Handle .to for 4bit quantized models (#186)

inseq/models/huggingface_model.py

@@ -155,13 +155,18 @@ def load(
     def device(self, new_device: str) -> None:
         check_device(new_device)
         self._device = new_device
+        is_loaded_in_8bit = getattr(self.model, "is_loaded_in_8bit", False)
+        is_loaded_in_4bit = getattr(self.model, "is_loaded_in_4bit", False)
+        is_quantized = is_loaded_in_8bit or is_loaded_in_4bit
+
         # Enable compatibility with 8bit models
         if self.model:
-            if not (hasattr(self.model, "is_loaded_in_8bit") and self.model.is_loaded_in_8bit):
+            if not is_quantized:
                 self.model.to(self._device)
             else:
+                mode = "8bit" if is_loaded_in_8bit else "4bit"
                 logger.warning(
-                    "The model is loaded in 8bit mode. The device cannot be changed after loading the model."
+                    f"The model is loaded in {mode} mode. The device cannot be changed after loading the model."
                 )
 
     @abstractmethod