diff --git a/icenet_mp/config/base_isambardai.yaml b/icenet_mp/config/base_isambardai.yaml
index 62f73011a..d47d1d808 100644
--- a/icenet_mp/config/base_isambardai.yaml
+++ b/icenet_mp/config/base_isambardai.yaml
@@ -2,4 +2,4 @@ defaults:
   - base
   - _self_
 
-base_path: /projects/u5gf/seaice/base/
+base_path: /projects/public/u6iz/shared/base/
diff --git a/icenet_mp/losses/weighted_mse_loss.py b/icenet_mp/losses/weighted_mse_loss.py
index 880ef24a2..84f8be6ea 100644
--- a/icenet_mp/losses/weighted_mse_loss.py
+++ b/icenet_mp/losses/weighted_mse_loss.py
@@ -20,6 +20,7 @@ def __init__(self, *args: Any, **kwargs: Any) -> None:
             **kwargs: Keyword arguments passed to torch.nn.MSELoss.
 
         """
+        kwargs["reduction"] = "none"
         super().__init__(*args, **kwargs)
 
     def forward(  # type: ignore[override]
@@ -43,5 +44,5 @@ def forward(  # type: ignore[override]
         targets = targets.squeeze()
         sample_weights = sample_weights.squeeze()
 
-        loss = super().forward(100 * y_hat, 100 * targets) * sample_weights
+        loss = super().forward(y_hat, targets) * sample_weights
         return loss.mean()
diff --git a/icenet_mp/model_service.py b/icenet_mp/model_service.py
index 9d60cc0e5..e456a3272 100644
--- a/icenet_mp/model_service.py
+++ b/icenet_mp/model_service.py
@@ -1,4 +1,5 @@
 import logging
+import os
 from pathlib import Path, PosixPath
 from typing import cast
 
@@ -23,8 +24,13 @@ class ModelService:
     def __init__(self, config: DictConfig) -> None:
         """Initialize the model service."""
         self.config_ = config
-        if seed := config.get("seed", None):
-            seed_everything(int(seed), workers=True)
+        if (seed := config.get("seed", None)) is not None:
+            seed = int(seed)
+            os.environ["PYTHONHASHSEED"] = str(seed)
+            os.environ["CUBLAS_WORKSPACE_CONFIG"] = ":4096:8"
+            seed_everything(seed, workers=True)
+            torch.use_deterministic_algorithms(True, warn_only=True)  # noqa: FBT003
+
         self.data_module_: CommonDataModule | None = None
         self.model_: BaseModel | None = None
 
@@ -184,6 +190,20 @@ def build_trainer(
                 )
             ),
         )
+        # Re-apply warn_only since Lightning may override it when deterministic=True
+        if self.config.get("seed", None):
+            torch.use_deterministic_algorithms(True, warn_only=True)  # noqa: FBT003
+
+        # Check warn_only survived Lightning's deterministic setup
+        log.debug(
+            "deterministic_algorithms_enabled: %s",
+            torch.are_deterministic_algorithms_enabled(),
+        )
+        log.debug(
+            "warn_only_enabled: %s",
+            torch.is_deterministic_algorithms_warn_only_enabled(),
+        )
+
         # Assign workers for data loading
         self.data_module.assign_workers(suggested_max_num_workers(trainer.num_devices))