Update app.py

app.py

@@ -51,19 +51,7 @@ def inference(text, voice):
     torchaudio.save('generated.wav', gen.squeeze(0).cpu(), 24000)
     return "generated.wav"
     
-def inference_own_voice(text, voice_1, voice_2, voice_3):
-    text = text[:256]
-    print(voice_1, voice_2, voice_3)
-    conds = [
-        load_audio(voice_1, 22050),
-        load_audio(voice_2, 22050),
-        load_audio(voice_3, 22050),
-    ]
-    print(text, conds, preset)
-    gen = tts.tts_with_preset(text, conds, preset)
-    print("gen")
-    torchaudio.save('generated.wav', gen.squeeze(0).cpu(), 24000)
-    return "generated.wav"
+
  
 text = "Joining two modalities results in a surprising increase in generalization! What would happen if we combined them all?"
 examples = [
@@ -77,28 +65,21 @@ with block:
     gr.Markdown("# TorToiSe")
     gr.Markdown("A multi-voice TTS system trained with an emphasis on quality")
     with gr.Tabs():
-        with gr.TabItem("Pre-recorded voices"):
-            iface = gr.Interface(
-                inference,
-                inputs=[
-                    gr.inputs.Textbox(type="str", default=text, label="Text", lines=3),
-                    gr.inputs.Dropdown(voices),
-                ],
-                outputs="audio",
-                examples=examples,
-            )
-        with gr.TabItem("Record your voice"):
-            iface = gr.Interface(
-              inference_own_voice,
-              inputs=[
-                  gr.inputs.Textbox(type="str", default=text, label="Text", lines=3),
-                  gr.inputs.Audio(source="microphone", label="Record yourself reading something out loud (audio 1)", type="filepath"),
-                  gr.inputs.Audio(source="microphone", label="Record yourself reading something out loud (audio 2)", type="filepath"),
-                  gr.inputs.Audio(source="microphone", label="Record yourself reading something out loud (audio 3)", type="filepath"),
-              ],
-              outputs="audio"
-            )
+      with gr.TabItem("Pre-recorded voices"):
+        iface = gr.Interface(
+          inference,
+          inputs=[
+              gr.inputs.Textbox(type="str", default=text, label="Text", lines=3),
+              gr.inputs.Dropdown(voices),
+          ],
+          outputs="audio",
+          enable_queue=True,
+          examples=examples,
+        )
 
     gr.Markdown("This demo shows the ultra fast option in the TorToiSe system. For more info check the <a href='https://github.com/neonbjb/tortoise-tts' target='_blank'>Repository</a>.",)
 
-    block.launch()
+block.launch()
+
+
+iface.launch(cache_examples=True)