fix bug

TFG/MuseTalk.py

@@ -4,7 +4,7 @@
 import os
 import time
 import re
-from huggingface_hub import snapshot_download
+# from huggingface_hub import snapshot_download
 import requests
 import numpy as np
 import cv2

requirements_webui.txt

@@ -1,13 +1,17 @@
 # PyTorch and its dependencies
 # These libraries include PyTorch and its related packages, supporting CUDA 11.8.
---extra-index-url https://download.pytorch.org/whl/cu118
-torch==2.0.1
-torchvision==0.15.2
-torchaudio==2.0.2
+--extra-index-url https://download.pytorch.org/whl/torch_stable.html
+torch
+torchvision
+torchaudio
+# torch==2.4.1+cu118
+# torchvision==0.19.1+cu118
+# torchaudio==2.4.1+cu118
+
 # Installation source for PyTorch: -f https://download.pytorch.org/whl/cu118
 # Example installation command:
 
-# pip install torch==2.0.1 torchvision==0.15.2 torchaudio==2.0.2 --index-url https://download.pytorch.org/whl/cu118
+# pip install torch==2.4.1 torchvision==0.19.1 torchaudio==2.4.1 --index-url https://download.pytorch.org/whl/cu118
 # pip install tb-nightly -i https://mirrors.aliyun.com/pypi/simple
 
 # General tools and libraries
@@ -23,7 +27,7 @@ spaces
 moviepy
 librosa==0.10.2
 ultralytics # for wav2lipv2
-gradio==4.16.0
+gradio==4.*
 scikit_learn==1.4.2
 
 # SadTalker related libraries
@@ -46,6 +50,7 @@ matplotlib==3.7.5
 
 # MuseTalk related libraries
 diffusers==0.27.2
+huggingface_hub==0.25.2
 accelerate==0.28.0
 opencv-python==4.9.0.80
 soundfile==0.12.1
@@ -65,7 +70,7 @@ transformers==4.39.2
 openai
 modelscope
 funasr>=1.0.0
-edge-tts>=6.1.9
+edge-tts>=6.1.18
 openai-whisper
 zhconv
 

webui.py

@@ -209,7 +209,7 @@ def LLM_response(
 
     # 生成VTT文件（如果TTS方法为'Edge-TTS'）
     tts_vtt = 'answer.vtt' if tts_method == 'Edge-TTS' else None
-
+    tts_vtt = None
     return tts_audio, tts_vtt, answer
 
 @calculate_time
@@ -267,7 +267,7 @@ def chat_response(system, message, history):
 
 def modify_system_session(system: str) -> str:
     if system is None or len(system) == 0:
-        system = default_system
+        system = DEFAULT_SYSTEM
     llm.clear_history()
     return system, system, []
 
@@ -312,6 +312,7 @@ def human_response(source_image, history, question_audio, talker_method, voice,
                                             cut_method, question_audio, question, use_mic_voice, 
                                             mode_checkbox_group, sft_dropdown, prompt_text_cv, prompt_wav_upload, prompt_wav_record, seed, speed_factor,tts_method)
     driven_vtt = 'answer.vtt' if tts_method == 'Edge-TTS' else None
+    driven_vtt = None
     if driven_audio is None:
         gr.Warning("音频没有正常生成，请检查TTS是否正确")
         return None
@@ -364,7 +365,7 @@ def MuseTalker_response(source_video, bbox_shift, question_audio, text, voice,
     # MuseTalker 视频生成
     video = musetalker.inference_noprepare(driven_audio, source_video, bbox_shift, batch_size, fps=25)
 
-    return video, driven_vtt if driven_vtt else video
+    return (video, driven_vtt) if driven_vtt else video
 
 GPT_SoVITS_ckpt = "GPT_SoVITS/pretrained_models"
 def load_vits_model(gpt_path, sovits_path, progress=gr.Progress(track_tqdm=True)):