def run(seed=1, ticks=100):
print('#', seed)
rseed(seed)
w = o(now=0)
while w.now < ticks:
alive = False
for machine in shuffle(Machine.Factory):
if not machine.here.stop():
alive = True
w.now += 1
machine.step(w)
Machine.report(machine.name)
break
if not alive: break
return w
train_y_valid = data["y_valid"]
train_y_not_valid = data["y_not_valid"]
max_length = data["max_length"][0]
#split data, make sure test data contains all types of samples (valid and not valid ones)
train_x_valid, test_x_valid, train_y_valid, test_y_valid = split_dataset(
train_x_valid, train_y_valid, 5)
train_x_not_valid, test_x_not_valid, train_y_not_valid, test_y_not_valid = split_dataset(
train_x_not_valid, train_y_not_valid, 5)
train_x = numpy.concatenate((train_x_valid, train_x_not_valid))
train_y = numpy.concatenate((train_y_valid, train_y_not_valid))
test_x = numpy.concatenate((test_x_valid, test_x_not_valid))
test_y = numpy.concatenate((test_y_valid, test_y_not_valid))
train_x, train_y = shuffle(train_x, train_y)
test_x, test_y = shuffle(test_x, test_y)
#convert y to one-hot encoding
#train_y = to_categorical(train_y, 2)
#test_y = to_categorical(test_y, 2)
train_x = train_x.astype("float32")
test_x = test_x.astype("float32")
#train_x /= 255
#test_x /= 255
train_x = numpy.expand_dims(train_x, axis=2)
test_x = numpy.expand_dims(test_x, axis=2)
def next(i, w):
for tran in shuffle(i.out):
if tran.gaurd(w, tran):
return tran.there
return i