In [1]:

    

import bhtsne
import h5py

import numpy as np
import pickle
from time import time
import matplotlib.pyplot as plt
%matplotlib inline


    

In [3]:

    

train_lbls = np.load('/media/raid_arr/data/ndsb/train_lbls.npy')
test_lbls = np.load('/media/raid_arr/data/ndsb/test_lbls.npy')

x2_concat = np.load('/media/raid_arr/data/ndsb/tsne2.npy')
x2_train = x2_concat[:len(train_lbls), :]
x2_test = x2_concat[-len(test_lbls):, :]

concat_lbls = np.r_[train_lbls, test_lbls]


    

In [4]:

    

from sklearn.semi_supervised import LabelPropagation
from sklearn.semi_supervised import LabelSpreading

clf = LabelPropagation()


    

In [14]:

    

# Fitting
samp = np.arange(0, len(x2_concat)-1,10)100

x2_sub = x2_concat[samp, :]
lbls_sub = concat_lbls[samp]

tic = time()
clf.fit(x2_sub, lbls_sub)
toc = time() - tic
print toc


    

    




5952.35704398

In [13]:

    

np.arange(0, len(x2_concat)-1,10).shape


    

    
Out[13]:





(16074,)