Update app.py

app.py

@@ -1,8 +1,9 @@
 import torch
-from transformers import pipeline, AutoModel, AutoProcessor
+import os
 from PIL import Image
+from transformers import pipeline, AutoModel
 
-# ✅ Use pipeline for image-text-to-text
+# ✅ Load the model using pipeline
 pipe = pipeline("image-text-to-text", model="deepseek-ai/deepseek-vl2-small", trust_remote_code=True)
 
 # ✅ Load model directly (alternative approach)
@@ -10,13 +11,20 @@ model = AutoModel.from_pretrained("deepseek-ai/deepseek-vl2-small", trust_remote
 
 # ✅ Function to process image and text
 def predict(image_path, text_prompt):
-    image = Image.open("C:\Users\70088531\Downloads\sample_img2.JPG").convert("RGB")
+    # Ensure correct path format for Windows/Linux
+    image_path = image_path.replace("\\", "/")  
+    image = Image.open(image_path).convert("RGB")
+    
     messages = [{"role": "user", "content": text_prompt}]
     result = pipe(image, messages)
     return result
 
 # ✅ Example usage
 if __name__ == "__main__":
-    test_image = "test.jpg"  # Replace with an actual image file
+    # Replace this with the correct image path
+    test_image = "sample_img2.JPG"  # Ensure this image exists in the same folder
     prompt = "Describe this image."
-    print("Generated Response:", predict(test_image, prompt))
+    
+    # Run prediction
+    output = predict(test_image, prompt)
+    print("Generated Response:", output)