Update app.py

app.py

@@ -4,65 +4,18 @@ import os
 from TTS.api import TTS
 import time
 
-tts = TTS(model_name="voice_conversion_models/multilingual/vctk/freevc24", progress_bar=False, gpu=False)
 logging.basicConfig(level=logging.INFO)
-count = 0
-
-def main():
-    global count
-    count += 1
-    if count > 50:
-        time.sleep(5)
-        os.system("rm -R /tmp/*")
-        count = 0
-        
-    with gr.Blocks() as demo:
-
-        with gr.Row():
-            with gr.Column(variant="panel"):
-                src_audio_mic = gr.Audio(source="microphone", label="Record your voice")
-                src_audio_file = gr.Audio(
-                    source="upload", type="filepath", label="Or upload audio to convert"
-                )
 
-            with gr.Column(variant="panel"):
-                tgt_audio_file = gr.Audio(
-                    source="upload", type="filepath", label="Select audio with target voice"
-                )
-
-        with gr.Row():
-            convert_btn = gr.Button("Convert")
-        with gr.Row():
-            result_audio = gr.Audio()
-
-        
-        def voice_conversion(src_from_mic_, src_from_file_, tgt_from_file_):
-            """
-            helper function which checks where source come from
-            """
-            src_ = None
-            if src_from_mic_:
-                src_ = src_from_mic_
-            elif src_from_file_:
-                src_ = src_from_file_
-            tgt_ = tgt_from_file_
-            if not src_ or not tgt_:
-                logging.warning("source or target are not provided")
-                return
+# TTS with on the fly voice conversion
+api = TTS("tts_models/deu/fairseq/vits")
+count = 0
 
-            print(src_)
-            print(tgt_)
-            tts.voice_conversion_to_file(source_wav=src_, target_wav=tgt_, file_path="output.wav")
-            return "output.wav"
-        
-        convert_btn.click(
-            voice_conversion,
-            inputs=[src_audio_mic, src_audio_file, tgt_audio_file],
-            outputs=result_audio,
-        )
+def audio_tts(txt, audio_file):
+    return {'cat': 0.3, 'dog': 0.7}
 
-    demo.queue(concurrency_count=1).launch(show_api=False, show_error=True)
+demo = gr.Interface(fn=image_classifier, inputs=[gr.Textbox(label="Input text TTS", value="Привет! Я Макс."),
+                                                gr.Audio(source="upload", type="filepath", label="Input audio")],
+                    outputs=gr.Audio(source="upload", type="filepath", label="Output audio"))
 
+demo.queue(concurrency_count=1).launch(show_error=True)
 
-if __name__ == "__main__":
-    main()