Ejemplo n.º 1
0
class GradientMergeOptimizer(MetaOptimizerBase):
    def __init__(self, optimizer):
        super(GradientMergeOptimizer, self).__init__(optimizer)
        self.inner_opt = optimizer
        self.wrapped_opt = None
        self.meta_optimizers_white_list = [
            "AMPOptimizer",
            "LarsOptimizer",
            "LambOptimizer",
            "GraphExecutionOptimizer",
            "RecomputeOptimizer",
        ]
        self.meta_optimizers_black_list = []

    def _set_basic_info(self, loss, role_maker, user_defined_optimizer,
                        user_defined_strategy):
        super(GradientMergeOptimizer,
              self)._set_basic_info(loss, role_maker, user_defined_optimizer,
                                    user_defined_strategy)

    def _init_wrapped_opt(self):
        config = self.user_defined_strategy.gradient_merge_configs
        self.wrapped_opt = GM(self.inner_opt)
        self.wrapped_opt._set_k_steps(
            self.user_defined_strategy.gradient_merge_configs["k_steps"])
        self.wrapped_opt._set_avg(
            self.user_defined_strategy.gradient_merge_configs["avg"])

    def _can_apply(self):
        if not self.role_maker._is_collective:
            return False

        can_apply = (self.user_defined_strategy.gradient_merge == True) and \
            self.user_defined_strategy.gradient_merge_configs["k_steps"] > 1
        return can_apply

    def _disable_strategy(self, dist_strategy):
        dist_strategy.gradient_merge = False
        dist_strategy.gradient_merge_configs = {}

    def _enable_strategy(self, dist_strategy, context):
        # we currently do not support auto-enable GradientMerge
        return

    def minimize_impl(self,
                      loss,
                      startup_program=None,
                      parameter_list=None,
                      no_grad_set=None):
        self._init_wrapped_opt()
        optimize_ops, params_grads = \
            self.wrapped_opt.minimize(loss, startup_program,
                                      parameter_list, no_grad_set)
        return optimize_ops, params_grads
 def _init_wrapped_opt(self):
     config = self.user_defined_strategy.gradient_merge_configs
     self.wrapped_opt = GM(self.inner_opt)
     self.wrapped_opt._set_k_steps(
         self.user_defined_strategy.gradient_merge_configs["k_steps"])
     self.wrapped_opt._set_avg(
         self.user_defined_strategy.gradient_merge_configs["avg"])
 def __init__(self, optimizer):
     super(GradientMergeOptimizer, self).__init__(optimizer)
     self.inner_opt = optimizer
     self.wrapped_opt = GM(optimizer)
     self.meta_optimizers_white_list = [
         "LarsOptimizer",
         "LambOptimizer",
         "GraphExecutionOptimizer",
     ]
     self.meta_optimizers_black_list = []