Speed up preloading by not loading pixel data

PIL lazy loads pixel data, so image size can be accessed without loading the full image.
This commit makes it so only image size and EXIF metadata are fetched from disk, speeding up the preload stage while still supporting transposing images.

data/image_train_item.py

@@ -162,6 +162,29 @@ class ImageTrainItem:
             pass
         return image
     
+    def _needs_transpose(self, image, print_error=False):
+        try:
+            exif = image.getexif()
+            orientation = exif.get(0x0112)
+            """
+                https://pillow.readthedocs.io/en/stable/_modules/PIL/ImageOps.html#exif_transpose
+                method = {
+                    2: Image.Transpose.FLIP_LEFT_RIGHT,
+                    3: Image.Transpose.ROTATE_180,
+                    4: Image.Transpose.FLIP_TOP_BOTTOM,
+                    5: Image.Transpose.TRANSPOSE,
+                    6: Image.Transpose.ROTATE_270,
+                    7: Image.Transpose.TRANSVERSE,
+                    8: Image.Transpose.ROTATE_90,
+                }.get(orientation)
+            """
+            return orientation in [5, 6, 7, 8]
+        except Exception as e:
+            logging.warning(F"Error rotating image: {e} on {self.pathname}, image will be loaded as is, EXIF may be corrupt") if print_error else None
+            pass
+        return False
+
+
     def _percent_random_crop(self, image, crop_jitter=0.02):
         """
         randomly crops the image by a percentage of the image size on each of the four sides
@@ -270,7 +293,10 @@ class ImageTrainItem:
         self.target_wh = None
         try:
             with PIL.Image.open(self.pathname) as image:
-                image = self._try_transpose(image, print_error=True).convert('RGB')
+                needs_transpose = self._needs_transpose(image)
+                if needs_transpose:
+                    height, width = image.size
+                else:
                     width, height = image.size
                 image_aspect = width / height
                 target_wh = min(self.aspects, key=lambda aspects:abs(aspects[0]/aspects[1] - image_aspect))