no message

webui.py

@@ -72,7 +72,7 @@ def generate_audio(tts_text, mode_checkbox_group, sft_dropdown, prompt_text, pro
         prompt_wav = None
     # if instruct mode, please make sure that model is iic/CosyVoice-300M-Instruct and not cross_lingual mode
     if mode_checkbox_group in ['自然语言控制']:
-        if cosyvoice.frontend.instruct is False:
+        if get_cosyvoice().frontend.instruct is False:
             gr.Warning('您正在使用自然语言控制模式, {}模型不支持此模式, 请使用iic/CosyVoice-300M-Instruct模型'.format(args.model_dir))
             yield (target_sr, default_data)
         if instruct_text == '':
@@ -82,7 +82,7 @@ def generate_audio(tts_text, mode_checkbox_group, sft_dropdown, prompt_text, pro
             gr.Info('您正在使用自然语言控制模式, prompt音频/prompt文本会被忽略')
     # if cross_lingual mode, please make sure that model is iic/CosyVoice-300M and tts_text prompt_text are different language
     if mode_checkbox_group in ['跨语种复刻']:
-        if cosyvoice.frontend.instruct is True:
+        if get_cosyvoice().frontend.instruct is True:
             gr.Warning('您正在使用跨语种复刻模式, {}模型不支持此模式, 请使用iic/CosyVoice-300M模型'.format(args.model_dir))
             yield (target_sr, default_data)
         if instruct_text != '':
@@ -114,24 +114,24 @@ def generate_audio(tts_text, mode_checkbox_group, sft_dropdown, prompt_text, pro
     if mode_checkbox_group == '预训练音色':
         logging.info('get sft inference request')
         set_all_random_seed(seed)
-        for i in cosyvoice.inference_sft(tts_text, sft_dropdown, stream=stream, speed=speed):
+        for i in get_cosyvoice().inference_sft(tts_text, sft_dropdown, stream=stream, speed=speed):
             yield (target_sr, i['tts_speech'].numpy().flatten())
     elif mode_checkbox_group == '3s极速复刻':
         logging.info('get zero_shot inference request')
         prompt_speech_16k = postprocess(load_wav(prompt_wav, prompt_sr))
         set_all_random_seed(seed)
-        for i in cosyvoice.inference_zero_shot(tts_text, prompt_text, prompt_speech_16k, stream=stream, speed=speed):
+        for i in get_cosyvoice().inference_zero_shot(tts_text, prompt_text, prompt_speech_16k, stream=stream, speed=speed):
             yield (target_sr, i['tts_speech'].numpy().flatten())
     elif mode_checkbox_group == '跨语种复刻':
         logging.info('get cross_lingual inference request')
         prompt_speech_16k = postprocess(load_wav(prompt_wav, prompt_sr))
         set_all_random_seed(seed)
-        for i in cosyvoice.inference_cross_lingual(tts_text, prompt_speech_16k, stream=stream, speed=speed):
+        for i in get_cosyvoice().inference_cross_lingual(tts_text, prompt_speech_16k, stream=stream, speed=speed):
             yield (target_sr, i['tts_speech'].numpy().flatten())
     else:
         logging.info('get instruct inference request')
         set_all_random_seed(seed)
-        for i in cosyvoice.inference_instruct(tts_text, sft_dropdown, instruct_text, stream=stream, speed=speed):
+        for i in get_cosyvoice().inference_instruct(tts_text, sft_dropdown, instruct_text, stream=stream, speed=speed):
             yield (target_sr, i['tts_speech'].numpy().flatten())
 
 # SDK模型下载
@@ -157,9 +157,10 @@ parser.add_argument('--model_dir',
 args = parser.parse_args()
 
 cosyvoice_instance = None
-
+model_dir=args.model_dir
 @spaces.GPU
-def create_cosyvoice(model_dir):
+def get_cosyvoice():
+    global cosyvoice_instance, model_dir
     # 在这里加入你需要的处理逻辑
     if cosyvoice_instance is not None:
         return cosyvoice_instance
@@ -168,9 +169,8 @@ def create_cosyvoice(model_dir):
 
 @spaces.GPU
 def load_sft_options():
-    return cosyvoice.list_avaliable_spks()
-# 将原来的初始化修改为函数调用
-cosyvoice = create_cosyvoice(args.model_dir)
+    return get_cosyvoice().list_avaliable_spks()
+
 
 prompt_sr, target_sr = 16000, 22050
 default_data = np.zeros(target_sr)