-
Notifications
You must be signed in to change notification settings - Fork 0
/
test_TNN.py
252 lines (216 loc) · 7.18 KB
/
test_TNN.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
# Test for TNN column implementation
# Based off of conv_mnist example in bindnet
# Uses our temporal neuron node implementation with k-WTA.
import math
import torch
import argparse
import matplotlib.pyplot as plt
from torchvision import transforms
from time import time as t
from tqdm import tqdm
from bindsnet.datasets import MNIST
from bindsnet.encoding import RankOrderEncoder
from bindsnet.network import Network
from bindsnet.network.monitors import Monitor
from bindsnet.network.nodes import Input
from bindsnet.network.topology import Conv2dConnection, Connection
from bindsnet.utils import get_square_weights
from bindsnet.analysis.plotting import (
plot_input,
plot_spikes,
plot_voltages,
plot_weights,
plot_conv2d_weights,
plot_voltages,
)
from TNN import *
from SpykeTorch import snn
from SpykeTorch import functional as sf
from SpykeTorch import visualization as vis
from SpykeTorch import utils
# Command line parsing
parser = argparse.ArgumentParser()
parser.add_argument("--seed", type=int, default=0)
parser.add_argument("--n_epochs", type=int, default=1)
parser.add_argument("--n_test", type=int, default=10000)
parser.add_argument("--examples", type=int, default=500)
parser.add_argument("--time", type=int, default=50)
parser.add_argument("--dt", type=int, default=1.0)
parser.add_argument("--intensity", type=float, default=128.0)
parser.add_argument("--train", dest="train", action="store_true")
parser.add_argument("--test", dest="train", action="store_false")
parser.add_argument("--plot", dest="plot", action="store_true")
parser.add_argument("--gpu", dest="gpu", action="store_true")
parser.add_argument("--device_id", type=int, default=0)
parser.set_defaults(plot=False, gpu=False, train=True)
args = parser.parse_args()
seed = args.seed
n_epochs = args.n_epochs
n_test = args.n_test
examples = args.examples
time = args.time
dt = args.dt
intensity = args.intensity
train = args.train
plot = args.plot
gpu = args.gpu
device_id = 0
gpu = False
if gpu and torch.cuda.is_available():
torch.cuda.set_device(device_id)
# torch.set_default_tensor_type('torch.cuda.FloatTensor')
else:
torch.manual_seed(seed)
# Parameters for TNN
input_size = 28*28
tnn_layer_sz = 25
num_timesteps = 256
tnn_thresh = 128
max_weight = num_timesteps/2
num_winners = 1
time = num_timesteps
# SpykeTorch specific parameters for on-off encoding
rf_size = 5 # Receptive field size that will be provided as input to the column
startposition = 13 # Start position of the receptive field w.r.t. top left corner of image
# On-Off Transform class for Image
class PreProcTransform:
def __init__(self, filter, timesteps=num_timesteps):
self.to_tensor = transforms.ToTensor() # Convert to tensor
self.filter = filter # Apply OnOff filtering
self.temporal_transform = utils.Intensity2Latency(timesteps) # Convert pixel values to time
# Higher value corresponds to earlier spiketime
self.crop = utils.Crop(startposition, rf_size) # Crop the image to form the receptive field
def __call__(self, image):
image = self.to_tensor(image) * 255
image.unsqueeze_(0) # Adds a temporal dimension at the beginning
image = self.filter(image)
temporal_image = self.temporal_transform(image)
temporal_image = temporal_image.sign() # This will create spikes
return self.crop(temporal_image)
kernels = [utils.OnKernel(3), utils.OffKernel(3)]
filter = utils.Filter(kernels, padding = 2, thresholds = 50)
preproc = PreProcTransform(filter)
