-
Notifications
You must be signed in to change notification settings - Fork 0
/
cnn.py
344 lines (283 loc) · 14.1 KB
/
cnn.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
102
103
104
105
106
107
108
109
110
111
112
113
114
115
116
117
118
119
120
121
122
123
124
125
126
127
128
129
130
131
132
133
134
135
136
137
138
139
140
141
142
143
144
145
146
147
148
149
150
151
152
153
154
155
156
157
158
159
160
161
162
163
164
165
166
167
168
169
170
171
172
173
174
175
176
177
178
179
180
181
182
183
184
185
186
187
188
189
190
191
192
193
194
195
196
197
198
199
200
201
202
203
204
205
206
207
208
209
210
211
212
213
214
215
216
217
218
219
220
221
222
223
224
225
226
227
228
229
230
231
232
233
234
235
236
237
238
239
240
241
242
243
244
245
246
247
248
249
250
251
252
253
254
255
256
257
258
259
260
261
262
263
264
265
266
267
268
269
270
271
272
273
274
275
276
277
278
279
280
281
282
283
284
285
286
287
288
289
290
291
292
293
294
295
296
297
298
299
300
301
302
303
304
305
306
307
308
309
310
311
312
313
314
315
316
317
318
319
320
321
322
323
324
325
326
327
328
329
330
331
332
333
334
335
336
337
338
339
340
341
342
343
344
import numpy as np
from nndl.layers import *
from nndl.conv_layers import *
from cs231n.fast_layers import *
from nndl.layer_utils import *
from nndl.conv_layer_utils import *
import pdb
"""
This code was originally written for CS 231n at Stanford University
(cs231n.stanford.edu). It has been modified in various areas for use in the
ECE 239AS class at UCLA. This includes the descriptions of what code to
implement as well as some slight potential changes in variable names to be
consistent with class nomenclature. We thank Justin Johnson & Serena Yeung for
permission to use this code. To see the original version, please visit
cs231n.stanford.edu.
"""
class ThreeLayerConvNet(object):
"""
A three-layer convolutional network with the following architecture:
conv - relu - 2x2 max pool - affine - relu - affine - softmax
The network operates on minibatches of data that have shape (N, C, H, W)
consisting of N images, each with height H and width W and with C input
channels.
"""
def __init__(self, input_dim=(3, 32, 32), num_filters=32, filter_size=7,
hidden_dim=100, num_classes=10, weight_scale=1e-3, reg=0.0,
dtype=np.float32, use_batchnorm=False):
"""
Initialize a new network.
Inputs:
- input_dim: Tuple (C, H, W) giving size of input data
- num_filters: Number of filters to use in the convolutional layer
- filter_size: Size of filters to use in the convolutional layer
- hidden_dim: Number of units to use in the fully-connected hidden layer
- num_classes: Number of scores to produce from the final affine layer.
- weight_scale: Scalar giving standard deviation for random initialization
of weights.
- reg: Scalar giving L2 regularization strength
- dtype: numpy datatype to use for computation.
"""
self.use_batchnorm = use_batchnorm
self.params = {}
self.reg = reg
self.dtype = dtype
# ================================================================ #
# YOUR CODE HERE:
# Initialize the weights and biases of a three layer CNN. To initialize:
# - the biases should be initialized to zeros.
# - the weights should be initialized to a matrix with entries
# drawn from a Gaussian distribution with zero mean and
# standard deviation given by weight_scale.
# ================================================================ #
C, H, W = input_dim
self.params['W1'] = np.random.randn(num_filters, C, filter_size, filter_size)*weight_scale
self.params['b1'] = np.zeros(num_filters)
self.params['W2'] = np.random.randn(num_filters*(H//2)*(W//2), hidden_dim)*weight_scale
self.params['b2'] = np.zeros(hidden_dim)
self.params['W3'] = np.random.randn(hidden_dim, num_classes)*weight_scale
self.params['b3'] = np.zeros(num_classes)
# ================================================================ #
# END YOUR CODE HERE
# ================================================================ #
for k, v in self.params.items():
self.params[k] = v.astype(dtype)
def loss(self, X, y=None):
"""
Evaluate loss and gradient for the three-layer convolutional network.
Input / output: Same API as TwoLayerNet in fc_net.py.
"""
W1, b1 = self.params['W1'], self.params['b1']
W2, b2 = self.params['W2'], self.params['b2']
W3, b3 = self.params['W3'], self.params['b3']
# pass conv_param to the forward pass for the convolutional layer
filter_size = W1.shape[2]
