Skip to content

Instantly share code, notes, and snippets.

@ikeyasu

ikeyasu/convert_to_onnx.py

Created June 23, 2018 12:11
Show Gist options
  • Star 0 You must be signed in to star a gist
  • Fork 0 You must be signed in to fork a gist
  • Save ikeyasu/971ece2e32ddf36bcc40a71762f68cd9 to your computer and use it in GitHub Desktop.
Save ikeyasu/971ece2e32ddf36bcc40a71762f68cd9 to your computer and use it in GitHub Desktop.
Download ZIP
Raw
convert_to_onnx.py
#!/usr/bin/env python
from __future__ import print_function
import chainer
import chainer.functions as F
import chainer.links as L
from chainer import training
from chainer.training import extensions
import onnx_chainer
import numpy as np
# Network definition
class VGG(chainer.Chain):
def __init__(self):
super(VGG, self).__init__()
with self.init_scope():
self.conv1_1=L.Convolution2D(3, 64, 3, stride=1, pad=1)
self.conv1_2=L.Convolution2D(64, 64, 3, stride=1, pad=1)
self.conv2_1=L.Convolution2D(64, 128, 3, stride=1, pad=1)
self.conv2_2=L.Convolution2D(128, 128, 3, stride=1, pad=1)
self.conv3_1=L.Convolution2D(128, 256, 3, stride=1, pad=1)
self.conv3_2=L.Convolution2D(256, 256, 3, stride=1, pad=1)
self.conv3_3=L.Convolution2D(256, 256, 3, stride=1, pad=1)
self.conv4_1=L.Convolution2D(256, 512, 3, stride=1, pad=1)
self.conv4_2=L.Convolution2D(512, 512, 3, stride=1, pad=1)
self.conv4_3=L.Convolution2D(512, 512, 3, stride=1, pad=1)
self.conv5_1=L.Convolution2D(512, 512, 3, stride=1, pad=1)
self.conv5_2=L.Convolution2D(512, 512, 3, stride=1, pad=1)
self.conv5_3=L.Convolution2D(512, 512, 3, stride=1, pad=1)
self.fc6=L.Linear(25088, 4096)
self.fc7=L.Linear(4096, 4096)
self.fc8=L.Linear(4096, 1000)
def __call__(self, x):
h = F.relu(self.conv1_1(x))
h = F.relu(self.conv1_2(h))
h = F.max_pooling_2d(h, 2, stride=2)
h = F.relu(self.conv2_1(h))
h = F.relu(self.conv2_2(h))
h = F.max_pooling_2d(h, 2, stride=2)
h = F.relu(self.conv3_1(h))
h = F.relu(self.conv3_2(h))
h = F.relu(self.conv3_3(h))
h = F.max_pooling_2d(h, 2, stride=2)
h = F.relu(self.conv4_1(h))
h = F.relu(self.conv4_2(h))
h = F.relu(self.conv4_3(h))
h = F.max_pooling_2d(h, 2, stride=2)
h = F.relu(self.conv5_1(h))
h = F.relu(self.conv5_2(h))
h = F.relu(self.conv5_3(h))
h = F.max_pooling_2d(h, 2, stride=2)
h = F.relu(self.fc6(h))
h = F.relu(self.fc7(h))
h = self.fc8(h)
return h
def main():
model = VGG()
#import pdb; pdb.set_trace()
data = np.zeros((1, 3, 224, 224), dtype=np.float32)
with chainer.using_config('train', False):
onnx_model = onnx_chainer.export(model, data, filename='vgg16.onnx')
if __name__ == '__main__':
main()
Raw
predict.py
#!/usr/bin/env python
from __future__ import print_function
import argparse
import chainer
import chainer.functions as F
import chainer.links as L
from chainer import training
from chainer.training import extensions
import numpy as np
import tqdm
import time
# Network definition
class VGG(chainer.Chain):
def __init__(self):
super(VGG, self).__init__()
with self.init_scope():
self.conv1_1=L.Convolution2D(3, 64, 3, stride=1, pad=1)
self.conv1_2=L.Convolution2D(64, 64, 3, stride=1, pad=1)
self.conv2_1=L.Convolution2D(64, 128, 3, stride=1, pad=1)
self.conv2_2=L.Convolution2D(128, 128, 3, stride=1, pad=1)
self.conv3_1=L.Convolution2D(128, 256, 3, stride=1, pad=1)
self.conv3_2=L.Convolution2D(256, 256, 3, stride=1, pad=1)
self.conv3_3=L.Convolution2D(256, 256, 3, stride=1, pad=1)
self.conv4_1=L.Convolution2D(256, 512, 3, stride=1, pad=1)
self.conv4_2=L.Convolution2D(512, 512, 3, stride=1, pad=1)
self.conv4_3=L.Convolution2D(512, 512, 3, stride=1, pad=1)
self.conv5_1=L.Convolution2D(512, 512, 3, stride=1, pad=1)
self.conv5_2=L.Convolution2D(512, 512, 3, stride=1, pad=1)
self.conv5_3=L.Convolution2D(512, 512, 3, stride=1, pad=1)
self.fc6=L.Linear(25088, 4096)
self.fc7=L.Linear(4096, 4096)
