Update src/streamlit_app.py

src/streamlit_app.py

@@ -109,7 +109,7 @@ if uploaded_file is not None:
                 # The BLIP pipeline expects a PIL Image object directly
                 caption_results = captioner(image)
                 # Extract the generated text from the pipeline's output [7]
-                generated_caption = caption_results['generated_text']
+                generated_caption = caption_results[0]['generated_text']
                 
                 # Optional: Enhance prompt for soundscape generation
                 # This helps guide the audio model towards environmental sounds