Fix tiling test (#3637)

* fix tile test

* add description to tile transform

src/otx/core/data/dataset/tile.py

@@ -64,6 +64,7 @@ class OTXTileTransform(Tile):
         extractor (DmDataset): Dataset subset to extract tiles from.
         tile_size (tuple[int, int]): Tile size.
         overlap (tuple[float, float]): Overlap ratio.
+            Overlap values are clipped between 0 and 0.9 to ensure the stride is not too small.
         threshold_drop_ann (float): Threshold to drop annotations.
         with_full_img (bool): Include full image in the tiles.
     """
@@ -76,6 +77,8 @@ def __init__(
         threshold_drop_ann: float,
         with_full_img: bool,
     ) -> None:
+        # NOTE: clip overlap to [0, 0.9]
+        overlap = max(0, min(overlap[0], 0.9)), max(0, min(overlap[1], 0.9))
         super().__init__(
             extractor,
             (0, 0),

src/otx/core/data/tile_adaptor.py

@@ -173,8 +173,8 @@ def adapt_tile_config(tile_config: TileConfig, dataset: Dataset) -> None:
 
         if tile_overlap >= 0.9:
             # Use the average object area if the tile overlap is too large to prevent 0 stride.
-            tile_overlap = avg_size / tile_size
-            log.info(f"----> (too big) tile_overlap: {avg_size} / {tile_size} = {tile_overlap}")
+            tile_overlap = min(avg_size / tile_size, 0.9)
+            log.info(f"----> (too big) tile_overlap: {avg_size} / {tile_size} = min[{tile_overlap}, 0.9]")
 
         # TODO(Eugene): how to validate lower/upper_bound? dataclass? pydantic?
         # https://github.com/openvinotoolkit/training_extensions/pull/2903

tests/unit/core/data/test_tiling.py

@@ -28,6 +28,7 @@
 from otx.core.model.detection import OTXDetectionModel
 from otx.core.model.instance_segmentation import OTXInstanceSegModel
 from otx.core.types.task import OTXTaskType
+from otx.core.types.transformer_libs import TransformLibType
 from torchvision import tv_tensors
 
 from tests.test_helpers import generate_random_bboxes
@@ -39,12 +40,12 @@ def mock_otx_det_model(self) -> OTXDetectionModel:
         return create_autospec(OTXDetectionModel)
 
     @pytest.fixture()
-    def fxt_mmcv_det_transform_config(self) -> list[DictConfig]:
-        mmdet_base = OmegaConf.load("src/otx/recipe/_base_/data/mmdet_base.yaml")
+    def fxt_tv_det_transform_config(self) -> list[DictConfig]:
+        mmdet_base = OmegaConf.load("src/otx/recipe/_base_/data/torchvision_base.yaml")
         return mmdet_base.config.train_subset.transforms
 
     @pytest.fixture()
-    def fxt_det_data_config(self, fxt_mmcv_det_transform_config) -> OTXDataModule:
+    def fxt_det_data_config(self, fxt_tv_det_transform_config) -> OTXDataModule:
         data_root = Path(__file__).parent.parent.parent.parent / "assets" / "car_tree_bug"
 
         batch_size = 8
@@ -56,35 +57,31 @@ def fxt_det_data_config(self, fxt_mmcv_det_transform_config) -> OTXDataModule:
                 subset_name="train",
                 batch_size=batch_size,
                 num_workers=num_workers,
-                transform_lib_type="MMDET",
-                transforms=fxt_mmcv_det_transform_config,
+                transform_lib_type=TransformLibType.TORCHVISION,
+                transforms=fxt_tv_det_transform_config,
             ),
             val_subset=SubsetConfig(
                 subset_name="val",
                 batch_size=batch_size,
                 num_workers=num_workers,
-                transform_lib_type="MMDET",
-                transforms=fxt_mmcv_det_transform_config,
+                transform_lib_type=TransformLibType.TORCHVISION,
+                transforms=fxt_tv_det_transform_config,
             ),
             test_subset=SubsetConfig(
                 subset_name="test",
                 batch_size=batch_size,
                 num_workers=num_workers,
-                transform_lib_type="MMDET",
-                transforms=fxt_mmcv_det_transform_config,
+                transform_lib_type=TransformLibType.TORCHVISION,
+                transforms=fxt_tv_det_transform_config,
             ),
             tile_config=TileConfig(),
             vpm_config=VisualPromptingConfig(),
         )
 
     @pytest.fixture()
-    def fxt_instseg_data_config(self, fxt_mmcv_det_transform_config) -> OTXDataModule:
+    def fxt_instseg_data_config(self, fxt_tv_det_transform_config) -> OTXDataModule:
         data_root = Path(__file__).parent.parent.parent.parent / "assets" / "car_tree_bug"
 
-        for transform in fxt_mmcv_det_transform_config:
-            if transform.type == "LoadAnnotations":
-                transform.with_mask = True
-
         batch_size = 8
         num_workers = 0
         return DataModuleConfig(
@@ -94,22 +91,22 @@ def fxt_instseg_data_config(self, fxt_mmcv_det_transform_config) -> OTXDataModul
                 subset_name="train",
                 batch_size=batch_size,
                 num_workers=num_workers,
-                transform_lib_type="MMDET",
-                transforms=fxt_mmcv_det_transform_config,
+                transform_lib_type=TransformLibType.TORCHVISION,
+                transforms=fxt_tv_det_transform_config,
             ),
             val_subset=SubsetConfig(
                 subset_name="val",
                 batch_size=batch_size,
                 num_workers=num_workers,
-                transform_lib_type="MMDET",
-                transforms=fxt_mmcv_det_transform_config,
+                transform_lib_type=TransformLibType.TORCHVISION,
+                transforms=fxt_tv_det_transform_config,
             ),
             test_subset=SubsetConfig(
                 subset_name="test",
                 batch_size=batch_size,
                 num_workers=num_workers,
-                transform_lib_type="MMDET",
-                transforms=fxt_mmcv_det_transform_config,
+                transform_lib_type=TransformLibType.TORCHVISION,
+                transforms=fxt_tv_det_transform_config,
             ),
             tile_config=TileConfig(),
             vpm_config=VisualPromptingConfig(),
@@ -239,6 +236,7 @@ def test_tile_transform(self):
         rng = np.random.default_rng()
         tile_size = rng.integers(low=100, high=500, size=(2,))
         overlap = rng.random(2)
+        overlap = overlap.clip(0, 0.9)
         threshold_drop_ann = rng.random()
         tiled_dataset = DmDataset.import_from("tests/assets/car_tree_bug", format="coco_instances")
         tiled_dataset.transform(