def get_out(out_state):
arr = []
for i in range(NUM_NEURONS):
arr.append(int(round(rkf.func_to_calc_sigmoid(out_state[i][2]))))
return arr
def get_out(out_state):
arr = []
for i in range(NUM_NEURONS):
arr.append(int(round(rkf.func_to_calc_sigmoid(out_state[i][2]))))
return arr
def calc_S(step, neuron):
sum_S = 0
# print("exp func : " + str(intern_state[step][j][2]))
for j in range(NUM_NEURONS):
sum_S += W[neuron][j] * rkf.func_to_calc_sigmoid(intern_state[step][j][2])
# print(sum_S)
return sum_S
def calc_S(step, neuron):
sum_S = 0
# print("exp func : " + str(intern_state[step][j][2]))
for j in range(NUM_NEURONS):
sum_S += W[neuron][j] * rkf.func_to_calc_sigmoid(
intern_state[step][j][2])
# print(sum_S)
return sum_S
def print_out(out_state):
st_outr = ''
idx = 0
for i in range(NUM_NEURONS):
st_outr += str(int(round(rkf.func_to_calc_sigmoid(out_state[i][2]))))
idx += 1
if idx == LINE_AMOUNT:
print(st_outr)
st_outr = ''
idx = 0
print(st_outr)
def print_out(out_state):
st_outr = ''
idx = 0
for i in range(NUM_NEURONS):
st_outr += str(int(round(rkf.func_to_calc_sigmoid(out_state[i][2]))))
idx += 1
if idx == LINE_AMOUNT:
print(st_outr)
st_outr=''
idx = 0
print(st_outr)
def run():
global intern_state
read_patterns()
print(W)
learning()
print(W)
read_in_pattern()
# show_picture(I)
fig1 = plt.figure()
idx = 1
show_picture(I, fig1, 5, 10, idx)
idx += 1
print(W)
for i in range(len(W[0])):
print(W[i])
# show_picture(patterns[0], fig1, 1, 4, 1)
# show_picture(patterns[1], fig1, 1, 4, 2)
# show_picture(patterns[2], fig1, 1, 4, 3)
# show_picture(patterns[3], fig1, 1, 4, 4)
# plt.show()
for i in range(NUM_NEURONS):
intern_state[0][i][2] = I[i]
# print(W)
for sstep in range(NUM_STEPS - 1):
for i in range(NUM_NEURONS):
#print(intern_state[sstep])
intern_state[sstep + 1][i] = rkf.solve_one_step(
intern_state[sstep][i], calc_S(sstep, i), I[i],
calc_Dxy(sstep), step_length)
if sstep % 100 == 0:
print("State on step=" + str(sstep))
print_out(intern_state[sstep + 1])
if sstep > (NUM_STEPS - 4900):
show_picture(get_out(intern_state[sstep + 1]), fig1, 5, 10,
idx)
idx += 1
print("Done")
show_picture(get_out(intern_state[-2]), fig1, 5, 10, idx)
# plt.show()
arr_val = []
arr_time = []
for i in range(NUM_STEPS):
arr_time.append(intern_state[i][0][0])
arr_val.append(intern_state[i][0][3])
print(arr_time)
print(arr_val)
plt.figure(2)
plt.plot(arr_time, arr_val)
plt.show()
print_out(intern_state[-2])
def run():
global intern_state
read_patterns()
print(W)
learning()
print(W)
read_in_pattern()
# show_picture(I)
fig1 = plt.figure()
idx = 1
show_picture(I, fig1, 5, 10, idx)
idx += 1
print(W)
for i in range(len(W[0])):
print(W[i])
# show_picture(patterns[0], fig1, 1, 4, 1)
# show_picture(patterns[1], fig1, 1, 4, 2)
# show_picture(patterns[2], fig1, 1, 4, 3)
# show_picture(patterns[3], fig1, 1, 4, 4)
# plt.show()
for i in range(NUM_NEURONS):
intern_state[0][i][2] = I[i]
# print(W)
for sstep in range(NUM_STEPS-1):
for i in range(NUM_NEURONS):
#print(intern_state[sstep])
intern_state[sstep+1][i] = rkf.solve_one_step(intern_state[sstep][i], calc_S(sstep, i), I[i], calc_Dxy(sstep), step_length)
if sstep % 100 == 0:
print("State on step=" + str(sstep))
print_out(intern_state[sstep+1])
if sstep > (NUM_STEPS - 4900):
show_picture(get_out(intern_state[sstep+1]), fig1, 5, 10, idx)
idx += 1
print("Done")
show_picture(get_out(intern_state[-2]), fig1, 5, 10, idx)
# plt.show()
arr_val = []
arr_time = []
for i in range(NUM_STEPS):
arr_time.append(intern_state[i][0][0])
arr_val.append(intern_state[i][0][3])
print(arr_time)
print(arr_val)
plt.figure(2)
plt.plot(arr_time, arr_val)
plt.show()
print_out(intern_state[-2])