Upload 2 files

N.zip

@@ -1,3 +1,3 @@
 version https://git-lfs.github.com/spec/v1
-oid sha256:b56e4467e1353d6475a17e6b4d67631103cd68df61108ddbe98ebb4962fbc1ec
-size 447896729
+oid sha256:c2ffdb4f06600470610cbaaf7e3aa9b75c32005707f42c545504b6a1d1588cfa
+size 298535552

app.py

@@ -49,9 +49,8 @@ def tts_fn(_text, _gender, _lang, _rate, _volume, sid, vc_transform, auto_f0,clu
         soundfile.write(input_audio, audio, 44100, format="wav")
 
         output_file_path = "tts_output.mp3"
-        print (_text, _gender, _lang, _rate, _volume, sid, vc_transform, auto_f0,cluster_ratio, slice_db,  f0_predictor)
         _audio = model.slice_inference(input_audio, sid, vc_transform, slice_db, cluster_ratio, auto_f0, 0.4,f0_predictor=f0_predictor,clip_seconds=40)
-        
+        print (_text, _gender, _lang, _rate, _volume, sid, vc_transform, auto_f0,cluster_ratio, slice_db,  f0_predictor)
         soundfile.write("tts_output.mp3", _audio, 44100, format="mp3")
         return "Success", output_file_path
 
@@ -136,7 +135,7 @@ def loadmodel(model_):
     spks = list(model.spk2id.keys())
     print(model_, "configs/" + model_name + ".json", "./kmeans/" + model_name + ".pt")
     
-    return "success",update_dropdown(spks)
+    return update_dropdown(spks)
 
 def update_dropdown(new_choices):
     global model
@@ -170,7 +169,7 @@ with app:
                 #f'<img style="width:auto;height:300px;" src="cover.png">'
                 #'</div>')
 
-            gr.Markdown(value=base64.b64decode( b'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').decode())
+            gr.Markdown(value=base64.b64decode( b'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').decode())
 
             with gr.Tabs():
                 with gr.TabItem("单个音频上传"):
@@ -188,20 +187,26 @@ with app:
                         tts_rate = gr.Slider(label = "TTS语音变速（倍速相对值）", minimum = -1, maximum = 3, value = 0, step = 0.1)
                         tts_volume = gr.Slider(label = "TTS语音音量（相对值）", minimum = -1, maximum = 1.5, value = 0, step = 0.1)
                     vc_tts_submit = gr.Button("文本转语音", variant="primary")
+                with gr.TabItem("使用方法说明："):
+                
+                    gr.Markdown(value=base64.b64decode( b'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').decode())
+
+            
             spks = list(model.spk2id.keys())
   
-            sid = gr.Dropdown(label="音色", choices=spks, value="HOSHINO_MINAMI")
+            sid = gr.Dropdown(label="音色(目前有58个)", choices=spks, value="HOSHINO_MINAMI")
             sid.change(fn=update_dropdown,inputs=[sid],outputs=[sid])
             sid.update(interactive=True)
-            with gr.Accordion(label="↓切换模型(默认58位成员v1,音色具有抽奖性质，可切换尝试。也有特化的个人模型可选择)", open=False):
+            with gr.Accordion(label="↓切换模型(默认58v1,音色具有抽奖性质,可尝试切换。也有特化的个人模型可选择)", open=False):
+                gr.Markdown(value="特殊说明:IKEDA_TERESA个人模型无聚类模型。44位成员的模型是9月份制作的，音质最好。")
                 modelstrs = gr.Dropdown(label = "模型", choices = modelPaths, value = modelPaths[0], type = "value")
                 btnMod = gr.Button("载入模型")
-                statusa = gr.TextArea()
-                btnMod.click(loadmodel, inputs=[modelstrs], outputs = [statusa,sid])
+                
+                btnMod.click(loadmodel, inputs=[modelstrs], outputs = [sid])
             with gr.Row():
                  slice_db = gr.Slider(label="切片阈值(较嘈杂时-30，保留呼吸声时-50)",maximum=-30, minimum=-70, step=1, value=-40)                              
                  vc_transform = gr.Slider(label="变调（整数，可以正负，半音数量，升高八度就是12）",maximum=16, minimum=-16, step=1, value=0)                       
-            f0_predictor = gr.Radio(label="f0预测器(如遇哑音可以尝试更换f0)凭干声干净程度选择。推荐fcpe和rmvpe", choices=["pm","dio","harvest","fcpe","rmvpe"], value="fcpe")
+            f0_predictor = gr.Radio(label="f0预测器(如遇哑音可以尝试更换f0)凭干声干净程度选择。只推荐fcpe音色最像或rmvpe音最准", choices=["pm","dio","harvest","fcpe","rmvpe"], value="fcpe")
             with gr.Row():    
                 cluster_ratio = gr.Slider(label="聚类模型混合比例，0-1之间，默认为0不启用聚类，能提升音色相似度，但会导致咬字下降（如果使用建议0.5左右）",maximum=1, minimum=0, step=0.1, value=0)   
                 output_format = gr.Radio(label="音频输出格式(MP3会导致时间轴多27ms,需合成请选flac)", choices=["flac", "mp3"], value = "mp3")#格式