JBed

# http://jmlr.org/papers/volume11/erhan10a/erhan10a.pdf
import cPickle as pickle
import numpy as np
 
from matplotlib import pyplot as plt
from os.path import join
from sklearn.cluster import KMeans
 
# download data here: http://www.cs.toronto.edu/~kriz/cifar.html
with open(join('data','cifar-10-batches-py','data_batch_1'),'rb') as f:

JBed

import re, math
from collections import Counter


def sentence_to_vector(text):
     words = WORD.findall(text)
     return Counter(words)

JBed

from sklearn.cluster.k_means_ import _k_init 
from operator import itemgetter
import numpy as np

def quasi_random_sampling(X, n_samples=25):
    def row_norms(X, squared=True):
        squared_norms = (X**2).sum(axis=1)
        if squared: return squared_norms 
        else: return np.sqrt(squared_norms)

JBed

# Inception-v3: http://arxiv.org/abs/1512.00567

from lasagne.layers import InputLayer
from lasagne.layers import Conv2DLayer
from lasagne.layers import Pool2DLayer
from lasagne.layers import DenseLayer
from lasagne.layers import GlobalPoolLayer
from lasagne.layers import ConcatLayer
from lasagne.layers.normalization import batch_norm
from lasagne.nonlinearities import softmax

JBed

JBed

Keybase proof

I hereby claim:

• I am jbed on github.
• I am jason128 (https://keybase.io/jason128) on keybase.
• I have a public key ASAKteCJXb25zZO9-tv9oZpD0S74odBhXcqpZfU2XEnx4Ao

To claim this, I am signing this object:

JBed

#PCA whitening involves finding the inverse square root of the covariance matrix 
#of a set of observations, which is prohibitively expensive when dealing 
#with natural images 

#starting with a path to a single image (img_path)
import numpy as np
from PIL import Image
from sklearn import preprocessing

JBed

from keras.models import Sequential
from keras.layers.core import Dense, Dropout, Activation, Flatten
from keras.layers.convolutional import Convolution2D, MaxPooling2D
from keras.layers.normalization import BatchNormalization

#AlexNet with batch normalization in Keras 
#input image is 224x224

model = Sequential()
model.add(Convolution2D(64, 3, 11, 11, border_mode='full'))