def test_grid_sample_gradient_1d(self):
for backend in BACKENDS:
if backend.supports(Backend.gradients):
with backend:
grid = math.tensor([0., 1, 2, 3], spatial('x'))
coords = math.tensor([0.5, 1.5], instance('points'))
with math.record_gradients(grid, coords):
sampled = math.grid_sample(grid, coords, extrapolation.ZERO)
loss = math.mean(math.l2_loss(sampled)) / 2
grad_grid, grad_coords = math.gradients(loss, grid, coords)
math.assert_close(grad_grid, math.tensor([0.125, 0.5, 0.375, 0], spatial('x')), msg=backend)
math.assert_close(grad_coords, math.tensor([0.25, 0.75], instance('points')), msg=backend)
def test_grid_sample_gradient_1d(self):
grads_grid = []
grads_coords = []
for backend in BACKENDS:
if backend.supports(Backend.gradients):
print(backend)
with backend:
grid = math.tensor([0., 1, 2, 3], 'x')
coords = math.tensor([0.5, 1.5], 'points')
with math.record_gradients(grid, coords):
sampled = math.grid_sample(grid, coords,
extrapolation.ZERO)
loss = math.l2_loss(sampled)
grad_grid, grad_coords = math.gradients(
loss, grid, coords)
grads_grid.append(grad_grid)
grads_coords.append(grad_coords)
math.assert_close(*grads_grid, math.tensor([0.125, 0.5, 0.375, 0],
'x'))
math.assert_close(*grads_coords, math.tensor([0.25, 0.75], 'points'))
def f(x: math.Tensor, y: math.Tensor):
pred = x
loss = math.l2_loss(pred - y)
return loss, pred
def l2_loss(field: SampledField, batch_norm=True):
""" L2 loss for the unweighted values of the field. See `phi.math.l2_loss()`. """
return math.l2_loss(field.values, batch_norm=batch_norm)
def loss_function(x):
print("Running loss_function")
assert isinstance(x, math.Tensor)
pred = math.native_call(net, x)
return math.l2_loss(pred)
def f(x: math.Tensor, y: math.Tensor):
assert isinstance(x, math.Tensor)
assert isinstance(y, math.Tensor)
pred = x
loss = math.l2_loss(pred - y)
return loss, pred
def loss_function(x):
return math.l2_loss(x)
def loss(x, y):
return math.l2_loss(x - 1) + math.l2_loss(y + 1)