diff --git a/engine/data/transforms/_transforms.py b/engine/data/transforms/_transforms.py
index 31588df5..b8318031 100644
--- a/engine/data/transforms/_transforms.py
+++ b/engine/data/transforms/_transforms.py
@@ -70,6 +70,10 @@ def _transform(self, inpt: Any, params: Dict[str, Any]) -> Any:
         fill = self._fill[type(inpt)]
         padding = params['padding']
         return F.pad(inpt, padding=padding, fill=fill, padding_mode=self.padding_mode)  # type: ignore[arg-type]
+    
+    # Compatibility with torchvision>0.20.0
+    def transform(self, inpt: Any, params: Dict[str, Any]) -> Any:
+        return self._transform(inpt, params)
 
     def __call__(self, *inputs: Any) -> Any:
         outputs = super().forward(*inputs)
@@ -91,6 +95,7 @@ def __call__(self, *inputs: Any) -> Any:
         return super().forward(*inputs)
 
 
+
 @register()
 class ConvertBoxes(T.Transform):
     _transformed_types = (
@@ -112,6 +117,19 @@ def _transform(self, inpt: Any, params: Dict[str, Any]) -> Any:
             inpt = inpt / torch.tensor(spatial_size[::-1]).tile(2)[None]
 
         return inpt
+    
+    # Compatibility with torchvision>0.20.0
+    def transform(self, inpt: Any, params: Dict[str, Any]) -> Any:
+        spatial_size = getattr(inpt, _boxes_keys[1])
+        if self.fmt:
+            in_fmt = inpt.format.value.lower()
+            inpt = torchvision.ops.box_convert(inpt, in_fmt=in_fmt, out_fmt=self.fmt.lower())
+            inpt = convert_to_tv_tensor(inpt, key='boxes', box_format=self.fmt.upper(), spatial_size=spatial_size)
+
+        if self.normalize:
+            inpt = inpt / torch.tensor(spatial_size[::-1]).tile(2)[None]
+
+        return inpt
 
 
 @register()
@@ -135,3 +153,16 @@ def _transform(self, inpt: Any, params: Dict[str, Any]) -> Any:
         inpt = Image(inpt)
 
         return inpt
+    
+    # Compatibility with torchvision>0.20.0
+    def transform(self, inpt: Any, params: Dict[str, Any]) -> Any:
+        inpt = F.pil_to_tensor(inpt)
+        if self.dtype == 'float32':
+            inpt = inpt.float()
+
+        if self.scale:
+            inpt = inpt / 255.
+
+        inpt = Image(inpt)
+
+        return inpt