kgrm

import numpy as np
from keras.models import Sequential
from keras.layers import Deconv2D, Lambda

def cropped_shape(inshape):
    samples, ch, w, h = inshape
    return (samples, ch, w - 2, h - 2)

mdl = Sequential()
mdl.add(Deconv2D(256, input_shape=(1, 7, 7), padding='valid', kernel_size=4, strides=2, data_format="channels_first"))

kgrm

from keras.layers import Input, Dense, Lambda
from keras.models import Model

def eucl_dist(inputs):
    x, y = inputs
    return ((x - y)**2).sum(axis=-1)

x = Input((32,))
y1 = Dense(8)(x)
y2 = Dense(8)(x)

kgrm

import numpy as np
import sys, os
if len(sys.argv) > 1 and sys.argv[1] == "gpu":
    os.environ["THEANO_FLAGS"] = "device=gpu2, "
    os.environ["THEANO_FLAGS"] += "dnn.enabled=True, "
    os.environ["THEANO_FLAGS"] += "lib.cnmem=0.95"
from keras.layers import Input, Convolution2D, MaxPooling2D, AveragePooling2D
from keras.layers import Flatten, BatchNormalization, Activation
from keras.regularizers import l2
from keras.models import Model

kgrm

import sys, os, cv2
import caffe
import numpy as np

model_file = "YOURNET.prototxt"
pretrained = "YOURNET.caffemodel"

# Load a pretrained CAFFE neural network and an image,
# use the network to extract features from the image,
# then save the feature vector to a text file.