def test_add_layer(self): config = { 'input_shape': [3, 5, 5], 'step_size': 1, 'mu': 0.9, 'step_decay': 0.9 } n = Net(config) l = FC(10) n.add(l)
def test_fit(self): config = { 'input_shape': [3, 5, 5], 'step_size': 1, 'mu': 0.9, 'step_decay': 0.9 } n = Net(config) l = FC(10) n.add(l) x = np.random.random([1, 3, 5, 5]).reshape(1, -1) y = np.array([0]) n.fit(x, y, 10)
def test_train_iteration(self): config = { 'input_shape': [3, 5, 5], 'step_size': 1, 'mu': 0.9, 'step_decay': 0.9 } n = Net(config) l = FC(10) n.add(l) x = np.random.random([1, 3, 5, 5]).reshape(1, -1) y = np.array([0]) loss = n.train_one_iteration(x, y)
from net import Net net = Net() x = net.variable(1) y = net.variable(2) z = net.variable(3) x2 = net.mul(x, x) y2 = net.mul(y, y) z2 = net.mul(z, z) o1 = net.add(x2, y2) o = net.add(o1, z2) net.gradient_descendent() print('x=', x.v, 'y=', y.v, 'z=', z.v)
from net import Net net = Net() x = net.variable(1) y = net.variable(2) z = net.variable(3) x2 = net.mul(x, x) # x*x=x2 y2 = net.mul(y, y) # y*y=y2 z2 = net.mul(z, z) # z*z=z2 o1 = net.add(x2, y2) # x2+y2=o1 o = net.add(o1, z2) # o1+z2=o net.gradient_descendent() print('x=', x.v, 'y=', y.v, 'z=', z.v)
from net import Net net = Net() x = net.variable(1) y = net.variable(3) x2 = net.mul(x, x) y2 = net.mul(y, y) o = net.add(x2, y2) ''' print('net.forward()=', net.forward()) print('net.backward()') net.backward() print('x=', x, 'y=', y, 'o=', o) print('gfx = x.g/o.g = ', x.g/o.g, 'gfy = y.g/o.g=', y.g/o.g) print('x2=', x2, 'y2=', y2) ''' net.gradient_descendent() print('x=', x.v, 'y=', y.v)
from net import Net net = Net() x = net.variable(1) y = net.variable(2) z = net.variable(3) x2 = net.mul(x, x) y2 = net.mul(y, y) z2 = net.mul(z, z) o = net.add(x2, y2) o = net.add(o, z2) net.gradient_descendent() print('x=', x.v, 'y=', y.v, 'z=', z.v)
from net import Net net = Net() x = net.variable(1) y = net.variable(2) z = net.variable(3) x2 = net.mul(x, x) y2 = net.mul(y, y) z2 = net.mul(z, z) A = net.add(x2, y2) B = net.add(A, z2) net.gradient_descendent() print('x=', x.v, 'y=', y.v, 'z=', z.v)
from net import Net net = Net() x = net.variable(1) y = net.variable(2) z = net.variable(3) x2 = net.mul(x, x) y2 = net.mul(y, y) z2 = net.mul(z, z) t = net.add(x2, y2) o = net.add(t, z2) net.gradient_descendent() print('x = ', x.v, 'y = ', y.v, 'z = ', z.v)
from net import Net n = Net() x = n.variable(1) y = n.variable(2) z = n.variable(3) x2 = n.mul(x, x) y2 = n.mul(y, y) z2 = n.mul(z, z) o = n.add(x2, y2) out = n.add(o, z2) n.gradient_descendent() print('x={0:.5f} y={1:.5f} z={2:.5f}'.format(x.v, y.v, z.v))
from net import Net net = Net() x = net.variable(1) y = net.variable(2) z = net.variable(3) x2 = net.mul(x, x) y2 = net.mul(y, y) z2 = net.mul(z, z) q = net.add(x2, y2) o = net.add(q, z2) ''' print('net.forward()=', net.forward()) print('net.backward()') net.backward() print('x=', x, 'y=', y, 'o=', o) print('gfx = x.g/o.g = ', x.g/o.g, 'gfy = y.g/o.g=', y.g/o.g) print('x2=', x2, 'y2=', y2) ''' net.gradient_descendent() print('x=', x.v, 'y=', y.v, 'z=', z.v)