Update recipes to use RandomResizedCrop instead of RandomResize a…

…nd `RandomCrop`
openvinotoolkit · Jul 12, 2024 · 2e32819 · 2e32819
1 parent 61e8550
commit 2e32819
Show file tree

Hide file tree

Showing 2 changed files with 16 additions and 40 deletions.
diff --git a/src/otx/recipe/_base_/data/semantic_segmentation.yaml b/src/otx/recipe/_base_/data/semantic_segmentation.yaml
@@ -13,34 +13,22 @@ train_subset:
   transform_lib_type: TORCHVISION
   to_tv_image: true
   transforms:
-    - class_path: otx.core.data.transform_libs.torchvision.RandomResize
+    - class_path: otx.core.data.transform_libs.torchvision.RandomResizedCrop
       init_args:
         scale:
-          - 544
-          - 544
-        ratio_range:
+          - 512
+          - 512
+        crop_ratio_range:
+          - 0.2
+          - 1.0
+        aspect_ratio_range:
           - 0.5
           - 2.0
         transform_mask: true
-    - class_path: otx.core.data.transform_libs.torchvision.RandomCrop
-      init_args:
-        crop_size:
-          - 512
-          - 512
-        cat_max_ratio: 0.75
+    - class_path: otx.core.data.transform_libs.torchvision.PhotoMetricDistortion
     - class_path: otx.core.data.transform_libs.torchvision.RandomFlip
       init_args:
         prob: 0.5
-    - class_path: otx.core.data.transform_libs.torchvision.PhotoMetricDistortion
-    - class_path: otx.core.data.transform_libs.torchvision.Pad
-      init_args:
-        size:
-          - 512
-          - 512
-        pad_val:
-          img: 0
-          mask: 255
-        transform_mask: true
         is_numpy_to_tvtensor: true
     - class_path: torchvision.transforms.v2.ToDtype
       init_args:

diff --git a/src/otx/recipe/semantic_segmentation/dino_v2.yaml b/src/otx/recipe/semantic_segmentation/dino_v2.yaml
@@ -42,34 +42,22 @@ overrides:
   data:
     train_subset:
       transforms:
-        - class_path: otx.core.data.transform_libs.torchvision.RandomResize
+        - class_path: otx.core.data.transform_libs.torchvision.RandomResizedCrop
           init_args:
             scale:
-              - 640
-              - 640
-            ratio_range:
+              - 560
+              - 560
+            crop_ratio_range:
+              - 0.2
+              - 1.0
+            aspect_ratio_range:
               - 0.5
               - 2.0
             transform_mask: true
-        - class_path: otx.core.data.transform_libs.torchvision.RandomCrop
-          init_args:
-            crop_size:
-              - 560
-              - 560
-            cat_max_ratio: 0.75
+        - class_path: otx.core.data.transform_libs.torchvision.PhotoMetricDistortion
         - class_path: otx.core.data.transform_libs.torchvision.RandomFlip
           init_args:
             prob: 0.5
-        - class_path: otx.core.data.transform_libs.torchvision.PhotoMetricDistortion
-        - class_path: otx.core.data.transform_libs.torchvision.Pad
-          init_args:
-            size:
-              - 560
-              - 560
-            pad_val:
-              img: 0
-              mask: 255
-            transform_mask: true
             is_numpy_to_tvtensor: true
         - class_path: torchvision.transforms.v2.ToDtype
           init_args: