Update app.py

app.py

@@ -3,8 +3,6 @@ import os
 import subprocess
 
 # Define the paths where the input and output files will be stored
-INPUT_AUDIO_PATH = "input_audio.wav"
-INPUT_VIDEO_PATH = "input_video.mp4"
 OUTPUT_VIDEO_PATH = "output_video.mp4"
 MODEL_PATH = "checkpoints/checkpoint.pt"
 
@@ -23,11 +21,7 @@ def get_sample_flags(sample_mode):
         return None
 
 # Function to run the model inference command
-def generate_video(audio_file, video_file):
-    # Save uploaded files to disk
-    audio_file.save(INPUT_AUDIO_PATH)
-    video_file.save(INPUT_VIDEO_PATH)
-
+def generate_video(audio_path, video_path):
     sample_input_flags = get_sample_flags(SAMPLE_MODE)
     if not sample_input_flags:
         return "Error: sample_mode can only be 'cross' or 'reconstruction'"
@@ -53,7 +47,7 @@ def generate_video(audio_file, video_file):
     )
     GEN_FLAGS = (
         f"--generate_from_filelist {GENERATE_FROM_FILELIST} "
-        f"--video_path={INPUT_VIDEO_PATH} --audio_path={INPUT_AUDIO_PATH} "
+        f"--video_path={video_path} --audio_path={audio_path} "
         f"--out_path={OUTPUT_VIDEO_PATH} --save_orig=False "
         f"--face_det_batch_size 16 --pads {PADS} --is_voxceleb2=False"
     )
@@ -77,8 +71,8 @@ with gr.Blocks() as demo:
     gr.Markdown("## Audio-Video Synthesis Model")
     
     with gr.Row():
-        audio_input = gr.Audio(label="Upload Audio", type="file")
-        video_input = gr.Video(label="Upload Video", type="file")
+        audio_input = gr.Audio(label="Upload Audio", type="filepath")
+        video_input = gr.Video(label="Upload Video", type="filepath")
 
     output_video = gr.Video(label="Generated Video")