# TNN Network Build
network = Network(dt=1)
input_layer = Input(n=input_size)
tnn_layer_1 = TemporalNeurons( \
n=tnn_layer_sz, \
timesteps=num_timesteps, \
threshold=tnn_thresh, \
num_winners=num_winners\
)
C1 = Connection(
source=input_layer,
target=tnn_layer_1,
w = 0.5 * max_weight * torch.rand(input_layer.n, tnn_layer_1.n),
update_rule=TNN_STDP,
ucapture = 10/128,
uminus = 10/128,
usearch = 4/128,
ubackoff = 96/128,
umin = 4/128,
timesteps = num_timesteps,
maxweight = max_weight
)
network.add_layer(input_layer, name="I")
network.add_layer(tnn_layer_1, name="TNN_1")
network.add_connection(C1, source="I", target="TNN_1")
spikes = {}
for l in network.layers:
spikes[l] = Monitor(network.layers[l], ["s"], time=num_timesteps)
network.add_monitor(spikes[l], name="%s_spikes" % l)
dataset = MNIST(
RampNoLeakTNNEncoder(time=num_timesteps, dt=1),
None,
root=os.path.join("..", "..", "data", "MNIST"),
download=True,
transform=transforms.Compose(
[transforms.ToTensor(), transforms.Lambda(lambda x: x * intensity)]
),
)
# Added by Kyle for On-Off Encoding
#dataset = MNIST(
# root=os.path.join("..", "..", "data", "MNIST"),
# download=True,
# transform=preproc,
#)
# Create a dataloader to iterate and batch data
dataloader = torch.utils.data.DataLoader(
dataset, batch_size=1, shuffle=False, num_workers=0, pin_memory=gpu
)
inpt_axes = None
inpt_ims = None
spike_axes = None
spike_ims = None
weights_im = None
weights_im2 = None
voltage_ims = None
voltage_axes = None
n_iters = examples
pbar = tqdm(enumerate(dataloader))
for (i, dataPoint) in pbar:
if i > n_iters:
break
datum = dataPoint["encoded_image"].view(time, 1, 1, 28, 28)#.to(device_id)
label = dataPoint["label"]
pbar.set_description_str("Train progress: (%d / %d)" % (i, n_iters))
network.run(inputs={"I": datum}, time=time)
n_iters = 10
training_pairs = []
pbar = tqdm(enumerate(dataloader))
for (i, dataPoint) in pbar:
if i > n_iters:
break
datum = dataPoint["encoded_image"].view(time, 1, 1, 28, 28)#.to(device_id)
label = dataPoint["label"]
pbar.set_description_str("Train progress: (%d / %d)" % (i, n_iters))
network.run(inputs={"I": datum}, time=time)
training_pairs.append([spikes["TNN_1"].get("s").int().squeeze(), label])
if plot:
inpt_axes, inpt_ims = plot_input(
dataPoint["image"].view(28, 28),
datum.view(time, 784).sum(0).view(28, 28),
label=label,
axes=inpt_axes,
ims=inpt_ims,
)
spike_ims, spike_axes = plot_spikes(
{layer: spikes[layer].get("s").view(time, -1) for layer in spikes},
axes=spike_axes,
ims=spike_ims,
)
weights_im = plot_weights(
get_square_weights(C1.w, int(math.ceil(math.sqrt(tnn_layer_sz))), 28),
im=weights_im, wmin=0, wmax=max_weight
)
plt.pause(1e-8)
network.reset_state_variables()
# TEST LOOP
n_iters = examples
table = torch.zeros((tnn_layer_sz, 10))
pred = torch.zeros(tnn_layer_sz)
totals = torch.zeros(tnn_layer_sz)
count = 0
pbar = tqdm(enumerate(dataloader))
for (i, dataPoint) in pbar:
if i > n_iters:
break
datum = dataPoint["encoded_image"].view(time, 1, 1, 28, 28)#.to(device_id)
label = dataPoint["label"]
pbar.set_description_str("Test progress: (%d / %d)" % (i, n_iters))
network.run(inputs={"I": datum}, time=time)
#training_pairs.append([spikes["TNN_1"].get("s").int().squeeze(), label])
count += 1
out = torch.sum(spikes["TNN_1"].get("s").int().squeeze(), dim=0)
temp = torch.nonzero(out)
if temp.size(0) != 0:
table[temp[0][0], label] += 1
network.reset_state_variables()
print("\n\n Confusion Matrix:")
print(table)
maxval = torch.max(table, 1)[0]
totals = torch.sum(table, 1)
pred = torch.sum(maxval)
covg_cnt = torch.sum(totals)
print("Purity: ", pred/covg_cnt)
print("Coverage: ", covg_cnt/count)