self.fc8=L.Linear(4096, 1000)
def __call__(self, x):
h = F.relu(self.conv1_1(x))
h = F.relu(self.conv1_2(h))
h = F.max_pooling_2d(h, 2, stride=2)
h = F.relu(self.conv2_1(h))
h = F.relu(self.conv2_2(h))
h = F.max_pooling_2d(h, 2, stride=2)
h = F.relu(self.conv3_1(h))
h = F.relu(self.conv3_2(h))
h = F.relu(self.conv3_3(h))
h = F.max_pooling_2d(h, 2, stride=2)
h = F.relu(self.conv4_1(h))
h = F.relu(self.conv4_2(h))
h = F.relu(self.conv4_3(h))
h = F.max_pooling_2d(h, 2, stride=2)
h = F.relu(self.conv5_1(h))
h = F.relu(self.conv5_2(h))
h = F.relu(self.conv5_3(h))
h = F.max_pooling_2d(h, 2, stride=2)
h = F.relu(self.fc6(h))
h = F.relu(self.fc7(h))
h = self.fc8(h)
return h
def main():
model = VGG()
nb_itr = 20
timings = []
for i in tqdm.tqdm(range(nb_itr)):
data = np.random.randn(1, 3, 224, 224).astype(np.float32)
start_time = time.time()
with chainer.using_config('train', False):
with chainer.using_config('enable_backprop', False):
ret = F.softmax(model(chainer.Variable(data)))
timings.append(time.time() - start_time)
print('%10s : %f (sd %f)'% ('mxnet-vgg-16', np.array(timings).mean(), np.array(timings).std()))
if __name__ == '__main__':
main()
Raw
predict_ideep.py
#!/usr/bin/env python
from __future__ import print_function
import argparse
import chainer
import chainer.functions as F
import chainer.links as L
from chainer import training
from chainer.training import extensions
import numpy as np
import tqdm
import time
# Network definition
class VGG(chainer.Chain):
def __init__(self):
super(VGG, self).__init__()
with self.init_scope():
self.conv1_1=L.Convolution2D(3, 64, 3, stride=1, pad=1)
self.conv1_2=L.Convolution2D(64, 64, 3, stride=1, pad=1)
self.conv2_1=L.Convolution2D(64, 128, 3, stride=1, pad=1)
self.conv2_2=L.Convolution2D(128, 128, 3, stride=1, pad=1)
self.conv3_1=L.Convolution2D(128, 256, 3, stride=1, pad=1)
self.conv3_2=L.Convolution2D(256, 256, 3, stride=1, pad=1)
self.conv3_3=L.Convolution2D(256, 256, 3, stride=1, pad=1)
self.conv4_1=L.Convolution2D(256, 512, 3, stride=1, pad=1)
self.conv4_2=L.Convolution2D(512, 512, 3, stride=1, pad=1)
self.conv4_3=L.Convolution2D(512, 512, 3, stride=1, pad=1)
self.conv5_1=L.Convolution2D(512, 512, 3, stride=1, pad=1)
self.conv5_2=L.Convolution2D(512, 512, 3, stride=1, pad=1)
self.conv5_3=L.Convolution2D(512, 512, 3, stride=1, pad=1)
self.fc6=L.Linear(25088, 4096)
self.fc7=L.Linear(4096, 4096)
self.fc8=L.Linear(4096, 1000)
def __call__(self, x):
h = F.relu(self.conv1_1(x))
h = F.relu(self.conv1_2(h))
h = F.max_pooling_2d(h, 2, stride=2)
h = F.relu(self.conv2_1(h))
h = F.relu(self.conv2_2(h))
h = F.max_pooling_2d(h, 2, stride=2)
h = F.relu(self.conv3_1(h))
h = F.relu(self.conv3_2(h))
h = F.relu(self.conv3_3(h))
h = F.max_pooling_2d(h, 2, stride=2)
h = F.relu(self.conv4_1(h))
h = F.relu(self.conv4_2(h))
h = F.relu(self.conv4_3(h))
h = F.max_pooling_2d(h, 2, stride=2)
h = F.relu(self.conv5_1(h))
h = F.relu(self.conv5_2(h))
h = F.relu(self.conv5_3(h))
h = F.max_pooling_2d(h, 2, stride=2)
h = F.relu(self.fc6(h))
h = F.relu(self.fc7(h))
h = self.fc8(h)
return h
def main():
model = VGG()
model.to_intel64()
nb_itr = 20
timings = []
for i in tqdm.tqdm(range(nb_itr)):
data = np.random.randn(1, 3, 224, 224).astype(np.float32)
start_time = time.time()
with chainer.using_config('train', False), chainer.using_config('enable_backprop', False), chainer.using_config('use_ideep', 'always'):
ret = F.softmax(model(chainer.Variable(data)))
timings.append(time.time() - start_time)
print('%10s : %f (sd %f)'% ('mxnet-vgg-16', np.array(timings).mean(), np.array(timings).std()))
if __name__ == '__main__':
main()
@ikeyasu
Copy link
Author

ikeyasu commented Jun 23, 2018

Forked from
https://github.com/peisuke/DeepLearningSpeedComparison

Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment