diff --git a/README.md b/README.md
index 2b2203373..95ba60a48 100644
--- a/README.md
+++ b/README.md
@@ -1,6 +1,6 @@
 <p align="center">
     <!-- community badges -->
-    <a href="https://discord.gg/uMbNqcraFc"><img src="https://img.shields.io/badge/Join-Discord-blue.svg"/></a>
+    <a href="https://discord.gg/uMbNqcraFc"><img src="https://dcbadge.vercel.app/api/server/uMbNqcraFc?style=plastic"/></a>
     <!-- doc badges -->
     <a href='https://docs.nerf.studio/'>
         <img src='https://readthedocs.com/projects/plenoptix-nerfstudio/badge/?version=latest' alt='Documentation Status' /></a>
diff --git a/docs/developer_guides/viewer/local_viewer.md b/docs/developer_guides/viewer/local_viewer.md
index 2931f8c0d..ef8b1ac1f 100644
--- a/docs/developer_guides/viewer/local_viewer.md
+++ b/docs/developer_guides/viewer/local_viewer.md
@@ -1,4 +1,8 @@
-# Local Server
+# (Legacy Viewer) Local Server
+
+**Note:** this doc only applies to the legacy version of the viewer, which was the default in in Nerfstudio versions `<=0.3.4`. It was deprecated starting Nerfstudio version `1.0.0`, where it needs to be opted into via the `--vis viewer_legacy` argument.
+
+---
 
 If you are unable to connect to `https://viewer.nerf.studio`, want to use Safari, or want to develop the viewer codebase, you can launch your own local viewer.
 
diff --git a/nerfstudio/models/splatfacto.py b/nerfstudio/models/splatfacto.py
index b23eabe00..a826599f9 100644
--- a/nerfstudio/models/splatfacto.py
+++ b/nerfstudio/models/splatfacto.py
@@ -155,6 +155,8 @@ class SplatfactoModelConfig(ModelConfig):
     """threshold of ratio of gaussian max to min scale before applying regularization
     loss from the PhysGaussian paper
     """
+    output_depth_during_training: bool = False
+    """If True, output depth during training. Otherwise, only output depth during evaluation."""
 
 
 class SplatfactoModel(Model):
@@ -667,7 +669,10 @@ class SplatfactoModel(Model):
         if self.crop_box is not None and not self.training:
             crop_ids = self.crop_box.within(self.means).squeeze()
             if crop_ids.sum() == 0:
-                return {"rgb": background.repeat(int(camera.height.item()), int(camera.width.item()), 1)}
+                rgb = background.repeat(int(camera.height.item()), int(camera.width.item()), 1)
+                depth = background.new_ones(*rgb.shape[:2], 1) * 10
+                accumulation = background.new_zeros(*rgb.shape[:2], 1)
+                return {"rgb": rgb, "depth": depth, "accumulation": accumulation}
         else:
             crop_ids = None
         camera_downscale = self._get_downscale_factor()
@@ -732,7 +737,10 @@ class SplatfactoModel(Model):
             tile_bounds,
         )  # type: ignore
         if (self.radii).sum() == 0:
-            return {"rgb": background.repeat(int(camera.height.item()), int(camera.width.item()), 1)}
+            rgb = background.repeat(int(camera.height.item()), int(camera.width.item()), 1)
+            depth = background.new_ones(*rgb.shape[:2], 1) * 10
+            accumulation = background.new_zeros(*rgb.shape[:2], 1)
+            return {"rgb": rgb, "depth": depth, "accumulation": accumulation}
 
         # Important to allow xys grads to populate properly
         if self.training:
@@ -766,7 +774,7 @@ class SplatfactoModel(Model):
         alpha = alpha[..., None]
         rgb = torch.clamp(rgb, max=1.0)  # type: ignore
         depth_im = None
-        if not self.training:
+        if self.config.output_depth_during_training or not self.training:
             depth_im = rasterize_gaussians(  # type: ignore
                 self.xys,
                 depths,