Remove unused prediction_type and upcast_attention from from_single_f…

…ile(...) calls.

invokeai/backend/model_manager/load/model_loaders/stable_diffusion.py

@@ -117,8 +117,6 @@ def _load_from_singlefile(
             load_class = load_classes[config.base][config.variant]
         except KeyError as e:
             raise Exception(f"No diffusers pipeline known for base={config.base}, variant={config.variant}") from e
-        prediction_type = config.prediction_type.value
-        upcast_attention = config.upcast_attention
 
         # Without SilenceWarnings we get log messages like this:
         # site-packages/huggingface_hub/file_download.py:1132: FutureWarning: `resume_download` is deprecated and will be removed in version 1.0.0. Downloads always resume when possible. If you want to force a new download, use `force_download=True`.
@@ -129,12 +127,7 @@ def _load_from_singlefile(
         # ['text_model.embeddings.position_ids']
 
         with SilenceWarnings():
-            pipeline = load_class.from_single_file(
-                config.path,
-                torch_dtype=self._torch_dtype,
-                prediction_type=prediction_type,
-                upcast_attention=upcast_attention,
-            )
+            pipeline = load_class.from_single_file(config.path, torch_dtype=self._torch_dtype)
 
         if not submodel_type:
             return pipeline