deepvisiontools 0.0.2__tar.gz → 0.0.4__tar.gz

{deepvisiontools-0.0.2 → deepvisiontools-0.0.4}/PKG-INFO

@@ -1,6 +1,6 @@
 Metadata-Version: 2.2
 Name: deepvisiontools
-Version: 0.0.2
+Version: 0.0.4
 Summary: Overlay of PyTorch to generalize trainning and inference processes for detection & instance segmentation tasks.
 Author-email: Lucas Bernigaud Samatan <lucas.bernigaud-samatan@inrae.fr>, Jordan Bernigaud Samatan <jordan.bernigaud-samatan@inrae.fr>
 Maintainer-email: Jordan Bernigaud Samatan <jordan.bernigaud-samatan@inrae.fr>

{deepvisiontools-0.0.2 → deepvisiontools-0.0.4}/src/deepvisiontools/__init__.py

@@ -4,7 +4,7 @@ from deepvisiontools.data.dataset import DeepVisionDataset, DeepVisionLoader
 from deepvisiontools.utils import visualization
 from deepvisiontools.inference.predictor import Predictor
 
-__version__ = "0.0.2"
+__version__ = "0.0.4"
 
 __all__ = (
     Configuration,

{deepvisiontools-0.0.2 → deepvisiontools-0.0.4}/src/deepvisiontools/formats/base_data.py

@@ -135,6 +135,13 @@ class SemanticMaskData(BaseData):
 
     def __init__(self, mask: Union[Tensor, Mask]):
         mask = mask.long() if isinstance(mask, Mask) else Mask(mask.long())
+        assert (
+            mask.dim() == 2 or mask.dim() == 3
+        ), f"SemanticMaskData : input tensor should be of dim 2 or 3 [1, H, W] (anyway casted to [H, W]), got {mask.shape}"
+
+        # cast mask to [Ncls, H, W] if dim == 2
+        if mask.dim() == 3:
+            mask = Mask(mask[0])
         self._value = copy.deepcopy(mask)  # convert to long
         self._canvas_size = (mask.shape[-2], mask.shape[-1])
         self._nb_object = torch.unique(mask).shape[0] - 1
@@ -153,6 +160,13 @@ class SemanticMaskData(BaseData):
             val, Tensor
         ), f"SemanticMaskData value must be Tensor, got {type(val)}"
         mask = val.long()
+        assert (
+            mask.dim() == 2 or mask.dim() == 3
+        ), f"SemanticMaskData : input tensor should be of dim 2 or 3 (anyway casted to [Ncls, H, W]), got {mask.shape}"
+
+        # cast mask to [Ncls, H, W] if dim == 2
+        if mask.dim() == 2:
+            mask = mask[None, :]
         # covert to torchvision tv_tensor
         mask = mask if isinstance(mask, Mask) else Mask(mask)
         mask = mask.to(self.device)

{deepvisiontools-0.0.2 → deepvisiontools-0.0.4}/src/deepvisiontools/train/trainer.py

@@ -259,7 +259,7 @@ class Trainer:
 
         log = ""
         for key, value in flattened_dict.items():
-            log += f"{key} : {str(round(value.item(), 3))} "
+            log += f"{key} : {str(round(value.item(), 4))} "
 
         return log
 

{deepvisiontools-0.0.2 → deepvisiontools-0.0.4}/src/deepvisiontools.egg-info/PKG-INFO

@@ -1,6 +1,6 @@
 Metadata-Version: 2.2
 Name: deepvisiontools
-Version: 0.0.2
+Version: 0.0.4
 Summary: Overlay of PyTorch to generalize trainning and inference processes for detection & instance segmentation tasks.
 Author-email: Lucas Bernigaud Samatan <lucas.bernigaud-samatan@inrae.fr>, Jordan Bernigaud Samatan <jordan.bernigaud-samatan@inrae.fr>
 Maintainer-email: Jordan Bernigaud Samatan <jordan.bernigaud-samatan@inrae.fr>