Fix model loading - use device param instead of .to()

AudioCraft's get_pretrained() accepts device parameter directly.
The model objects don't have a .to() method.

Fixed in all adapters:
- MusicGen
- AudioGen
- MAGNeT
- MusicGen Style
- JASCO

🤖 Generated with [Claude Code](https://claude.com/claude-code)

Co-Authored-By: Claude <noreply@anthropic.com>

src/models/audiogen/adapter.py

@@ -104,8 +104,7 @@ class AudioGenAdapter(BaseAudioModel):
             from audiocraft.models import AudioGen
 
             self._device = torch.device(device)
-            self._model = AudioGen.get_pretrained(self._config["hf_id"])
+            self._model = AudioGen.get_pretrained(self._config["hf_id"], device=device)
-            self._model.to(self._device)
 
             logger.info(
                 f"AudioGen {self._variant} loaded successfully "

src/models/jasco/adapter.py

@@ -116,8 +116,7 @@ class JASCOAdapter(BaseAudioModel):
             from audiocraft.models import JASCO
 
             self._device = torch.device(device)
-            self._model = JASCO.get_pretrained(self._config["hf_id"])
+            self._model = JASCO.get_pretrained(self._config["hf_id"], device=device)
-            self._model.to(self._device)
 
             logger.info(
                 f"JASCO {self._variant} loaded successfully "

src/models/magnet/adapter.py

@@ -142,8 +142,7 @@ class MAGNeTAdapter(BaseAudioModel):
             from audiocraft.models import MAGNeT
 
             self._device = torch.device(device)
-            self._model = MAGNeT.get_pretrained(self._config["hf_id"])
+            self._model = MAGNeT.get_pretrained(self._config["hf_id"], device=device)
-            self._model.to(self._device)
 
             logger.info(
                 f"MAGNeT {self._variant} loaded successfully "

src/models/musicgen/adapter.py

@@ -168,8 +168,8 @@ class MusicGenAdapter(BaseAudioModel):
             from audiocraft.models import MusicGen
 
             self._device = torch.device(device)
-            self._model = MusicGen.get_pretrained(self._config["hf_id"])
+            # MusicGen.get_pretrained() accepts device parameter directly
-            self._model.to(self._device)
+            self._model = MusicGen.get_pretrained(self._config["hf_id"], device=device)
 
             logger.info(
                 f"MusicGen {self._variant} loaded successfully "

src/models/musicgen_style/adapter.py

@@ -105,8 +105,7 @@ class MusicGenStyleAdapter(BaseAudioModel):
             from audiocraft.models import MusicGen
 
             self._device = torch.device(device)
-            self._model = MusicGen.get_pretrained(self._config["hf_id"])
+            self._model = MusicGen.get_pretrained(self._config["hf_id"], device=device)
-            self._model.to(self._device)
 
             logger.info(
                 f"MusicGen Style {self._variant} loaded successfully "