415be78cc0d1275a29d0ceda550d0d7a7a5224ea,python/ray/util/sgd/torch/training_operator.py,TrainingOperator,train_epoch,#TrainingOperator#Any#Any#,120

Before Change

            metric_meters.update(metrics, n=metrics.pop(NUM_SAMPLES, 1))
            self.global_step += 1

        if self.scheduler and info.get(SCHEDULER_STEP) == SCHEDULER_STEP_EPOCH:
            self.scheduler.step()

        return metric_meters.summary()

After Change

        Returns:
            A dict of metrics from training.
        
        if not hasattr(self, "model"):
            raise RuntimeError("Either set self.model in setup function or "
                               "override this method to implement a custom "
                               "training loop.")
        model = self.model
        scheduler = None
        if hasattr(self, "scheduler"):
            scheduler = self.scheduler

        if self.use_tqdm and self.world_rank == 0:
            desc = ""
            if info is not None and "epoch_idx" in info:
                if "num_epochs" in info:
                    desc = f"{info["epoch_idx"] + 1}/{info["num_epochs"]}e"
                else:
                    desc = f"{info["epoch_idx"] + 1}e"

            // TODO: Implement len for Dataset?
            total = info[NUM_STEPS]
            if total is None:
                if hasattr(iterator, "__len__"):
                    total = len(iterator)

            _progress_bar = tqdm(
                total=total, desc=desc, unit="batch", leave=False)

        metric_meters = AverageMeterCollection()

Instances