add cpu offload / xformers

app.py

@@ -45,15 +45,19 @@ class Instance:
     def load_model(self, model_id='j-min/IterInpaint-CLEVR'):
         pipe = StableDiffusionInpaintPipeline.from_pretrained(model_id)
 
-        if torch.cuda.is_available():
-            pipe = pipe.to('cuda')
-
         def dummy(images, **kwargs):
             return images, False
         pipe.safety_checker = dummy
         print("Disabled safety checker")
 
         print("Loaded model")
+
+        # This command loads the individual model components on GPU on-demand. So, we don't
+        # need to explicitly call pipe.to("cuda").
+        pipe.enable_model_cpu_offload()
+
+        # xformers
+        pipe.enable_xformers_memory_efficient_attention()
         return pipe
 
 instance = Instance()