- fix issues in tensorboard logger

inferno/trainers/callbacks/logging/tensorboard.py

@@ -199,6 +199,8 @@ def log_object(self, tag, object_,
                                 allow_image_logging,
                                 allow_histogram_logging)
             return
+
+        # FIXME this can throw ugly warnings
         # Check whether object is a scalar
         if tu.is_scalar_tensor(object_) and allow_scalar_logging:
             # Log scalar
@@ -211,9 +213,10 @@ def log_object(self, tag, object_,
             # Add a channel axis and log as images
             self.log_image_or_volume_batch(tag, object_[:, None, ...],
                                            self.trainer.iteration_count)
-        elif tu.is_image_or_volume_tensor(object_) and allow_image_logging:
-            # Log images
-            self.log_image_or_volume_batch(tag, object_, self.trainer.iteration_count)
+        elif tu.is_image_or_volume_tensor(object_):
+            if allow_image_logging:
+                # Log images
+                self.log_image_or_volume_batch(tag, object_, self.trainer.iteration_count)
         elif tu.is_vector_tensor(object_) and allow_histogram_logging:
             # Log histograms
             values = tu.unwrap(object_, as_numpy=True)
@@ -388,12 +391,15 @@ def log_images(self, tag, images, step, image_format='CHW'):
             if image.shape[-1] == 1:
                 image = image[..., [0, 0, 0]]
             image = self._normalize_image(image)
+            # print(image.dtype, image.shape)
             self.writer.add_image(tag, img_tensor=image, global_step=step)
 
     @staticmethod
     def _normalize_image(image):
         normalized_image = image - image.min()
-        normalized_image = normalized_image / normalized_image.max()
+        maxval = normalized_image.max()
+        if maxval > 0:
+            normalized_image = normalized_image / maxval
         return normalized_image
 
     def log_histogram(self, tag, values, step, bins=1000):