Update app.py

app.py

@@ -1,4 +1,4 @@
-import streamlit as st
+import gradio as gr
 import edge_tts
 import asyncio
 import librosa
@@ -24,24 +24,25 @@ def tts_get_voices_list():
     
     return voices
 
+def tts_mode(txt, voice):
+    tts = asyncio.run(edge_tts.Communicate(txt, voice).save('temp\\test.mp3'))
+    audio, sr = librosa.load('temp\\test.mp3', sr=16000, mono=True)
+    raw_path = io.BytesIO()
+    soundfile.write(raw_path, audio, 16000, format="wav")
+    raw_path.seek(0)
+    model = Svc(fr"Herta-Svc/G_10000.pth", f"Herta-Svc/config.json", device = 'cpu')
+    out_audio, out_sr = model.infer('speaker0', 0, raw_path, auto_predict_f0 = True,)
+    soundfile.write('temp\\xxx.wav', out_audio.cpu().numpy(), 44100)
+    audio_file = open('temp\\xxx.wav', 'rb')
+    audio_bytes = audio_file.read()
+
+    return audio_bytes
+
 loop = asyncio.new_event_loop()
 asyncio.set_event_loop(loop)
 
-with st.form(key = 'tts', clear_on_submit=False):
-    txt = st.text_input('your text message (your text message should be < 100 characters)', max_chars = 100)
-    voice = str(st.selectbox('voices', tts_get_voices_list()))
-    summitted = st.form_submit_button('Summit')
-
-    if summitted:
-        tts = asyncio.run(edge_tts.Communicate(txt, voice).save('temp\\test.mp3'))
-        audio, sr = librosa.load('temp\\test.mp3', sr=16000, mono=True)
-        raw_path = io.BytesIO()
-        soundfile.write(raw_path, audio, 16000, format="wav")
-        raw_path.seek(0)
-        model = Svc(fr"Herta-Svc/G_10000.pth", f"Herta-Svc/config.json", device = 'cpu')
-        out_audio, out_sr = model.infer('speaker0', 0, raw_path, auto_predict_f0 = True,)
-        soundfile.write('temp\\xxx.wav', out_audio.cpu().numpy(), 44100)
-        audio_file = open('temp\\xxx.wav', 'rb')
-        audio_bytes = audio_file.read()
-        st.audio(audio_bytes, format = 'audio/wav')
+txt = gr.Textbox(label = 'TTS text')
+voice = gr.Dropdown(tts_get_voices_list(), value = str)
 
+if gr.Button():
+    tts_mode(txt, voice)