Update app.py

app.py

@@ -120,7 +120,7 @@ def extrac(text):
     final_list = [x for x in final_list if x != '']
     print(final_list)
     return final_list
-def infer(language,text,speaker_id, n_scale= 0.667,n_scale_w = 0.8, l_scale = 1 ):
+def infer(text ,language, speaker_id,n_scale= 0.667,n_scale_w = 0.8, l_scale = 1):
     speaker_id = int(selection(speaker_id))
     a = ['【','[','(','（']
     b = ['】',']',')','）']
@@ -152,7 +152,9 @@ def infer(language,text,speaker_id, n_scale= 0.667,n_scale_w = 0.8, l_scale = 1
             print(time_end)
             f1.write(str(c-1)+'\n'+time_start+' --> '+time_end+'\n'+sentence+'\n\n')
             audio_fin.append(audio)
-    return (hps_ms.data.sampling_rate, np.concatenate(audio_fin))
+    with open("subtitles.srt", 'r', encoding='utf-8') as file:
+        subtitles = file.read()
+    return (hps_ms.data.sampling_rate, np.concatenate(audio_fin)),subtitles
 lan = ["中文","日文","自动"]
 idols = ["高咲侑","歩夢","かすみ","しずく","果林","愛","せつ菜","璃奈","栞子","エマ","ランジュ","ミア","派蒙"]
 hps_ms = utils.get_hparams_from_file("lovelive/config.json")
@@ -164,20 +166,18 @@ net_g_ms = SynthesizerTrn(
     **hps_ms.model).to(dev)
 _ = net_g_ms.eval()
 _ = utils.load_checkpoint("lovelive/G_936000.pth", net_g_ms)
-app = gr.Blocks()
-with app:
-    with gr.Tabs():
-
-        with gr.TabItem("虹团vits模型，现可按句分割实现长文本合成，可自行用export_to_onnx.py导出"):
-
-            tts_input1 = gr.TextArea(label="如需实现快速合成，建议在colab上克隆后运行本仓库", value="为什么你会那么熟练啊？你和雪菜亲过多少次了？我想做只属于你一个人的学院偶像，所以，请只注视我一个人，好吗？【中文】\nなんでそんなに慣れてんだよっ？せつ菜と…何回キスしたんだよ？どこまであたしを置いてきぼりにすれば気が済むんだよ？[日文]\nI can't choose just one（English）")
-            language = gr.Dropdown(label="选择语言，目前勉强可以做到自动识别",choices=lan, value="自动", interactive=True)
-            para_input1 = gr.Slider(minimum= 0,maximum=1.0,label="更改噪声比例，以控制情感", value=0.667)
-            para_input2 = gr.Slider(minimum= 0,maximum=1.0,label="更改噪声偏差，以控制音素长短", value=0.7)
-            para_input3 = gr.Slider(minimum= 0.1,maximum=10,label="更改时间比例", value=1)
-            tts_submit = gr.Button("Generate", variant="primary")
-            speaker1 = gr.Dropdown(label="选择说话人",choices=idols, value="歩夢", interactive=True)
-            tts_output2 = gr.Audio(label="字幕见 https://huggingface.co/spaces/Mahiruoshi/Lovelive_Nijigasaki_VITS/subtitles.srt")
-            tts_submit.click(infer, [language,tts_input1,speaker1,para_input1,para_input2,para_input3], [tts_output2])
-    #app.launch(share=True)
-    app.launch()
+inputs = [gr.TextArea(label="如需实现快速合成，建议在colab上克隆后运行本仓库", value="为什么你会那么熟练啊？你和雪菜亲过多少次了？我想做只属于你一个人的学院偶像，所以，请只注视我一个人，好吗？【中文】\nなんでそんなに慣れてんだよっ？せつ菜と…何回キスしたんだよ？どこまであたしを置いてきぼりにすれば気が済むんだよ？[日文]\nI can't choose just one（English）"),
+           gr.Dropdown(label="选择语言，目前勉强可以做到自动识别",choices=lan, value="自动", interactive=True),
+           gr.Dropdown(label="选择说话人",choices=idols, value="歩夢", interactive=True),
+           gr.Slider(minimum= 0,maximum=1.0,label="更改噪声比例，以控制情感", value=0.267),
+           gr.Slider(minimum= 0,maximum=1.0,label="更改噪声偏差，以控制音素长短", value=0.7),
+           gr.Slider(minimum= 0.1,maximum=10,label="更改时间比例", value=1)]
+outputs=[gr.Audio(label="采样率22050"), gr.outputs.File(label="字幕文件：subtitles.srt")]
+iface = gr.Interface(
+    fn=infer,
+    inputs=inputs,
+    outputs=outputs,
+    title="Vits",
+    description="虹团11人模型",
+)
+iface.launch()