Method update

python/singa/opt.py:738–746

Performs a single optimization step. Args: param(Tensor): param values to be update grad(Tensor): param gradients

(self, param, grad)

        self.backward_and_update(loss)

    def update(self, param, grad):
        """Performs a single optimization step.

        Args:
                param(Tensor): param values to be update
                grad(Tensor): param gradients
        """
        grad /= self.world_size
        self.opt.update(param, grad)

    def all_reduce(self, tensor):
        """Performs all reduce of a tensor for distributed training.