Added analyze.class_name_id_map property

pylabel/analyze.py

@@ -82,6 +82,19 @@ def num_images(self):
         """
         return self.dataset.df["img_filename"].nunique()
 
+    @property
+    def class_name_id_map(self):
+        """Returns a dict where the class name is the key and class id is the value.
+
+        Returns:
+            Dict
+
+        Example:
+            >>> dataset.analyze.class_name_id_map()
+            {('Squirrel', 0),('Nut', 1)}
+        """
+        return dict(zip(self.dataset.df.cat_name, self.dataset.df.cat_id))
+
     def ShowClassSplits(self, normalize=True):
         """Show the distribution of classes across train, val, and
         test splits of the dataset.

pylabel/exporter.py

@@ -452,7 +452,9 @@ def ExportToYoloV5(
         # Note, having zero annotates can still be considered annotated
         # in cases when are no objects in the image thats should be indentified
         yolo_dataset = yolo_dataset.loc[yolo_dataset["annotated"] == 1]
-        yolo_dataset["cat_id"] = yolo_dataset["cat_id"].astype(pd.Int32Dtype())
+        yolo_dataset["cat_id"] = (
+            yolo_dataset["cat_id"].astype("float").astype(pd.Int32Dtype())
+        )
 
         if cat_id_index != None:
             assert isinstance(cat_id_index, int), "cat_id_index must be an int."