def test_model_broadcast(self):
init_executor_gateway(self.sc)
model = Linear(3, 2)
broadcasted = broadcast_model(self.sc, model)
input_data = np.random.rand(3)
output = self.sc.parallelize([input_data], 1)\
.map(lambda x: broadcasted.value.forward(x)).first()
expected = model.forward(input_data)
assert_allclose(output, expected)
def test_model_broadcast(self):
init_executor_gateway(self.sc)
model = Linear(3, 2)
broadcasted = broadcast_model(self.sc, model)
input_data = np.random.rand(3)
output = self.sc.parallelize([input_data], 1)\
.map(lambda x: broadcasted.value.forward(x)).first()
expected = model.forward(input_data)
assert_allclose(output, expected)
def test_forward_backward(self):
from bigdl.nn.layer import Linear
rng = RNG()
rng.set_seed(100)
linear = Linear(4, 5)
input = rng.uniform(0.0, 1.0, [4])
output = linear.forward(input)
assert_allclose(output,
np.array([0.41366524,
0.009532653,
-0.677581,
0.07945433,
-0.5742568]),
atol=1e-6, rtol=0)
mse = MSECriterion()
target = rng.uniform(0.0, 1.0, [5])
loss = mse.forward(output, target)
print("loss: " + str(loss))
grad_output = mse.backward(output, rng.uniform(0.0, 1.0, [5]))
l_grad_output = linear.backward(input, grad_output)
def test_forward_backward(self):
from bigdl.nn.layer import Linear
rng = RNG()
rng.set_seed(100)
linear = Linear(4, 5)
input = rng.uniform(0.0, 1.0, [4])
output = linear.forward(input)
assert_allclose(output,
np.array([0.41366524,
0.009532653,
-0.677581,
0.07945433,
-0.5742568]),
atol=1e-6, rtol=0)
mse = MSECriterion()
target = rng.uniform(0.0, 1.0, [5])
loss = mse.forward(output, target)
print("loss: " + str(loss))
grad_output = mse.backward(output, rng.uniform(0.0, 1.0, [5]))
l_grad_output = linear.backward(input, grad_output)