conv_param = {'stride': 1, 'pad': (filter_size - 1) / 2}
# pass pool_param to the forward pass for the max-pooling layer
pool_param = {'pool_height': 2, 'pool_width': 2, 'stride': 2}
scores = None
# ================================================================ #
# YOUR CODE HERE:
# Implement the forward pass of the three layer CNN. Store the output
# scores as the variable "scores".
# ================================================================ #
out1, cache1 = conv_relu_pool_forward(X, W1, b1,conv_param, pool_param)
out2, cache2 = affine_relu_forward(out1, W2, b2)
out3, cache3 = affine_forward(out2, W3, b3)
scores = out3
# ================================================================ #
# END YOUR CODE HERE
# ================================================================ #
if y is None:
return scores
loss, grads = 0, {}
# ================================================================ #
# YOUR CODE HERE:
# Implement the backward pass of the three layer CNN. Store the grads
# in the grads dictionary, exactly as before (i.e., the gradient of
# self.params[k] will be grads[k]). Store the loss as "loss", and
# don't forget to add regularization on ALL weight matrices.
# ================================================================ #
loss, dscores = softmax_loss(scores, y)
loss += 0.5 * self.reg * (np.sum(W1*W1) + np.sum(W2*W2) + np.sum(W3*W3))
dX3, dW3, db3 = affine_backward(dscores, cache3)
dX2, dW2, db2 = affine_relu_backward(dX3, cache2)
dX1, dW1, db1 = conv_relu_pool_backward(dX2, cache1)
grads['W1'] = dW1 + self.reg * W1
grads['W2'] = dW2 + self.reg * W2
grads['W3'] = dW3 + self.reg * W3
grads['b1'] = db1
grads['b2'] = db2
grads['b3'] = db3
# ================================================================ #
# END YOUR CODE HERE
# ================================================================ #
return loss, grads
pass
'''
#---------------------------------------------------------------------------------#
#---------------------------------------------------------------------------------#
#---------------------------------------------------------------------------------#
#---------------------------------------------------------------------------------#
#---------------------------------------------------------------------------------#
#---------------------------------------------------------------------------------#
'''
class ConvNet(object):
"""
# Designed 6 Layer networks ConvNet
# 3x3 filter, 2x2 max pool
# Applied batchnorm after every layers
# Network archicecture = [conv-relu-norm]-[conv-relu-pool-norm]x3-affine-relu-norm-affine-softmax
# 1st conv Layer used 8 filters
# 2nd conv Layer used 16 filters
# 3rd conv Layer used 32 filters
# 4th conv Layer used 64 filters
"""
def __init__(self, input_dim=(3, 32, 32), num_filters=8, filter_size=3,
hidden_dim=100, num_classes=10, weight_scale=1e-3, reg=0.25,
dtype=np.float32, use_batchnorm=True):
"""
Initialize a new network.
Inputs:
- input_dim: Tuple (C, H, W) giving size of input data
- num_filters: Number of filters to use in the convolutional layer
- filter_size: Size of filters to use in the convolutional layer
- hidden_dim: Number of units to use in the fully-connected hidden layer
- num_classes: Number of scores to produce from the final affine layer.
- weight_scale: Scalar giving standard deviation for random initialization
of weights.
- reg: Scalar giving L2 regularization strength
- dtype: numpy datatype to use for computation.
"""
self.use_batchnorm = use_batchnorm
self.params = {}
self.reg = reg
self.dtype = dtype
# ================================================================ #
# YOUR CODE HERE:
# Initialize the weights and biases of a three layer CNN. To initialize:
# - the biases should be initialized to zeros.
# - the weights should be initialized to a matrix with entries
# drawn from a Gaussian distribution with zero mean and
# standard deviation given by weight_scale.
# ================================================================ #
C, H, W = input_dim
self.params['W1'] = np.random.randn(num_filters, C, filter_size, filter_size)*weight_scale
self.params['b1'] = np.zeros(num_filters)
self.params['gamma1'] = np.ones(num_filters)
self.params['beta1'] = np.zeros(num_filters)
self.params['W2'] = np.random.randn(num_filters*2,num_filters,filter_size,filter_size)*weight_scale
self.params['b2'] = np.zeros(num_filters*2)
self.params['gamma2'] = np.ones(num_filters*2)
self.params['beta2'] = np.zeros(num_filters*2)
self.params['W3'] = np.random.randn(num_filters*4,num_filters*2,filter_size,filter_size)*weight_scale
self.params['b3'] = np.zeros(num_filters*4)
self.params['gamma3'] = np.ones(num_filters*4)
self.params['beta3'] = np.zeros(num_filters*4)
self.params['W4'] = np.random.randn(num_filters*8,num_filters*4,filter_size,filter_size)*weight_scale
self.params['b4'] = np.zeros(num_filters*8)
self.params['gamma4'] = np.ones(num_filters*8)
self.params['beta4'] = np.zeros(num_filters*8)
self.params['W5'] = np.random.randn(8*num_filters*(H//8)*(W//8), hidden_dim)*weight_scale
self.params['b5'] = np.zeros(hidden_dim)
self.params['gamma5'] = np.ones(hidden_dim)
self.params['beta5'] = np.zeros(hidden_dim)
self.params['W6'] = np.random.randn(hidden_dim, num_classes)*weight_scale
self.params['b6'] = np.zeros(num_classes)
# ================================================================ #
# END YOUR CODE HERE
# ================================================================ #
self.bn_params = []
self.bn_params = [{'mode': 'train'} for i in np.arange(6)]
for k, v in self.params.items():
self.params[k] = v.astype(dtype)
def loss(self, X, y=None):
"""
Evaluate loss and gradient for the three-layer convolutional network.
