microsoft
/
MMdnn
зеркало из https://github.com/microsoft/MMdnn.git
1
0
Форкнуть 0
Код Задачи Пакеты Проекты Релизы Вики Активность

cntk parser and extractor

This commit is contained in:
Ru ZHANG 2018-02-09 15:21:50 +08:00
Родитель a0e25cc46a 236f1675cb
Коммит 0f45220261
3 изменённых файлов: 63 добавлений и 12 удалений
Показать статистику Скачать Patch файл Скачать Diff файл Показать все файлы Свернуть все файлы

7
mmdnn/conversion/examples/imagenet_test.py
Просмотреть файл

@ -64,7 +64,7 @@ class TestKit(object):
'caffe' : {
'alexnet' : lambda path : TestKit.ZeroCenter(path, 227, True),
'vgg19' : lambda path : TestKit.ZeroCenter(path, 224, True),
'inception_v1' : lambda path : TestKit.ZeroCenter(path, 224, True),
'inception_v1' : lambda path : TestKit.ZeroCenter(path, 227, True),
'resnet152' : lambda path : TestKit.ZeroCenter(path, 224, True),
'squeezenet' : lambda path : TestKit.ZeroCenter(path, 227, False)
},
@ -74,7 +74,12 @@ class TestKit(object):
'inception_v1' : lambda path : TestKit.Standard(path, 224),
'inception_v3' : lambda path : TestKit.Standard(path, 299),
'resnet' : lambda path : TestKit.Standard(path, 299),
'resnet_v1_50' : lambda path : TestKit.ZeroCenter(path, 224, False),
'resnet_v1_101' : lambda path : TestKit.ZeroCenter(path, 224, False),
'resnet_v1_152' : lambda path : TestKit.ZeroCenter(path, 224, False),
'resnet_v2_50' : lambda path : TestKit.Standard(path, 299),
'resnet_v2_152' : lambda path : TestKit.Standard(path, 299),
'resnet_v2_200' : lambda path : TestKit.Standard(path, 299),
'resnet152' : lambda path : TestKit.Standard(path, 299),
'mobilenet' : lambda path : TestKit.Standard(path, 224)
},

42
mmdnn/conversion/examples/tensorflow/extractor.py
Просмотреть файл

@ -43,7 +43,47 @@ class tensorflow_extractor(base_extractor):
'arg_scope' : inception.inception_v3_arg_scope,
'input' : lambda : tf.placeholder(name='input', dtype=tf.float32, shape=[None, 299, 299, 3]),
'num_classes' : 1001,
}
},
'resnet_v1_50' : {
'url' : 'http://download.tensorflow.org/models/resnet_v1_50_2016_08_28.tar.gz',
'filename' : 'resnet_v1_50.ckpt',
'builder' : lambda : resnet_v1.resnet_v1_50,
'arg_scope' : resnet_v2.resnet_arg_scope,
'input' : lambda : tf.placeholder(name='input', dtype=tf.float32, shape=[None, 224, 224, 3]),
'num_classes' : 1000,
},
'resnet_v1_152' : {
'url' : 'http://download.tensorflow.org/models/resnet_v1_152_2016_08_28.tar.gz',
'filename' : 'resnet_v1_152.ckpt',
'builder' : lambda : resnet_v1.resnet_v1_152,
'arg_scope' : resnet_v2.resnet_arg_scope,
'input' : lambda : tf.placeholder(name='input', dtype=tf.float32, shape=[None, 224, 224, 3]),
'num_classes' : 1000,
},
'resnet_v2_50' : {
'url' : 'http://download.tensorflow.org/models/resnet_v2_50_2017_04_14.tar.gz',
'filename' : 'resnet_v2_50.ckpt',
'builder' : lambda : resnet_v2.resnet_v2_50,
'arg_scope' : resnet_v2.resnet_arg_scope,
'input' : lambda : tf.placeholder(name='input', dtype=tf.float32, shape=[None, 299, 299, 3]),
'num_classes' : 1001,
},
'resnet_v2_152' : {
'url' : 'http://download.tensorflow.org/models/resnet_v2_152_2017_04_14.tar.gz',
'filename' : 'resnet_v2_152.ckpt',
'builder' : lambda : resnet_v2.resnet_v2_152,
'arg_scope' : resnet_v2.resnet_arg_scope,
'input' : lambda : tf.placeholder(name='input', dtype=tf.float32, shape=[None, 299, 299, 3]),
'num_classes' : 1001,
},
'resnet_v2_200' : {
'url' : 'http://download.tensorflow.org/models/resnet_v2_200_2017_04_14.tar.gz',
'filename' : 'resnet_v2_200.ckpt',
'builder' : lambda : resnet_v2.resnet_v2_200,
'arg_scope' : resnet_v2.resnet_arg_scope,
'input' : lambda : tf.placeholder(name='input', dtype=tf.float32, shape=[None, 299, 299, 3]),
'num_classes' : 1001,
},
}

