self.parent.params():
self.assertTrue((param.data == 1).all())
def setUp(self):
fd, path = tempfile.mkstemp()
os.close(fd)
self.temp_file_path = path
child = link.Chain(child_linear=links.Linear(2, 3))
parent = link.Chain(
parent_linear=links.Linear(3, 2), child=child)
npz.save_npz(path, parent, self.compress)
self.source_child = child
self.source_parent = parent
def test_load_optimizer_without_strict(self):
for param in self.parent.params():
param.data.fill(1)
npz.save_npz(self.temp_file_path, self.parent, self.compress)
# Remove a param
del self.parent.child.linear.b
for param in self.parent.params():
param.data.fill(0)
npz.load_npz(self.temp_file_path, self.parent, strict=False)
for param in self.parent.params():
self.assertTrue((param.data == 1).all())
self.assertFalse(hasattr(self.parent.child.linear, 'b'))
def setUp(self):
fd, path = tempfile.mkstemp()
os.close(fd)
self.temp_file_path = path
child = link.Chain(linear=links.Linear(2, 3))
parent = link.Chain(linear=links.Linear(3, 2), child=child)
npz.save_npz(self.temp_file_path, parent)
self.source = parent
self.npzfile = numpy.load(path)
self.deserializer = npz.NpzDeserializer(self.npzfile, strict=False)
def convert_caffemodel_to_npz(cls, path_caffemodel, path_npz):
"""Converts a pre-trained caffemodel to a chainer model.
Args:
path_caffemodel (str): Path of the pre-trained caffemodel.
path_npz (str): Path of the converted chainer model.
"""
# As CaffeFunction uses shortcut symbols,
# we import CaffeFunction here.
from chainer.links.caffe.caffe_function import CaffeFunction
caffemodel = CaffeFunction(path_caffemodel)
npz.save_npz(path_npz, caffemodel, compression=False)
def test_serialization(self):
lin1 = links.SimplifiedDropconnect(None, self.out_size)
x = chainer.Variable(self.x)
# Must call the link to initialize weights.
lin1(x)
w1 = lin1.W.data
fd, temp_file_path = tempfile.mkstemp()
os.close(fd)
npz.save_npz(temp_file_path, lin1)
lin2 = links.SimplifiedDropconnect(None, self.out_size)
npz.load_npz(temp_file_path, lin2)
w2 = lin2.W.data
self.assertEqual((w1 == w2).all(), True)
def setUp(self):
fd, path = tempfile.mkstemp()
os.close(fd)
self.temp_file_path = path
child = link.Chain()
with child.init_scope():
child.child_linear = links.Linear(2, 3)
parent = link.Chain()
with parent.init_scope():
parent.parent_linear = links.Linear(3, 2)
parent.child = child
npz.save_npz(path, parent, self.compress)
self.source_child = child
self.source_parent = parent
def setUp(self):
fd, path = tempfile.mkstemp()
os.close(fd)
self.temp_file_path = path
child = link.Chain()
with child.init_scope():
child.linear2 = links.Linear(2, 3)
parent = link.Chain()
with parent.init_scope():
parent.linear = links.Linear(3, 2)
parent.child = child
npz.save_npz(self.temp_file_path, parent)
self.source = parent
self.npzfile = numpy.load(path)
self.deserializer = npz.NpzDeserializer(
self.npzfile, ignore_names=self.ignore_names)
def convert_caffemodel_to_npz(cls, path_caffemodel, path_npz, n_layers=50):
"""Converts a pre-trained caffemodel to a chainer model.
Args:
path_caffemodel (str): Path of the pre-trained caffemodel.
path_npz (str): Path of the converted chainer model.
"""
# As CaffeFunction uses shortcut symbols,
# we import CaffeFunction here.
from chainer.links.caffe.caffe_function import CaffeFunction
caffemodel = CaffeFunction(path_caffemodel)
chainermodel = cls(pretrained_model=None, n_layers=n_layers)
if n_layers == 50:
_transfer_resnet50(caffemodel, chainermodel)
elif n_layers == 101:
_transfer_resnet101(caffemodel, chainermodel)
elif n_layers == 152:
_transfer_resnet152(caffemodel, chainermodel)
else:
raise ValueError('The n_layers argument should be either 50, 101,'
' or 152, but {} was given.'.format(n_layers))
npz.save_npz(path_npz, chainermodel, compression=False)
def setUp(self):
if self.file_type == 'filename':
fd, path = tempfile.mkstemp()
os.close(fd)
self.file = path
elif self.file_type == 'bytesio':
self.file = six.BytesIO()
else:
assert False
child = link.Chain()
with child.init_scope():
child.child_linear = links.Linear(2, 3)
parent = link.Chain()
with parent.init_scope():
parent.parent_linear = links.Linear(3, 2)
parent.child = child
npz.save_npz(self.file, parent, self.compress)
if self.file_type == 'bytesio':
self.file.seek(0)
self.source_child = child
self.source_parent = parent
from chainer.links.caffe.caffe_function import CaffeFunction
from chainer.serializers import npz
caffemodel = CaffeFunction('./src/VGG_ILSVRC_19_layers.caffemodel')
npz.save_npz('VGG_ILSVRC_19_layers.npz', caffemodel, compression=False)
def _convert_caffemodel_to_npz(path_caffemodel, path_npz):
