Update handler.py

handler.py

@@ -14,21 +14,25 @@ if device.type != 'cuda':
 
 class EndpointHandler():
     def __init__(self, path=""):
+
+        self.pipe = AutoPipelineForInpainting.from_pretrained("diffusers/stable-diffusion-xl-1.0-inpainting-0.1", torch_dtype=torch.float16, variant="fp16").to("cuda")
+        self.generator = torch.Generator(device="cuda").manual_seed(0)
+        
         # load StableDiffusionInpaintPipeline pipeline
-        self.pipe = AutoPipelineForInpainting.from_pretrained(
+        #self.pipe = AutoPipelineForInpainting.from_pretrained(
             "runwayml/stable-diffusion-inpainting",
             revision="fp16",
             torch_dtype=torch.float16,
         )
         # use DPMSolverMultistepScheduler
-        self.pipe.scheduler = DPMSolverMultistepScheduler.from_config(self.pipe.scheduler.config)
+        #self.pipe.scheduler = DPMSolverMultistepScheduler.from_config(self.pipe.scheduler.config)
         # move to device        
-        self.pipe = self.pipe.to(device)
+        #self.pipe = self.pipe.to(device)
 
-        self.pipe2 = AutoPipelineForInpainting.from_pretrained("stabilityai/stable-diffusion-xl-refiner-1.0", torch_dtype=torch.float16, variant="fp16", use_safetensors=True)
-        self.pipe2.to("cuda")
+        #self.pipe2 = AutoPipelineForInpainting.from_pretrained("stabilityai/stable-diffusion-xl-refiner-1.0", torch_dtype=torch.float16, variant="fp16", use_safetensors=True)
+        #self.pipe2.to("cuda")
 
-        self.pipe3 = AutoPipelineForImage2Image.from_pipe(self.pipe2)
+        #self.pipe3 = AutoPipelineForImage2Image.from_pipe(self.pipe2)
         
 
 
@@ -52,6 +56,22 @@ class EndpointHandler():
             image = None
             mask_image = None 
 
+
+        image = self.pipe(
+          prompt=prompt,
+          image=image,
+          mask_image=mask_image,
+          guidance_scale=8.0,
+          num_inference_steps=20,  # steps between 15 and 30 work well for us
+          strength=0.99,  # make sure to use `strength` below 1.0
+          generator=generator,
+        ).images[0]
+
+        return image
+
+        """
+        pipe = AutoPipelineForInpainting.from_pretrained("diffusers/stable-diffusion-xl-1.0-inpainting-0.1", torch_dtype=torch.float16, variant="fp16").to("cuda")
+
         self.pipe.enable_xformers_memory_efficient_attention()
         
         # run inference pipeline
@@ -94,6 +114,7 @@ class EndpointHandler():
             
         # return first generate PIL image
         return result
+        """
     
     # helper to decode input image
     def decode_base64_image(self, image_string):