def test_mm_001(self):
a1 = Tensor(np.array([[1.0], [2.0], [3.0]]), autograd=True)
w_2 = Tensor(np.array([[11.0, 21.0, 31.0], [12.0, 22.0, 32.0],
[13.0, 23.0, 33.0], [14.0, 24.0, 34.0]]),
autograd=True)
z2 = w_2.mm(a1)
print('z2: {0};\r\n{1}'.format(z2.data.shape, z2))
z2.backward(Tensor(np.ones_like(z2.data)))
print('a1.grad: {0};'.format(a1.grad))
print('w_2.grad: {0};'.format(w_2.grad))
def test_train_nn_002(self):
np.random.seed(0)
data = Tensor(np.array([[0, 0], [0, 1], [1, 0], [1, 1]]),
autograd=True)
target = Tensor(np.array([[0], [1], [0], [1]]), autograd=True)
w = list()
w.append(Tensor(np.random.rand(2, 3), autograd=True))
w.append(Tensor(np.random.rand(3, 1), autograd=True))
optim = OptimSgd(parameters=w, alpha=0.1)
for i in range(10):
pred = data.mm(w[0]).mm(w[1])
loss = ((pred - target) * (pred - target)).sum(0)
loss.backward(Tensor(np.ones_like(loss.data)))
optim.step()
print('epoch_{0}: loss={1};'.format(i, loss))
print(loss)
def test_train_nn_001(self):
np.random.seed(0)
data = Tensor(np.array([[0, 0], [0, 1], [1, 0], [1, 1]]),
autograd=True)
target = Tensor(np.array([[0], [1], [0], [1]]), autograd=True)
w = list()
w.append(Tensor(np.random.rand(2, 3), autograd=True))
w.append(Tensor(np.random.rand(3, 1), autograd=True))
for i in range(10):
pred = data.mm(w[0]).mm(w[1])
loss = ((pred - target) * (pred - target)).sum(0)
loss.backward(Tensor(np.ones_like(loss.data)))
for w_ in w:
w_.data -= w_.grad.data * 0.1
w_.grad.data *= 0
print('epoch_{0}: loss={1};'.format(i, loss))
print(loss)