7/21 - [17/10K] - preds = model(sweep_imgs, mats, ...)

쉽지 않다.......

아래 처럼 실행 후 preds를 구해도 128x128의 값이기 때문에 다시 박스를 계산해야한다.

model = model.to(device)        
preds = model(sweep_imgs, mats,
                pts_pv=pts_pv,
                is_train=False)

 

너무 복잡하다 이게 하나의 로직이다.

단순 그냥 박스 그리는것이 아니였다. 쩝......

    def eval_step(self, batch, batch_idx, prefix: str):
        (sweep_imgs, mats, img_metas, _, _, _, _, pts_pv) = batch
        if torch.cuda.is_available():
            if self.return_image:
                sweep_imgs = sweep_imgs.cuda()
                for key, value in mats.items():
                    mats[key] = value.cuda()
            if self.return_radar_pv:
                pts_pv = pts_pv.cuda()
        preds = self(sweep_imgs, mats,
                     pts_pv=pts_pv,
                     is_train=False)
        if isinstance(self.model, torch.nn.parallel.DistributedDataParallel):
            results = self.model.module.get_bboxes(preds, img_metas)
        else:
            results = self.model.get_bboxes(preds, img_metas)
        for i in range(len(results)):
            results[i][0] = results[i][0].tensor.detach().cpu().numpy()
            results[i][1] = results[i][1].detach().cpu().numpy()
            results[i][2] = results[i][2].detach().cpu().numpy()
            results[i].append(img_metas[i])
        return results

   def test_epoch_end(self, test_step_outputs):
        all_pred_results = list()
        all_img_metas = list()
        for test_step_output in test_step_outputs:
            for i in range(len(test_step_output)):
                all_pred_results.append(test_step_output[i][:3])
                all_img_metas.append(test_step_output[i][3])
        synchronize()
        # TODO: Change another way.
        dataset_length = len(self.val_dataloader().dataset)
        all_pred_results = sum(
            map(list, zip(*all_gather_object(all_pred_results))),
            [])[:dataset_length]
        all_img_metas = sum(map(list, zip(*all_gather_object(all_img_metas))),
                            [])[:dataset_length]
        if self.global_rank == 0:
            self.evaluator.evaluate(all_pred_results, all_img_metas)

    def evaluate(
        self,
        results,
        img_metas,
        metric='bbox',
        logger=None,
        jsonfile_prefix=None,
        result_names=['img_bbox'],
        show=False,
        out_dir=None,
        pipeline=None,
    ):
        """Evaluation in nuScenes protocol.

        Args:
            results (list[dict]): Testing results of the dataset.
            metric (str | list[str]): Metrics to be evaluated.
            logger (logging.Logger | str | None): Logger used for printing
                related information during evaluation. Default: None.
            jsonfile_prefix (str | None): The prefix of json files. It includes
                the file path and the prefix of filename, e.g., "a/b/prefix".
                If not specified, a temp file will be created. Default: None.
            show (bool): Whether to visualize.
                Default: False.
            out_dir (str): Path to save the visualization results.
                Default: None.
            pipeline (list[dict], optional): raw data loading for showing.
                Default: None.

        Returns:
            dict[str, float]: Results of each evaluation metric.
        """
        result_files, tmp_dir = self.format_results(results, img_metas,
                                                    result_names,
                                                    jsonfile_prefix)