chainerx.force_backprop_mode

chainerx.force_backprop_mode()

Creates a context manager which temporarily enables backpropagation.

This context re-enables backpropagation that is disabled by any surrounding no_backprop_mode() context.

x = chainerx.array([4, 3], numpy.float32)
x.require_grad()

with chainerx.no_backprop_mode():
    with chainerx.force_backprop_mode():
        y = 2 * x + 1

y.backward()
x.grad
# array([2., 2.], shape=(2,), dtype=float32, device='native:0')

See also

• chainerx.no_backprop_mode()

• chainerx.is_backprop_required()

• chainer.force_backprop_mode()