handler.py

handler.py

@@ -0,0 +1,36 @@
+
+from typing import  Dict, List, Any
+import torch
+from torch import autocast
+from diffusers import StableDiffusionPipeline
+import base64
+from io import BytesIO
+# set device
+device = torch.device('cuda' if torch.cuda.is_available() else 'cpu')
+
+if device.type != 'cuda':
+    raise ValueError("need to run on GPU")
+
+class EndpointHandler():
+    def __init__(self, path=""):
+        # load the optimized model
+        self.pipe = StableDiffusionPipeline.from_pretrained(path, torch_dtype=torch.float16)
+        self.pipe = self.pipe.to(device)
+
+
+    def __call__(self, data: Any) -> "PIL.Image":
+        """
+        Args:
+            data (:obj:):
+                includes the input data and the parameters for the inference.
+        Return:
+            A :obj:`dict`:. base64 encoded image
+        """
+        inputs = data.pop("inputs", data)
+
+        # run inference pipeline
+        with autocast(device.type):
+            image = self.pipe(inputs, guidance_scale=7.5)["sample"][0]
+
+        # encoding image as base 64 is done by the default toolkit
+        return image

requirements.txt

@@ -0,0 +1,3 @@
+torch
+diffusers
+typing