Update app.py

app.py

@@ -61,14 +61,29 @@ def align_embeddings(prompt_embeds, negative_prompt_embeds):
     return torch.nn.functional.pad(prompt_embeds, (0, 0, 0, max_length - prompt_embeds.shape[1])), \
            torch.nn.functional.pad(negative_prompt_embeds, (0, 0, 0, max_length - negative_prompt_embeds.shape[1]))
 
-def preprocess_image(image, target_width, target_height): # Преобразует изображение в формат, подходящий для модели.
+# def preprocess_image(image, target_width, target_height): # Преобразует изображение в формат, подходящий для модели.
+#     if isinstance(image, np.ndarray):
+#         image = Image.fromarray(image)
+#     image = image.resize((target_width, target_height), Image.LANCZOS)
+#     image = np.array(image).astype(np.float32) / 255.0  # Нормализация [0, 1]
+#     image = image[None].transpose(0, 3, 1, 2)  # Преобразуем в (batch, channels, height, width)
+#     image = torch.from_numpy(image).to(device)
+#     return image
+
+def preprocess_image(image, target_width, target_height, resize_to_224=False):
     if isinstance(image, np.ndarray):
         image = Image.fromarray(image)
-    image = image.resize((target_width, target_height), Image.LANCZOS)
+    
+    # Если resize_to_224=True, изменяем размер до 224x224
+    if resize_to_224:
+        image = image.resize((224, 224), Image.LANCZOS)
+    else:
+        image = image.resize((target_width, target_height), Image.LANCZOS)
+    
     image = np.array(image).astype(np.float32) / 255.0  # Нормализация [0, 1]
     image = image[None].transpose(0, 3, 1, 2)  # Преобразуем в (batch, channels, height, width)
     image = torch.from_numpy(image).to(device)
-    return image
+    return image    
 
 pipe_default = get_lora_sd_pipeline(lora_dir='./lora_man_animestyle', base_model_name_or_path=model_default, dtype=torch_dtype).to(device)
 # pipe_controlnet = StableDiffusionControlNetPipeline.from_pretrained(
@@ -115,8 +130,11 @@ def infer(
         pipe_ip_adapter.set_ip_adapter_scale(ip_adapter_strength)
 
         # Преобразование изображений
-        ip_source_image = preprocess_image(ip_source_image, width, height)
-        ip_adapter_image = preprocess_image(ip_adapter_image, width, height)
+        #ip_source_image = preprocess_image(ip_source_image, width, height)
+        #ip_adapter_image = preprocess_image(ip_adapter_image, width, height)
+        # Преобразование изображений для IP-Adapter (размер 224x224)
+        ip_source_image = preprocess_image(ip_source_image, width, height, resize_to_224=True)
+        ip_adapter_image = preprocess_image(ip_adapter_image, width, height, resize_to_224=True)
 
         # Получение эмбеддингов изображения для IP-Adapter
         #image_embeds = pipe_ip_adapter.get_image_embeds(ip_adapter_image)