revert adding multispeaker demo until further investigated

app.py

@@ -6,7 +6,6 @@ import numpy as np
 import torch
 
 from InferenceInterfaces.Meta_FastSpeech2 import Meta_FastSpeech2
-from Preprocessing.ProsodicConditionExtractor import ProsodicConditionExtractor
 
 
 def float2pcm(sig, dtype='int16'):
@@ -37,7 +36,7 @@ class TTS_Interface:
         self.device = "cuda" if torch.cuda.is_available() else "cpu"
         self.model = Meta_FastSpeech2(device=self.device)
 
-    def read(self, prompt, language, audio):
+    def read(self, prompt, language):
         language_id_lookup = {
             "English"  : "en",
             "German"   : "de",
@@ -50,8 +49,6 @@ class TTS_Interface:
             "French"   : "fr"
             }
         self.model.set_language(language_id_lookup[language])
-        if audio is not None:
-            self.model.default_utterance_embedding = ProsodicConditionExtractor(sr=audio[0]).extract_condition_from_reference_wave(audio[1]).to(self.device)
         wav = self.model(prompt)
         return 48000, float2pcm(wav.cpu().numpy())
 
@@ -69,8 +66,7 @@ iface = gr.Interface(fn=meta_model.read,
                                                  'Russian',
                                                  'Hungarian',
                                                  'Dutch',
-                                                 'French'], type="value", default='English', label="Language Selection"),
-                             gr.inputs.Audio(source="microphone", optional=True, label="[optional] Make the TTS imitate your Voice")],
+                                                 'French'], type="value", default='English', label="Language Selection")],
                      outputs=gr.outputs.Audio(type="numpy", label=None),
                      layout="vertical",
                      title="IMS Toucan Multilingual Multispeaker Demo",