26
tests/test_conversion_imagenet.py
Просмотреть файл

@ -178,8 +178,6 @@ class TestModels(CorrectnessTest):
@staticmethod
def CntkEmit(original_framework, architecture_name, architecture_path, weight_path, image_path):
print("Testing {} from {} to CNTK.".format(architecture_name, original_framework))
# IR to code
emitter = CntkEmitter((architecture_path, weight_path))
emitter.run("converted_model.py", None, 'test')
@ -202,8 +200,6 @@ class TestModels(CorrectnessTest):
@staticmethod
def TensorflowEmit(original_framework, architecture_name, architecture_path, weight_path, image_path):
print("Testing {} from {} to TensorFlow.".format(architecture_name, original_framework))
# IR to code
emitter = TensorflowEmitter((architecture_path, weight_path))
emitter.run("converted_model.py", None, 'test')
@ -232,9 +228,10 @@ class TestModels(CorrectnessTest):
@staticmethod
def PytorchEmit(original_framework, architecture_name, architecture_path, weight_path, image_path):
import torch
print("Testing {} from {} to PyTorch.".format(architecture_name, original_framework))
# IR to code
converted_file = original_framework + '_pytorch_' + architecture_name + "_converted"
converted_file = converted_file.replace('.', '_')
emitter = PytorchEmitter((architecture_path, weight_path))
emitter.run("converted_model.py", "pytorch_weight.npy", 'test')
del emitter
@ -266,8 +263,6 @@ class TestModels(CorrectnessTest):
@staticmethod
def KerasEmit(original_framework, architecture_name, architecture_path, weight_path, image_path):
print("Testing {} from {} to Keras.".format(architecture_name, original_framework))
# IR to code
emitter = Keras2Emitter((architecture_path, weight_path))
emitter.run("converted_model.py", None, 'test')
@ -340,6 +335,7 @@ class TestModels(CorrectnessTest):
'mobilenet' : [TensorflowEmit, KerasEmit],
'nasnet' : [TensorflowEmit, KerasEmit],
},
'mxnet' : {
'vgg19' : [CntkEmit, TensorflowEmit, KerasEmit, PytorchEmit],
'imagenet1k-inception-bn' : [CntkEmit, TensorflowEmit, KerasEmit, PytorchEmit],
@ -348,17 +344,24 @@ class TestModels(CorrectnessTest):
'imagenet1k-resnext-101-64x4d' : [CntkEmit, TensorflowEmit, PytorchEmit], # Keras is too slow
'imagenet1k-resnext-50' : [CntkEmit, TensorflowEmit, KerasEmit, PytorchEmit],
},
'caffe' : {
'vgg19' : [CntkEmit, TensorflowEmit, KerasEmit, PytorchEmit],
'alexnet' : [CntkEmit],
'inception_v1' : [CntkEmit, TensorflowEmit, KerasEmit, PytorchEmit],
'resnet152' : [CntkEmit, TensorflowEmit, KerasEmit, PytorchEmit],
'squeezenet' : [CntkEmit, PytorchEmit],
'squeezenet' : [CntkEmit, PytorchEmit]
},
'tensorflow' : {
'vgg19' : [CntkEmit, TensorflowEmit, KerasEmit, PytorchEmit],
'inception_v1' : [TensorflowEmit, KerasEmit, PytorchEmit], # TODO: CntkEmit
}
'inception_v3' : [CntkEmit, TensorflowEmit, KerasEmit], # TODO: PytorchEmit
'resnet_v1_50' : [TensorflowEmit, KerasEmit, PytorchEmit], # TODO: CntkEmit
'resnet_v1_152' : [TensorflowEmit, KerasEmit, PytorchEmit], # TODO: CntkEmit
'resnet_v2_50' : [TensorflowEmit, KerasEmit, PytorchEmit], # TODO: CntkEmit
'resnet_v2_152' : [TensorflowEmit, KerasEmit, PytorchEmit], # TODO: CntkEmit
},
}
@ -367,12 +370,13 @@ class TestModels(CorrectnessTest):
ensure_dir(self.tmpdir)
for network_name in self.test_table[original_framework].keys():
print("Testing {} model {} start.".format(original_framework, network_name))
print("Test {} from {} start.".format(network_name, original_framework), file=sys.stderr, flush=True)
# get original model prediction result
original_predict = parser(network_name, self.image_path)
for emit in self.test_table[original_framework][network_name]:
print('Testing conversion {} from {} to {}.'.format(network_name, original_framework, emit.__func__.__name__[:-4]), file=sys.stderr, flush=True)
converted_predict = emit.__func__(
original_framework,
network_name,
@ -382,6 +386,8 @@ class TestModels(CorrectnessTest):
self._compare_outputs(original_predict, converted_predict)
print('Conversion {} from {} to {} passed.'.format(network_name, original_framework, emit.__func__.__name__[:-4]), file=sys.stderr, flush=True)
try:
os.remove(self.tmpdir + network_name + "_converted.json")
except OSError: