-
Notifications
You must be signed in to change notification settings - Fork 2
/
circuit_training.py
178 lines (155 loc) · 5.88 KB
/
circuit_training.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
"""
.. module:: circuit_training
:synopsis: Module containing functions to organize the training of circuits
.. moduleauthor:: Marc Javin
"""
import sys
import numpy as np
import scipy as sp
from odynn import neuron as nr
from odynn import circuit
from odynn import datas
from odynn import utils
from odynn.neuron import PyBioNeuron
from odynn.circuit import CircuitTf
from odynn.coptim import CircuitOpt
import odynn.csimul as sim
p = PyBioNeuron.default_params
def inhibit():
inhib =circuit.SYNAPSE_inhib
connections = {(0,1) : inhib, (1,0) : inhib}
t = np.array(sp.arange(0.0, 2500.,datas.DT))
i0 = 10.*((t>300)&(t<350)) + 20.*((t>900)&(t<950))
i1 = 10.*((t>500)&(t<550)) + 20.*((t>700)&(t<750)) + 14*((t>1100)&(t<1800))+ 22*((t>1900)&(t<2000))
i_injs = np.array([i0, i1]).transpose()
sim.simul(t, i_injs, [p, p], connections, show=True, save=False)
def dual():
inhib =circuit.SYNAPSE_inhib
connections = {(1,0) : circuit.SYNAPSE2, (0,1) : inhib}
t = np.array(sp.arange(0.0, 2500.,datas.DT))
i0 = 10.*((t>300)&(t<350)) + 20.*((t>900)&(t<950))
i1 = 10.*((t>500)&(t<550)) + 20.*((t>700)&(t<750)) + 14*((t>1100)&(t<1800))+ 22*((t>1900)&(t<2000))
i_injs = np.array([i0, i1]).transpose()
sim.simul(t, i_injs, [p, p], connections, show=True, save=False)
def opt_neurons():
connections = {(0, 1):circuit.SYNAPSE_inhib,
(1, 0):circuit.SYNAPSE}
t = np.array(sp.arange(0.0, 1000.,datas.DT))
i0 = 10. * ((t > 200) & (t < 400)) + 30. * ((t > 500) & (t < 600))
i1 = 30. * ((t > 700) & (t < 800))
i_injs = np.array([i0, i1]).transpose()
f = sim.simul(t, i_injs, [p, p], connections, dump=True)
c = CircuitOpt([p, p], connections)
c.opt_neurons(f)
def test(nb_neuron, conns, conns_opt, dir, t, i_injs, n_out=[1]):
pars = [p for _ in range(nb_neuron)]
dir = utils.set_dir(dir)
print("Feed with current of shape : ", i_injs.shape)
# train = sim.simul(t, i_injs, pars, conns, n_out=n_out, show=False)
n = nr.BioNeuronTf(init_p=pars, dt=t[1]-t[0])
cr = CircuitTf(n, synapses=conns_opt)
cr.plot(save=True);exit()
c = CircuitOpt(cr)
c.optimize(dir, n_out=n_out, train=train)
def full4to1():
t,i = datas.full4(nb_neuron_zero=1)
n_neuron = 5
conns = {(0, 4):circuit.SYNAPSE,
(1, 4):circuit.SYNAPSE,
(2, 4):circuit.SYNAPSE,
(3, 4):circuit.SYNAPSE,
}
conns_opt = {(0, 4):circuit.get_syn_rand(),
(1, 4):circuit.get_syn_rand(),
(2, 4):circuit.get_syn_rand(),
(3, 4):circuit.get_syn_rand(),
}
dir = '4to1-test'
