concatenate videos

app.py

@@ -45,9 +45,9 @@ def process_images(diary_image, writer_image):
     # Generate the video based on the summaries
     video_path = generate_video(scene_list, writer_summary, fps=24)
 
-    captions = scenes_caption(scene_list, openai_api_key)
+    caption = scenes_caption(scene_list, openai_api_key)
 
-    return video_path, captions
+    return video_path, caption
 
 
 # Define the Gradio interface
@@ -65,20 +65,22 @@ with gr.Blocks() as interface:
     gr.Markdown("# Handwritten Diary to Video")
 
     with gr.Row():
-        diary_image_input = gr.Image(label="Upload your handwritten diary image", type="pil")
-        writer_image_input = gr.Image(label="Upload a photo of the writer", type="pil")
+        # Left column for user inputs
+        with gr.Column():
+            diary_image_input = gr.Image(label="Upload your handwritten diary image", type="pil")
+            writer_image_input = gr.Image(label="Upload a photo of the writer", type="pil")
+            submit_button = gr.Button("Generate Video")
 
-    submit_button = gr.Button("Generate Video")
-
-    with gr.Row():
-        video_output = gr.Video(label="Generated Video")
-        captions_output = gr.Markdown(label="Scene Captions")
+        # Right column for generated video and caption
+        with gr.Column():
+            video_output = gr.Video(label="Generated Video")
+            caption_output = gr.Markdown(label="Scene Caption")
 
     # Bind the submit button click to trigger the video generation and display
     submit_button.click(
         fn=gradio_interface,
         inputs=[diary_image_input, writer_image_input],
-        outputs=[video_output, captions_output]
+        outputs=[video_output, caption_output]
     )
 
 # Launch the interface

baseline_utils.py

@@ -126,7 +126,7 @@ def scenes_caption(scenes, api_key):
         # Append the generated caption to the list
         captions.append(response.choices[0].message.content)
 
-    return captions
+    return "\n\n".join(captions)
 
 
 @spaces.GPU