Update inference_from_video.py

inference_from_video.py

@@ -68,6 +68,10 @@ def parse_args():
         "--sample_rate", type=int, default=16000,
         help="Sample rate for audio output.",
     )
+    parser.add_argument(
+        "--max_duration", type=int, default=10,
+        help="Maximum length duration for generated audio."
+    )
     parser.add_argument(
         "--save_dir", type=str, default="./outputs/tmp",
         help="output save dir"
@@ -140,7 +144,7 @@ def main():
     sample_rate = args.sample_rate
 
      # Define max_len_in_seconds globally for consistency
-    max_len_in_seconds = 8
+    max_len_in_seconds = args.max_duration
 
     def audio_text_matching(waveforms, text, sample_freq=24000, max_len_in_seconds=max_len_in_seconds):
         new_freq = 48000