def create_model():
model = Model(input_shape=(256, 256, 3))
#16
model.add(Conv2D(16, (3, 3), (1, 1), padding='same'))
model.add(Conv2D(16, (3, 3), (1, 1), padding='same'))
model.add(MaxPooling2D((2, 2), strides=(2, 2)))
#32
model.add(Conv2D(32, (3, 3), (1, 1), padding='same'))
model.add(Conv2D(32, (3, 3), (1, 1), padding='same'))
model.add(MaxPooling2D((2, 2), strides=(2, 2)))
#64
model.add(Conv2D(64, (3, 3), (1, 1), padding='same'))
model.add(Conv2D(64, (3, 3), (1, 1), padding='same'))
model.add(MaxPooling2D((2, 2), strides=(2, 2)))
#96
model.add(Conv2D(96, (3, 3), (1, 1), padding='same'))
model.add(Conv2D(96, (3, 3), (1, 1), padding='same'))
model.add(MaxPooling2D((2, 2), strides=(2, 2)))
#128
model.add(Conv2D(128, (3, 3), (1, 1), padding='same'))
model.add(Conv2D(128, (3, 3), (1, 1), padding='same'))
model.add(MaxPooling2D((2, 2), strides=(2, 2)))
model.add(Flatten())
model.add(Dense(96))
model.add(Dense(5))
# save as svg file
model.save_fig("example5.svg")
def main():
model = Model(input_shape=(224, 224, 3))
model.add(Conv2D(64, (3, 3), padding="same"))
model.add(Conv2D(64, (3, 3), padding="same"))
model.add(MaxPooling2D((2, 2)))
model.add(Conv2D(128, (3, 3), padding="same"))
model.add(Conv2D(128, (3, 3), padding="same"))
model.add(MaxPooling2D((2, 2)))
model.add(Conv2D(256, (3, 3), padding="same"))
model.add(Conv2D(256, (3, 3), padding="same"))
model.add(Conv2D(256, (3, 3), padding="same"))
model.add(MaxPooling2D((2, 2)))
model.add(Conv2D(512, (3, 3), padding="same"))
model.add(Conv2D(512, (3, 3), padding="same"))
model.add(Conv2D(512, (3, 3), padding="same"))
model.add(MaxPooling2D((2, 2)))
model.add(Conv2D(512, (3, 3), padding="same"))
model.add(Conv2D(512, (3, 3), padding="same"))
model.add(Conv2D(512, (3, 3), padding="same"))
model.add(MaxPooling2D((2, 2)))
model.add(Flatten())
model.add(Dense(4096))
model.add(Dense(4096))
model.add(Dense(1000))
model.save_fig(os.path.splitext(os.path.basename(__file__))[0] + ".svg")
save_model_to_pptx(
model,
os.path.splitext(os.path.basename(__file__))[0] + ".pptx")
def convert_drawer_model(model):
_input_shape = model.input_shape
# figure = Model(input_shape=_input_shape[1:])
# CNN Model
if len(_input_shape) == 4:
figure = Model(input_shape=_input_shape[1:])
# MLP Model
else:
dims = int(np.sqrt(_input_shape[1]))
_input_shape = (1,dims, dims,1)
figure = Model(input_shape=_input_shape[1:])
figure.add(Flatten())
for config in model.get_config()["layers"]:
class_name = config.get("class_name", False)
class_config = config.get("config", False)
print(class_name,class_config)
if class_name and class_config:
# class_obj = is_class_object(class_name)
print(class_name)
if class_name == "Conv2D":
class_obj = is_class_object(class_name)
conv_2d = get_conv2d_obj(class_obj, class_config)
figure.add(conv_2d)
elif class_name == "MaxPooling2D":
class_obj = is_class_object(class_name)
max_pooling_2d = get_maxpooling2d_obj(class_obj, class_config)
figure.add(max_pooling_2d)
elif class_name == "AveragePooling2D":
class_obj = is_class_object(class_name)
avg_pooling_2d = get_avgpooling2d_obj(class_obj, class_config)
figure.add(avg_pooling_2d)
elif class_name == "Dense":
class_obj = is_class_object(class_name)
dense = get_dense_obj(class_obj, class_config)
figure.add(dense)
elif class_name == "Activation":
pass
elif class_name == "Dropout":
pass
elif class_name == "BatchNormalizationV1":
pass
elif class_name == "BatchNormalization":
pass
elif class_name == "InputLayer":
pass
elif class_name == "Sequential":
pass
elif class_name == "Add":
pass
else:
class_obj = is_class_object(class_name)
figure.add(class_obj())
else:
raise ValueError
return figure
def main():
configure(inter_layer_margin=65, channel_scale=4 / 5)
model = Model(input_shape=(32, 32, 1))
model.add(Conv2D(6, (5, 5), (1, 1)))
model.add(MaxPooling2D((2, 2)))
model.add(Conv2D(16, (5, 5)))
model.add(MaxPooling2D((2, 2)))
model.add(Flatten())
model.add(Dense(120))
model.add(Dense(84))
model.add(Dense(10))
model.save_fig(os.path.splitext(os.path.basename(__file__))[0] + ".svg")
def main():
model = Model(input_shape=(227, 227, 3))
model.add(Conv2D(96, (11, 11), (4, 4)))
model.add(MaxPooling2D((3, 3), strides=(2, 2)))
model.add(Conv2D(256, (5, 5), padding="same"))
model.add(MaxPooling2D((3, 3), strides=(2, 2)))
model.add(Conv2D(384, (3, 3), padding="same"))
model.add(Conv2D(384, (3, 3), padding="same"))
model.add(Conv2D(256, (3, 3), padding="same"))
model.add(MaxPooling2D((3, 3), strides=(2, 2)))
model.add(Flatten())
model.add(Dense(4096))
model.add(Dense(4096))
model.add(Dense(1000))
model.save_fig(os.path.splitext(os.path.basename(__file__))[0] + ".svg")
# -
# ### モデルのアーキテクチャを可視化
# +
import sys
sys.path.append('../../python_lib/convnet-drawer')
# +
from keras.models import Sequential
from convnet_drawer import Model, Conv2D, MaxPooling2D, Flatten, Dense
drawer_model = Model(input_shape=(28, 28, 3))
drawer_model.add(Flatten())
drawer_model.add(Dense(1500))
drawer_model.add(Dense(1000))
drawer_model.add(Dense(500))
drawer_model.add(Dense(10))
drawer_model.add(Dense(500))
drawer_model.add(Dense(1000))
drawer_model.add(Dense(1500))
drawer_model.add(Dense(784))
drawer_model.save_fig('alexnet.svg')
# + {"colab_type": "text", "id": "mvFdHRtk2GGY", "cell_type": "markdown"}
# ### モデルの概要を表示
# + {"colab": {"base_uri": "https://localhost:8080/", "height": 442}, "colab_type": "code", "id": "cr3WlpfZBd7a", "outputId": "75a74782-6a78-4ed2-bbb1-34dc3d770b83"}
from convnet_drawer import Model, Conv2D, MaxPooling2D, Flatten, Dense
model = Model(input_shape=(50, 50, 3))
model.add(Conv2D(32, (3, 3), padding="same"))
model.add(Conv2D(32, (3, 3)))
model.add(MaxPooling2D(pool_size=(2, 2)))
model.add(Conv2D(64, (3, 3), padding='same'))
model.add(Conv2D(64, (3, 3)))
model.add(MaxPooling2D(pool_size=(2, 2)))
model.add(Flatten())
model.add(Dense(512))
model.add(Dense(4))
model.save_fig("example.svg")