Update app.py

app.py

@@ -7,9 +7,9 @@ model_id = "Salesforce/blip-image-captioning-base"
 model = BlipForConditionalGeneration.from_pretrained(model_id)
 processor = BlipProcessor.from_pretrained(model_id)
 
-def generate_caption(image_path):
-    # Load the image directly from the path
-    image = Image.open(image_path).convert('RGB')
+def generate_caption(image_array):
+    # Convert numpy array to PIL Image
+    image = Image.fromarray(image_array.astype('uint8')).convert('RGB')
     
     # Process the image to generate tensor inputs
     inputs = processor(image, return_tensors="pt")
@@ -20,6 +20,7 @@ def generate_caption(image_path):
     # Decode and return the generated caption
     return processor.decode(out[0], skip_special_tokens=True)
 
+
 # Gradio interface setup to accept image input and produce text output
 iface = gr.Interface(generate_caption, inputs="image", outputs="text")