gradio

F5-TTS/src/f5_tts/infer/infer_cli_test.py

@@ -14,6 +14,7 @@ from cached_path import cached_path
 from omegaconf import OmegaConf
 
 from moviepy.editor import VideoFileClip, AudioFileClip
+import traceback
 
 
 from f5_tts.infer.utils_infer import (
@@ -495,6 +496,7 @@ if __name__ == "__main__":
                 wav_gen = torch.from_numpy(wav_gen).unsqueeze(0)
                 assert(sr_gen == 24000)
             except:
+                traceback.print_exc()
                 print("error generation", i+args.start, txt_p, txt)
                 wav_gen = torch.zeros(1, 24000)
                 sr_gen = 24000

app.py

@@ -71,10 +71,11 @@ def video_to_audio_and_speech(video: gr.Video, prompt: str, text: str, audio_pro
     print("v2a command", command)
     os.system(command)
     
+    video_gen = video_save_path[:-4]+".mp4.gen.mp4"
+    
     command = "python ./F5-TTS/src/f5_tts/infer/infer_cli_test.py --output_dir %s --start 0 --end 1 --ckpt_file ./F5-TTS/ckpts/v2c/v2c_s44.pt --v2a_path %s --wav_p %s --txt_p \"%s\" --video %s --v2a_wav %s --txt \"%s\"" % (output_dir, output_dir, audio_p_path, text_prompt, video_save_path, video_save_path[:-4]+".flac", text)
-    print("v2s command", command)
+    print("v2s command", command, video_gen)
     os.system(command)
-    video_gen = output_dir + "/videos/gen/0001"
     
     return video_save_path, video_gen