diff --git a/src/batdetect2/api_v2.py b/src/batdetect2/api_v2.py
index 8e8cae5..a714ec6 100644
--- a/src/batdetect2/api_v2.py
+++ b/src/batdetect2/api_v2.py
@@ -484,11 +484,19 @@ class BatDetect2API:
             transform=output_transform,
         )
 
+        # NOTE: Build separate instances of preprocessor and postprocessor
+        # to avoid device mismatch errors
         model = build_model(
             config=config.model,
-            targets=targets,
-            preprocessor=preprocessor,
-            postprocessor=postprocessor,
+            targets=build_targets(config=config.model.targets),
+            preprocessor=build_preprocessor(
+                input_samplerate=audio_loader.samplerate,
+                config=config.model.preprocess,
+            ),
+            postprocessor=build_postprocessor(
+                preprocessor,
+                config=config.model.postprocess,
+            ),
         )
 
         return cls(
diff --git a/src/batdetect2/preprocess/types.py b/src/batdetect2/preprocess/types.py
index dbf3758..39485e9 100644
--- a/src/batdetect2/preprocess/types.py
+++ b/src/batdetect2/preprocess/types.py
@@ -29,6 +29,3 @@ class PreprocessorProtocol(Protocol):
 
     def process_numpy(self, wav: np.ndarray) -> np.ndarray:
         return self(torch.tensor(wav)).numpy()
-
-    def cpu(self) -> "PreprocessorProtocol":
-        return self
diff --git a/src/batdetect2/train/dataset.py b/src/batdetect2/train/dataset.py
index 570106d..e9e1883 100644
--- a/src/batdetect2/train/dataset.py
+++ b/src/batdetect2/train/dataset.py
@@ -272,7 +272,7 @@ def build_train_dataset(
         audio_loader=audio_loader,
         labeller=labeller,
         clipper=clipper,
-        preprocessor=preprocessor.cpu(),
+        preprocessor=preprocessor,
         audio_augmentation=audio_augmentation,
         spectrogram_augmentation=spectrogram_augmentation,
     )
@@ -305,7 +305,7 @@ def build_val_dataset(
         clip_annotations,
         audio_loader=audio_loader,
         labeller=labeller,
-        preprocessor=preprocessor.cpu(),
+        preprocessor=preprocessor,
         clipper=clipper,
     )
 
diff --git a/src/batdetect2/train/train.py b/src/batdetect2/train/train.py
index 1f1745a..2842738 100644
--- a/src/batdetect2/train/train.py
+++ b/src/batdetect2/train/train.py
@@ -73,9 +73,6 @@ def run_train(
 
     audio_loader = audio_loader or build_audio_loader(config=audio_config)
 
-    if model is not None:
-        preprocessor = preprocessor or model.preprocessor
-
     preprocessor = preprocessor or build_preprocessor(
         input_samplerate=audio_loader.samplerate,
         config=model_config.preprocess,