test(n_neuron, conns, conns_opt, dir, t, i, n_out=[4])
def full441():
t, i =datas.full4(nb_neuron_zero=6)
print(i.shape)
n_neuron = 10
conns = {(0, 4):circuit.SYNAPSE,
(1, 4):circuit.SYNAPSE,
(2, 4):circuit.SYNAPSE,
(3, 4):circuit.SYNAPSE,
(0, 5):circuit.SYNAPSE1,
(1, 5):circuit.SYNAPSE1,
(2, 5):circuit.SYNAPSE1,
(3, 5):circuit.SYNAPSE1,
(0, 6):circuit.SYNAPSE2,
(1, 6):circuit.SYNAPSE2,
(2, 6):circuit.SYNAPSE2,
(3, 6):circuit.SYNAPSE2,
(0, 7):circuit.SYNAPSE,
(1, 7):circuit.SYNAPSE1,
(2, 7):circuit.SYNAPSE,
(3, 7):circuit.SYNAPSE1,
(4, 8):circuit.SYNAPSE1,
(5, 8):circuit.SYNAPSE,
(6, 8):circuit.SYNAPSE,
(7, 8):circuit.SYNAPSE2,
(4, 9):circuit.SYNAPSE,
(5, 9):circuit.SYNAPSE2,
(6, 9):circuit.SYNAPSE1,
(7, 9):circuit.SYNAPSE,
}
conns_opt = dict([(k,circuit.get_syn_rand()) for k in conns.keys()])
dir = '4to4to2-test'
test(n_neuron, conns, conns_opt, dir, t, i, n_out=[8, 9])
exit(0)
def with_LSTM():
dir = utils.set_dir('withLSTM')
conns = {(0, 1): circuit.SYNAPSE}
conns_opt = {(0, 1): circuit.get_syn_rand(True)}
dt = 0.5
t, i = datas.give_train(dt=dt)
i_1 = np.zeros(i.shape)
i_injs = np.stack([i, i_1], axis=2)
train = sim.simul(t, i_injs, [p, p], conns, n_out=[0, 1], show=False)
neurons = nr.Neurons([nr.BioNeuronTf(PyBioNeuron.get_random(), fixed=[], dt=dt), nr.BioNeuronTf(p, fixed='all', dt=dt)])
c = CircuitTf(neurons=neurons, synapses=conns_opt)
co = CircuitOpt(circuit=c)
co.optimize(dir, train=train, n_out=[0, 1], l_rate=(0.01, 9, 0.95))
if __name__ == '__main__':
xp = sys.argv[1]
if(xp == '21exc'):
n_neuron = 2
conns = {(0,1) :circuit.SYNAPSE}
conns_opt = {(0,1) :circuit.get_syn_rand(True)}
dir = '2n-1exc-test'
elif(xp=='21inh'):
n_neuron = 2
conns = {(0,1):circuit.SYNAPSE_inhib}
conns_opt = {(0,1):circuit.get_syn_rand(False)}
dir = '2n-1inh-test'
elif(xp == '22exc'):
n_neuron = 2
conns = {(0,1):circuit.SYNAPSE,
(1,0):circuit.SYNAPSE}
conns_opt = [{(0,1):circuit.get_syn_rand(True),
(1,0):circuit.get_syn_rand(True)} for _ in range(100)]
dir = '2n-2exc-test'
elif(xp == '21exc1inh'):
n_neuron = 2
conns = {(0, 1):circuit.SYNAPSE,
(1, 0):circuit.SYNAPSE_inhib}
conns_opt = {(0, 1):circuit.get_syn_rand(True),
(1, 0):circuit.get_syn_rand(False)}
dir = '2n-1exc1inh-test'
elif (xp == '22inh'):
n_neuron = 2
conns = {(0, 1):circuit.SYNAPSE_inhib,
(1, 0):circuit.SYNAPSE_inhib}
conns_opt = {(0, 1):circuit.get_syn_rand(False),
(1, 0):circuit.get_syn_rand(False)}
dir = '2n-2inh-test'
elif xp == '41':
full4to1()
elif xp=='441':
full441()
t, i =datas.give_train(dt=0.5)
i_1 = np.zeros(i.shape)
i_injs = np.stack([i, i_1], axis=2)
test(n_neuron, conns, conns_opt, dir, t, i_injs)