Update handler.py

handler.py

@@ -13,6 +13,8 @@ from insightface.app import FaceAnalysis
 from style_template import styles
 from pipeline_stable_diffusion_xl_instantid_full import StableDiffusionXLInstantIDPipeline, draw_kps
 from controlnet_aux import OpenposeDetector
+import torch.nn.functional as F
+from torchvision.transforms import Compose
 import os
 from huggingface_hub import hf_hub_download
 import base64
@@ -56,13 +58,14 @@ class EndpointHandler:
         controlnet_model = hf_hub_download(repo_id="InstantX/InstantID", filename="ControlNetModel/diffusion_pytorch_model.safetensors", local_dir=os.path.join(model_dir, "checkpoints"))
         face_adapter = hf_hub_download(repo_id="InstantX/InstantID", filename="ip-adapter.bin", local_dir=os.path.join(model_dir, "checkpoints"))
 
-        # Load face detection model
-        self.face_net = cv2.dnn.readNet(os.path.join(model_dir, "models", "face_detection_yunet_2023mar_int8.onnx"))
-
-        # Initialize InsightFace for face analysis with a model available for commercial use
-        self.face_analyzer = FaceAnalysis(name='model', root=model_dir, providers=['CUDAExecutionProvider', 'CPUExecutionProvider'])
-        self.face_analyzer.prepare(ctx_id=0, det_size=(640, 640))
+        dir_path = os.path.join(model_dir, "models", "face_detection_yunet_2023mar_int8.onnx")
+        if not os.path.exists(dir_path):
+            raise RuntimeError(f"Model path {dir_path} does not exist.")
+        else:
+            self.face_net = cv2.dnn.readNet(dir_path)
 
+        self.app = FaceAnalysis(name='model', root=model_dir, providers=['CUDAExecutionProvider', 'CPUExecutionProvider'])
+        self.app.prepare(ctx_id=0, det_size=(640, 640))
         openpose = OpenposeDetector.from_pretrained("lllyasviel/ControlNet")
 
         # Path to InstantID models
@@ -259,7 +262,7 @@ class EndpointHandler:
         face_kps = draw_kps(face_image, np.array([[x, y], [x1, y1]]))  # Placeholder keypoints
 
         # Analyze the face using InsightFace
-        face_info = self.face_analyzer.get(face_image_cv2)
+        face_info = self.app.get(face_image_cv2)
 
         if not face_info:
             return {"error": "Face analysis failed."}