Input / output: Same API as TwoLayerNet in fc_net.py.
"""
W1, b1 = self.params['W1'], self.params['b1']
W2, b2 = self.params['W2'], self.params['b2']
W3, b3 = self.params['W3'], self.params['b3']
W4, b4 = self.params['W4'], self.params['b4']
W5, b5 = self.params['W5'], self.params['b5']
W6, b6 = self.params['W6'], self.params['b6']
gamma1, beta1 = self.params['gamma1'], self.params['beta1']
gamma2, beta2 = self.params['gamma2'], self.params['beta2']
gamma3, beta3 = self.params['gamma3'], self.params['beta3']
gamma4, beta4 = self.params['gamma4'], self.params['beta4']
gamma5, beta5 = self.params['gamma5'], self.params['beta5']
mode = 'test' if y is None else 'train'
for bn_param in self.bn_params:
bn_param[mode] = mode
# pass conv_param to the forward pass for the convolutional layer
filter_size = W1.shape[2]
conv_param = {'stride': 1, 'pad': (filter_size - 1) / 2}
# pass pool_param to the forward pass for the max-pooling layer
pool_param = {'pool_height': 2, 'pool_width': 2, 'stride': 2}
scores = None
# ================================================================ #
# YOUR CODE HERE:
# Implement the forward pass of the three layer CNN. Store the output
# scores as the variable "scores".
# ================================================================ #
out1, cache1 = conv_relu_forward(X, W1, b1, conv_param)
out1, bncache1 = spatial_batchnorm_forward(out1, gamma1, beta1, self.bn_params[1])
out2, cache2 = conv_relu_pool_forward(out1, W2, b2, conv_param, pool_param)
out2, bncache2 = spatial_batchnorm_forward(out2, gamma2, beta2, self.bn_params[2])
out3, cache3 = conv_relu_pool_forward(out2, W3, b3, conv_param, pool_param)
out3, bncache3 = spatial_batchnorm_forward(out3, gamma3, beta3, self.bn_params[3])
out4, cache4 = conv_relu_pool_forward(out3, W4, b4, conv_param, pool_param)
out4, bncache4 = spatial_batchnorm_forward(out4, gamma4, beta4, self.bn_params[4])
out5, cache5 = affine_relu_forward(out4, W5, b5)
out5, bncache5 = batchnorm_forward(out5, gamma5, beta5, self.bn_params[5])
out6, cache6 = affine_forward(out5, W6, b6)
scores = out6
# ================================================================ #
# END YOUR CODE HERE
# ================================================================ #
if y is None:
return scores
loss, grads = 0, {}
# ================================================================ #
# YOUR CODE HERE:
# Implement the backward pass of the three layer CNN. Store the grads
# in the grads dictionary, exactly as before (i.e., the gradient of
# self.params[k] will be grads[k]). Store the loss as "loss", and
# don't forget to add regularization on ALL weight matrices.
# ================================================================ #
loss, dscores = softmax_loss(scores, y)
loss += 0.5 * self.reg * (np.sum(W1*W1) + np.sum(W2*W2) + np.sum(W3*W3)+np.sum(W6*W6) + np.sum(W4*W4) + np.sum(W5*W5))
dX6, dW6, db6 = affine_backward(dscores, cache6)
dX6, dgamma5, dbeta5 = batchnorm_backward(dX6, bncache5)
dX5, dW5, db5 = affine_relu_backward(dX6, cache5)
dX5, dgamma4, dbeta4 = spatial_batchnorm_backward(dX5, bncache4)
dX4, dW4, db4 = conv_relu_pool_backward(dX5, cache4)
dX4, dgamma3, dbeta3 = spatial_batchnorm_backward(dX4, bncache3)
dX3, dW3, db3 = conv_relu_pool_backward(dX4, cache3)
dX3, dgamma2, dbeta2 = spatial_batchnorm_backward(dX3, bncache2)
dX2, dW2, db2 = conv_relu_pool_backward(dX3, cache2)
dX2, dgamma1, dbeta1 = spatial_batchnorm_backward(dX2, bncache1)
dX1, dW1, db1 = conv_relu_backward(dX2, cache1)
grads['W1'] = dW1 + self.reg * W1
grads['W2'] = dW2 + self.reg * W2
grads['W3'] = dW3 + self.reg * W3
grads['W4'] = dW4 + self.reg * W4
grads['W5'] = dW5 + self.reg * W5
grads['W6'] = dW6 + self.reg * W6
grads['b1'] = db1
grads['b2'] = db2
grads['b3'] = db3
grads['b4'] = db4
grads['b5'] = db5
grads['b6'] = db6
grads['gamma1'], grads['beta1'] = dgamma1, dbeta1
grads['gamma2'], grads['beta2'] = dgamma2, dbeta2
grads['gamma3'], grads['beta3'] = dgamma3, dbeta3
grads['gamma4'], grads['beta4'] = dgamma4, dbeta4
grads['gamma5'], grads['beta5'] = dgamma5, dbeta5
# ================================================================ #
# END YOUR CODE HERE
# ================================================================ #
return loss, grads
pass