Rework UNet loader

coreml_suite/loaders.py

@@ -1,3 +1,5 @@
+import os.path
+
 from coremltools import ComputeUnit
 from python_coreml_stable_diffusion.coreml_model import CoreMLModel
 
@@ -9,6 +11,7 @@ from comfy.model_patcher import ModelPatcher
 from coreml_suite.logger import logger
 from coreml_suite.model import CoreMLModelWrapper
 
+
 class CoreMLLoader:
     PACKAGE_DIRNAME = ""
 
@@ -26,21 +29,22 @@ class CoreMLLoader:
             }
         }
 
-    RETURN_TYPES = ("COREML_MODEL",)
     FUNCTION = "load"
     CATEGORY = "CoreML Suite"
 
     @classmethod
     def coreml_filenames(cls):
+        extensions = [".mlmodelc", ".mlpackage"]
+        all_paths = folder_paths.get_filename_list_(cls.PACKAGE_DIRNAME)[1]
+        coreml_paths = folder_paths.filter_files_extensions(all_paths, extensions)
+
         return {
-            p.split('/')[-1]:
-                p for p in
-            folder_paths.get_filename_list_(cls.PACKAGE_DIRNAME)[1]
-            if p.endswith((".mlpackage", ".mlmodelc"))
+            os.path.split(p)[-1]: p
+            for p in coreml_paths
         }
 
     def load(self, coreml_name, compute_unit):
-        logger.info(f"Loading {coreml_name}")
+        logger.info(f"Loading {coreml_name} to {compute_unit}")
 
         coreml_path = self.coreml_filenames()[coreml_name]
 
@@ -73,25 +77,25 @@ class CoreMLLoaderTextEncoder(CoreMLLoader):
 
 class CoreMLLoaderUNet(CoreMLLoader):
     PACKAGE_DIRNAME = "unet"
-    RETURN_TYPES = ("MODEL",)
+    RETURN_TYPES = ("coreml_model",)
 
-    def _load(self, coreml_path, compute_unit, sources):
-        # TODO: This is a dummy model config, but it should be enough to
-        #  get the model to load - implement a proper model config
-        model_config = supported_models_base.BASE({})
-        model_config.latent_format = SD15()
-        model_config.unet_config = {
-            "disable_unet_model_creation": True,
-            "num_res_blocks": 2,
-            "attention_resolutions": [1, 2, 4],
-            "channel_mult": [1, 2, 4, 4],
-            "transformer_depth": [1, 1, 1, 0],
-        }
-        coreml_model = CoreMLModelWrapper(model_config, coreml_path,
-                                          compute_unit, sources)
-
-        return (ModelPatcher(coreml_model, model_management.get_torch_device(),
-                             None),)
+    # def _load(self, coreml_path, compute_unit, sources):
+    #     # TODO: This is a dummy model config, but it should be enough to
+    #     #  get the model to load - implement a proper model config
+    #     model_config = supported_models_base.BASE({})
+    #     model_config.latent_format = SD15()
+    #     model_config.unet_config = {
+    #         "disable_unet_model_creation": True,
+    #         "num_res_blocks": 2,
+    #         "attention_resolutions": [1, 2, 4],
+    #         "channel_mult": [1, 2, 4, 4],
+    #         "transformer_depth": [1, 1, 1, 0],
+    #     }
+    #     coreml_model = CoreMLModelWrapper(model_config, coreml_path,
+    #                                       compute_unit, sources)
+    #
+    #     return (ModelPatcher(coreml_model, model_management.get_torch_device(),
+    #                          None),)
 
 
 class CoreMLLoaderVAE(CoreMLLoader):