def _build_model(sequence_length):
model = Sequential()
model.add(Embedding(20, 10, input_length=sequence_length))
model.add(Convolution1D(4, 3))
model.add(Flatten())
model.add(Dense(5, activation="softmax"))
return model
def _build_mode(self):
print("Now we build the model")
model = Sequential()
model.add(Convolution2D(32, 8, 8, subsample=(4, 4), border_mode='same',
input_shape=(IMAGE_ROWS, IMAGE_COLS, IMAGE_CHANNELS))) # 80*80*4
model.add(Activation('relu'))
model.add(Convolution2D(64, 4, 4, subsample=(2, 2), border_mode='same'))
model.add(Activation('relu'))
model.add(Convolution2D(64, 3, 3, subsample=(1, 1), border_mode='same'))
model.add(Activation('relu'))
model.add(Flatten())
model.add(Dense(512))
model.add(Activation('relu'))
model.add(Dense(2))
# get the 1 * 2 output represent each action's probability
model.add(Activation('softmax'))
return model
ImageResize(256, 256),
ImageCenterCrop(224, 224),
ImageChannelNormalize(123.0, 117.0, 104.0),
ImageMatToTensor(),
ImageFeatureToTensor()
])
full_model = Net.load_bigdl(model_path)
# create a new model by remove layers after pool5/drop_7x7_s1
model = full_model.new_graph(["pool5/drop_7x7_s1"])
# freeze layers from input to pool4/3x3_s2 inclusive
model.freeze_up_to(["pool4/3x3_s2"])
inputNode = Input(name="input", shape=(3, 224, 224))
inception = model.to_keras()(inputNode)
flatten = Flatten()(inception)
logits = Dense(2)(flatten)
lrModel = Model(inputNode, logits)
classifier = NNClassifier(lrModel, CrossEntropyCriterion(), transformer) \
.setLearningRate(0.003).setBatchSize(40).setMaxEpoch(1).setFeaturesCol("image")
pipeline = Pipeline(stages=[classifier])
catdogModel = pipeline.fit(trainingDF)
predictionDF = catdogModel.transform(validationDF).cache()
predictionDF.show()
correct = predictionDF.filter("label=prediction").count()
overall = predictionDF.count()
convolve_net.add(
Convolution2D(
nb_filter=LAYER_2_NUM_CHANNEL, # 通道: 8 -> 2.
nb_row=CONVOLVE_2_KERNEL_SIZE, # 尺寸: 12 - 5 + 1 = 8.
nb_col=CONVOLVE_2_KERNEL_SIZE,
activation="relu",
W_regularizer=L2Regularizer(args.penalty_rate)))
convolve_net.add(
AveragePooling2D(
pool_size=(
POOLING_2_WINDOW_SIZE, # 尺寸: 8 / 2 = 4.
POOLING_2_WINDOW_SIZE),
strides=(POOLING_2_STRIDE_SIZE, POOLING_2_STRIDE_SIZE),
))
convolve_net.add(BatchNormalization())
convolve_net.add(Flatten()) # 尺寸: 4 * 4 * 2 -> 32
convolve_net.add(
Dense(
output_dim=FC_LINEAR_DIMENSION, # 尺寸: 32 -> 64.
activation="sigmoid",
W_regularizer=L2Regularizer(args.penalty_rate)))
convolve_net.add(Dropout(args.dropout_rate))
# BigDL 不支持 parameter sharing, 不得已而为之.
both_feature = TimeDistributed(layer=convolve_net,
input_shape=input_shape)(both_input)
encode_left = both_feature.index_select(1, 0)
encode_right = both_feature.index_select(1, 1)
distance = autograd.abs(encode_left - encode_right)
convolve_net.add(
Convolution2D(
nb_filter=LAYER_2_NUM_CHANNEL, # 8 -> 2.
nb_row=CONVOLVE_2_KERNEL_SIZE, # Size: 12 - 5 + 1 = 8.
nb_col=CONVOLVE_2_KERNEL_SIZE,
activation="relu",
W_regularizer=L2Regularizer(args.penalty_rate)))
convolve_net.add(
AveragePooling2D(
pool_size=(
POOLING_2_WINDOW_SIZE, # Size: 8 / 2 = 4.
POOLING_2_WINDOW_SIZE),
strides=(POOLING_2_STRIDE_SIZE, POOLING_2_STRIDE_SIZE),
))
convolve_net.add(BatchNormalization())
convolve_net.add(Flatten()) # Size: 4 * 4 * 2 -> 32
convolve_net.add(
Dense(
output_dim=FC_LINEAR_DIMENSION, # Size: 32 -> 64.
activation="sigmoid",
W_regularizer=L2Regularizer(args.penalty_rate)))
convolve_net.add(Dropout(args.dropout_rate))
# BigDL Parameter Sharing and laying out the final model.
both_feature = TimeDistributed(layer=convolve_net,
input_shape=input_shape)(both_input)
encode_left = both_feature.index_select(1, 0)
encode_right = both_feature.index_select(1, 1)
distance = autograd.abs(encode_left - encode_right)