# As CaffeFunction uses shortcut symbols,
# we import CaffeFunction here.
from chainer.links.caffe.caffe_function import CaffeFunction
caffemodel = CaffeFunction(path_caffemodel)
npz.save_npz(path_npz, caffemodel, compression=False)
def test_save_optimizer2(self):
npz.save_npz(self.temp_file_path, self.optimizer, self.compress)
with numpy.load(self.temp_file_path) as f:
self._check_optimizer_group(f, ("Wp/msg", "Wp/msdx", "epoch", "t"))
def test_save_chain2(self):
npz.save_npz(self.temp_file_path, self.parent, self.compress)
with numpy.load(self.temp_file_path) as f:
self._check_chain_group(f, ("Wp",))
def caffe2npz(caffe_path):
caffe_model = CaffeFunction(caffe_path)
npz_filename = Path(caffe_path).stem + '.npz'
npz.save_npz(npz_filename, caffe_model, compression=False)
return npz_filename
import argparse
from chainer.links.caffe.caffe_function import CaffeFunction
from chainer.serializers import npz
if __name__ == '__main__':
parser = argparse.ArgumentParser(description='Convert caffemodel to npz')
parser.add_argument('--input_caffemodel',
'-i',
default='VGG_ILSVRC_16_layers.caffemodel')
parser.add_argument('--output_npz',
'-o',
default='VGG_ILSVRC_16_layers.npz')
args = parser.parse_args()
caffemodel = CaffeFunction(args.input_caffemodel)
npz.save_npz(args.output_npz, caffemodel, compression=False)
def test_save_optimizer2(self):
npz.save_npz(self.temp_file_path, self.optimizer, self.compress)
with numpy.load(self.temp_file_path) as f:
self._check_optimizer_group(f, ('Wp/msg', 'Wp/msdx', 'epoch', 't'))
def test_save_chain2(self):
npz.save_npz(self.temp_file_path, self.parent, self.compress)
with numpy.load(self.temp_file_path) as f:
self._check_chain_group(f, ('Wp', ))
def _save_npz(self, file, obj, compress):
npz.save_npz(file, obj, compress)
if self.file_type == 'bytesio':
self.file.seek(0)
def convert_caffemodel_to_npz(path_caffemodel, path_npz):
from chainer.links.caffe.caffe_function import CaffeFunction
caffemodel = CaffeFunction(path_caffemodel)
npz.save_npz(path_npz, caffemodel, compression=False)
def convert_tf_to_chainer(cls, tf_npz_path, chainer_npz_path, depth_mult):
tf_model = np.load(tf_npz_path)
chainermodel = cls(depth_multiplier=depth_mult)
cls._transfer_mobile_net_v1(tf_model, chainermodel)
npz.save_npz(chainer_npz_path, chainermodel, compression=False)
def _save_npz(self, file, obj, compress):
npz.save_npz(file, obj, compress)
if self.file_type == 'bytesio':
self.file.seek(0)
from chainer.serializers import npz
import os, tkinter.filedialog, tkinter, tkinter.messagebox
from chainer.links.caffe import CaffeFunction
root = tkinter.Tk()
root.withdraw()
fTyp = [("", "*")]
iDir = os.path.abspath(os.path.dirname(__file__))
path = tkinter.filedialog.askopenfilename(filetypes=fTyp, initialdir=iDir)
model = CaffeFunction(path)
npz.save_npz('vgg16.npz', model, compression=False)
self.conv3 = L.Convolution2D(None, 384, 3, pad=1)
self.conv4 = L.Convolution2D(None, 384, 3, pad=1)
self.conv5 = L.Convolution2D(None, 256, 3, pad=1)
self.fc6 = L.Linear(None, 4096)
self.fc7 = L.Linear(None, 4096)
self.fc8 = L.Linear(None, nb_class)
def __call__(self, x, t):
h = F.max_pooling_2d(F.local_response_normalization(
F.relu(self.conv1(x))),
3,
stride=2)
h = F.max_pooling_2d(F.local_response_normalization(
F.relu(self.conv2(h))),
3,
stride=2)
h = F.relu(self.conv3(h))
h = F.relu(self.conv4(h))
h = F.max_pooling_2d(F.relu(self.conv5(h)), 3, stride=2)
h = F.dropout(F.relu(self.fc6(h)))
h = F.dropout(F.relu(self.fc7(h)))
h = self.fc8(h)
return h
if __name__ == "__main__":
caffemodel = CaffeFunction("bvlc_alexnet.caffemodel")
npz.save_npz("alexnet.npz", caffemodel, compression=False)
alexnet = Alex()
npz.load_npz("alexnet.npz", alexnet)
def test_save_optimizer2(self):
npz.save_npz(self.temp_file_path, self.optimizer, self.compress)
with numpy.load(self.temp_file_path) as f:
self._check_optimizer_group(f, ('Wp/msg', 'Wp/msdx', 'epoch', 't'))
