update api module

sgm/inference/api.py

@@ -5,8 +5,8 @@ import os
 from sgm.inference.helpers import (
     do_sample,
     do_img2img,
-    BaseDeviceModelLoader,
-    CudaModelLoader,
+    DeviceModelManager,
+    CudaModelManager,
     Img2ImgDiscretizationWrapper,
     Txt2NoisyDiscretizationWrapper,
 )
@@ -167,7 +167,7 @@ class SamplingPipeline:
         model_path: Optional[str] = None,
         config_path: Optional[str] = None,
         use_fp16: bool = True,
-        model_loader: BaseDeviceModelLoader = CudaModelLoader(device="cuda"),
+        device_manager: DeviceModelManager = CudaModelManager(device="cuda"),
     ) -> None:
         """
         Sampling pipeline for generating images from a model.
@@ -204,17 +204,18 @@ class SamplingPipeline:
             raise ValueError(
                 f"Checkpoint {self.ckpt} not found, check model spec or config_path"
             )
-        self.device = device
-        self.swap_device = swap_device
-        load_device = device if swap_device is None else swap_device
-        self.model = self._load_model(device=load_device, use_fp16=use_fp16)
 
-    def _load_model(self, device="cuda", use_fp16=True):
+        self.model_manager = device_manager
+        self.model = self._load_model(
+            device_manager=self.model_manager, use_fp16=use_fp16
+        )
+
+    def _load_model(self, device_manager: DeviceModelManager, use_fp16=True):
         config = OmegaConf.load(self.config)
         model = load_model_from_config(config, self.ckpt)
         if model is None:
             raise ValueError(f"Model {self.model_id} could not be loaded")
-        model.to(device)
+        device_manager.load(model)
         if use_fp16:
             model.conditioner.half()
             model.model.half()
@@ -256,7 +257,7 @@ class SamplingPipeline:
             force_uc_zero_embeddings=["txt"] if not self.specs.is_legacy else [],
             return_latents=return_latents,
             filter=filter,
-            device=self.device,
+            model_manager=self.model_manager,
         )
 
     def image_to_image(

sgm/inference/helpers.py

@@ -91,10 +91,12 @@ class DeviceModelManager(object):
         The default model loader does not perform any swapping, so the model will
         stay on device.
         """
-        model.to(self.device)
-        yield
-        if self.device != self.swap_device:
-            model.to(self.swap_device)
+        try:
+            model.to(self.device)
+            yield
+        finally:
+            if self.device != self.swap_device:
+                model.to(self.swap_device)
 
 
 class CudaModelManager(DeviceModelManager):