-
Notifications
You must be signed in to change notification settings - Fork 0
/
KNN_score.py
101 lines (96 loc) · 3.33 KB
/
KNN_score.py
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
61
62
63
64
65
66
67
68
69
70
71
72
73
74
75
76
77
78
79
80
81
82
83
84
85
86
87
88
89
90
91
92
93
94
95
96
97
98
99
100
101
#!/usr/bin/env python2
# -*- coding: utf-8 -*-
"""
Created on Wed Nov 30 22:19:30 2016
@author: XFZ
"""
import mxnet as mx
import cPickle
import numpy as np
from random import shuffle
from sklearn.neighbors import KNeighborsClassifier
import cv2
from magnetLoss import *
def unpickle(file):
fo = open(file, 'rb')
dict = cPickle.load(fo)
fo.close()
return dict
def provideBatch(data,batchSize=128):
nRound = len(data)/batchSize
nSample = len(data)
for start in range(nRound+1):
batch = np.zeros((batchSize,3,224,224))
for i in range(batchSize) :
ii=start*batchSize+i
if ii<nSample:
img = data[ii]
img = img.reshape(32,32,3,order='F')
img = cv2.resize(img,(224,224))
img=img.swapaxes(2,0)
batch[i]=img
yield batch
def KNN_test(data,label,featureExector,splitRatio,n_neighbors,hash_len):
#extrat feature by provide featureEx
batchSize = 128
nSample = len(data)
featureSize=hash_len
remains = nSample%batchSize
features = np.zeros((nSample,featureSize))
i=0
print 'extract features'
for batch in provideBatch(data,batchSize):
f = featureExector.predict(batch)
f = np.squeeze(f)
if (i+1)*batchSize <= nSample :
features[i*batchSize:(i+1)*batchSize]=f
else:
features[i*batchSize:nSample-1]=f[0:(remains-1)]
i+=1
print i
data_label = zip(features,label)
shuffle(data_label)
splitPoint = int(len(data_label)*splitRatio)
test = data_label[0:splitPoint]
train = data_label[splitPoint:]
test_data = [x for (x,y) in test]
test_label = [y for (x, y) in test]
train_data = [x for (x, y) in train]
train_label = [y for (x, y) in train]
neigh = KNeighborsClassifier(metric='euclidean',\
n_neighbors=n_neighbors)
neigh.fit(train_data, train_label)
score = neigh.score(test_data,test_label)
return score
#def main():
# prefix = 'model/'
# num_round = 5
# hash_len = 32
# model = mx.model.FeedForward.load(prefix, num_round, ctx=mx.gpu(), numpy_batch_size=128)
# internals = model.symbol.get_internals()
# fea_symbol = internals["fch_output"]
# feature_extractor = mx.model.FeedForward(ctx=mx.gpu(), symbol=fea_symbol, numpy_batch_size=128,
# arg_params=model.arg_params, aux_params=model.aux_params,
# allow_extra_params=True)
#
# dic=unpickle('../test/data/cifa10/test_batch')
# data = dic['data']
# label = dic['labels']
# score = KNN_test(data,label,feature_extractor,0.2,1,hash_len)
# print score
#if __name__ == '__main__':
# main()
prefix = 'model/'
num_round = 10
hash_len = 128
model = mx.model.FeedForward.load(prefix, num_round, ctx=mx.gpu(), numpy_batch_size=128)
internals = model.symbol.get_internals()
fea_symbol = internals["fch_output"]
feature_extractor = mx.model.FeedForward(ctx=mx.gpu(), symbol=fea_symbol, numpy_batch_size=128,
arg_params=model.arg_params, aux_params=model.aux_params,
allow_extra_params=True)
dic=unpickle('../test/data/cifa10/test_batch')
data = dic['data']
label = dic['labels']
score = KNN_test(data,label,feature_extractor,0.2,5,hash_len)
print score