if args.resume:
model.optimizer = opt
model.optimizer.optimize(model.layers_to_optimize, epoch=model.epoch_index)
# run fprop again as a measure of the model state
out_fprop = model.fprop(im)
out_fprop_save2 = [x.get() for x in out_fprop]
if not args.resume:
save_obj([out_fprop_save, out_fprop_save2], 'serial_test_out1.pkl')
else:
# load up the saved file and compare
run1 = load_obj('serial_test_out1.pkl')
# compare the initial fprops
for x, y in zip(run1[0], out_fprop_save):
assert np.max(np.abs(
x - y)) == 0.0, 'Deserialized model not matching serialized model'
# and post extra training fprops
for x, y in zip(run1[1], out_fprop_save2):
if np.max(np.abs(x - y)) != 0.0:
neon_logger.error('Max Diff: {}'.format(np.max(np.abs(x - y))))
raise ValueError(
'Deserialized training not matching serialized training')
# see if the single epoch of optimization had any real effect
for x, y in zip(out_fprop_save, out_fprop_save2):
assert np.max(np.abs(x - y)) > 0.0, 'Training had no effect on model'
neon_logger.display('passed')
# You may obtain a copy of the License at
#
# http://www.apache.org/licenses/LICENSE-2.0
#
# Unless required by applicable law or agreed to in writing, software
# distributed under the License is distributed on an "AS IS" BASIS,
# WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
# See the License for the specific language governing permissions and
# limitations under the License.
# ----------------------------------------------------------------------------
import sys
from neon import logger as neon_logger
from neon.data.dataloaderadapter import DataLoaderAdapter
try:
from aeon import DataLoader as AeonLoader
except ImportError:
neon_logger.error('Unable to load Aeon data loading module.')
neon_logger.error('Please follow installation instructions at:')
neon_logger.error('https://github.com/NervanaSystems/aeon')
sys.exit(1)
def AeonDataLoader(config, adapter=True):
if adapter:
return DataLoaderAdapter(AeonLoader(config))
else:
return AeonLoader(config)
"""
Helper function to compare two serialized model files
This is only comparing the model weights and states and layer
config parameters
Returns:
bool: True if the two file match
"""
models = []
for fn in [file1, file2]:
assert os.path.exists(fn), 'Could not find file %s' % fn
with open(fn, 'r') as fid:
models.append(ModelDescription(pickle.load(fid)))
return models[0] == models[1]
if __name__ == '__main__':
parser = argparse.ArgumentParser(description='Compare two serialized model files.')
parser.add_argument('file1')
parser.add_argument('file2')
args = parser.parse_args()
if not compare_files(args.file1, args.file2):
neon_logger.error('Models do not match!')
sys.exit(1)
else:
neon_logger.display('Models match')
delta = model.cost.get_errors(im, l)
model.bprop(delta)
if args.resume:
model.optimizer = opt
model.optimizer.optimize(model.layers_to_optimize, epoch=model.epoch_index)
# run fprop again as a measure of the model state
out_fprop = model.fprop(im)
out_fprop_save2 = [x.get() for x in out_fprop]
if not args.resume:
save_obj([out_fprop_save, out_fprop_save2], 'serial_test_out1.pkl')
else:
# load up the saved file and compare
run1 = load_obj('serial_test_out1.pkl')
# compare the initial fprops
for x, y in zip(run1[0], out_fprop_save):
assert np.max(np.abs(x - y)) == 0.0, 'Deserialized model not matching serialized model'
# and post extra training fprops
for x, y in zip(run1[1], out_fprop_save2):
if np.max(np.abs(x - y)) != 0.0:
neon_logger.error('Max Diff: {}'.format(np.max(np.abs(x - y))))
raise ValueError('Deserialized training not matching serialized training')
# see if the single epoch of optimization had any real effect
for x, y in zip(out_fprop_save, out_fprop_save2):
assert np.max(np.abs(x - y)) > 0.0, 'Training had no effect on model'
neon_logger.display('passed')
# you may not use this file except in compliance with the License.
# You may obtain a copy of the License at
#
# http://www.apache.org/licenses/LICENSE-2.0
#
# Unless required by applicable law or agreed to in writing, software
# distributed under the License is distributed on an "AS IS" BASIS,
# WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
# See the License for the specific language governing permissions and
# limitations under the License.
# ----------------------------------------------------------------------------
import sys
from neon import logger as neon_logger
from neon.data.dataloaderadapter import DataLoaderAdapter
try:
from aeon import DataLoader as AeonLoader
except ImportError:
neon_logger.error('Unable to load Aeon data loading module.')
neon_logger.error('Please follow installation instructions at:')
neon_logger.error('https://github.com/NervanaSystems/aeon')
sys.exit(1)
def AeonDataLoader(config, adapter=True):
if adapter:
return DataLoaderAdapter(AeonLoader(config))
else:
return AeonLoader(config)
def __init__(self, *args, **kwargs):
logger.error('DataIterator class has been deprecated and renamed'
'"ArrayIterator" please use that name.')
super(DataIterator, self).__init__(*args, **kwargs)
def __init__(self, *args, **kwargs):
logger.error('DataIterator class has been deprecated and renamed'
'"ArrayIterator" please use that name.')
super(DataIterator, self).__init__(*args, **kwargs)
