Fix bug

app.py

@@ -74,10 +74,12 @@ def greet(image, audio):
 title = "Audio-Grounded Contrastive Learning"
 
 description = """<p>
-This is a simple demo of Can CLIP Help Sound Source Localization? (WACV2024), zero-shot visual sound localization<br><br>.
+This is a simple demo of Can CLIP Help Sound Source Localization? (WACV2024), zero-shot visual sound localization.<br><br>
 To use it simply upload an image and corresponding audio to mask (identify in the image), or use one of the examples below and click ‘submit’.
 </p>"""
 
+article = "<p style='text-align: center'><a href='https://arxiv.org/abs/2311.04066'>Can CLIP Help Sound Source Localization?</a> | <a href='https://github.com/swimmiing/ACL-SSL'>Offical Github repo</a></p>"
+
 demo = gr.Interface(
     fn=greet,
     inputs=[gr.Image(type='pil'), gr.Audio()],

requirements.txt

@@ -3,4 +3,4 @@ torchvision==0.14.0
 torchaudio==0.13.0
 transformers==4.25.1
 Pillow
-opencv-python==4.8.1.78
+opencv-python