fix(dpmodel): rmse_e loss normalized by n_atoms (#5124)

Consistent with other backends.

<!-- This is an auto-generated comment: release notes by coderabbit.ai
-->

## Summary by CodeRabbit

* **Bug Fixes**
* Improved energy RMSE metric display by normalizing reported values per
atom, ensuring more accurate representation in loss monitoring.

<sub>✏️ Tip: You can customize this high-level summary in your review
settings.</sub>

<!-- end of auto-generated comment: release notes by coderabbit.ai -->

Author: njzjz

deepmd/dpmodel/loss/ener.py

@@ -180,7 +180,7 @@ def call(
                 )
                 loss += pref_e * l_huber_loss
             more_loss["rmse_e"] = self.display_if_exist(
-                xp.sqrt(l2_ener_loss), find_energy
+                xp.sqrt(l2_ener_loss) * atom_norm_ener, find_energy
             )
         if self.has_f:
             l2_force_loss = xp.mean(xp.square(diff_f))