Update app.py

app.py

@@ -33,6 +33,7 @@ class TTS_Interface:
         self.model = Meta_FastSpeech2(device=self.device)
         self.current_speaker = "English Speaker's Voice"
         self.current_language = "English"
+        self.current_accent = "English"
         self.language_id_lookup = {
             "English"   : "en",
             "German"    : "de",
@@ -62,7 +63,9 @@ class TTS_Interface:
             "Italian Speaker's Voice"   : "reference_audios/italian.flac",
             }
 
-    def read(self, prompt, language, speaker):
+    def read(self, prompt, language, accent, speaker):
+        language = language.rstrip(" Text")
+        accent = accent.rstrip(" Accent")
         if len(prompt) > 2000:
             if language == "English":
                 prompt = "Your input was too long. Please try either a shorter text or split it into several parts."
@@ -90,8 +93,11 @@ class TTS_Interface:
                 prompt = "Il tuo input era troppo lungo. Per favore, prova un testo più corto o dividilo in più parti."
 
         if self.current_language != language:
-            self.model.set_language(self.language_id_lookup[language])
+            self.model.set_phonemizer_language(self.language_id_lookup[language])
             self.current_language = language
+        if self.current_accent != accent:
+            self.model.set_accent_language(self.language_id_lookup[language])
+            self.current_accent = accent
         if self.current_speaker != speaker:
             self.model.set_utterance_embedding(self.speaker_path_lookup[speaker])
             self.current_speaker = speaker
@@ -105,19 +111,31 @@ article = "<p style='text-align: left'>This is still a work in progress, models
 iface = gr.Interface(fn=meta_model.read,
                      inputs=[gr.inputs.Textbox(lines=2,
                                                placeholder="write what you want the synthesis to read here... (2000 character maximum)",
-                                               label=" "),
-                             gr.inputs.Dropdown(['English',
-                                                 'German',
-                                                 'Greek',
-                                                 'Spanish',
-                                                 'Finnish',
-                                                 'Russian',
-                                                 'Hungarian',
-                                                 'Dutch',
-                                                 'French',
-                                                 'Polish',
-                                                 'Portuguese',
-                                                 'Italian'], type="value", default='English', label="Language Selection"),
+                                               label="Text input"),
+                             gr.inputs.Dropdown(['English Text',
+                                                 'German Text',
+                                                 'Greek Text',
+                                                 'Spanish Text',
+                                                 'Finnish Text',
+                                                 'Russian Text',
+                                                 'Hungarian Text',
+                                                 'Dutch Text',
+                                                 'French Text',
+                                                 'Polish Text',
+                                                 'Portuguese Text',
+                                                 'Italian Text'], type="value", default='English Text', label="Select the Language of the Text"),
+                             gr.inputs.Dropdown(['English Accent',
+                                                 'German Accent',
+                                                 'Greek Accent',
+                                                 'Spanish Accent',
+                                                 'Finnish Accent',
+                                                 'Russian Accent',
+                                                 'Hungarian Accent',
+                                                 'Dutch Accent',
+                                                 'French Accent',
+                                                 'Polish Accent',
+                                                 'Portuguese Accent',
+                                                 'Italian Accent'], type="value", default='English Accent', label="Select the Accent of the Speaker"),
                              gr.inputs.Dropdown(["English Speaker's Voice",
                                                  "German Speaker's Voice",
                                                  "Greek Speaker's Voice",
@@ -129,7 +147,7 @@ iface = gr.Interface(fn=meta_model.read,
                                                  "French Speaker's Voice",
                                                  "Polish Speaker's Voice",
                                                  "Portuguese Speaker's Voice",
-                                                 "Italian Speaker's Voice"], type="value", default="English Speaker's Voice", label="Speaker Selection")],
+                                                 "Italian Speaker's Voice"], type="value", default="English Speaker's Voice", label="Select the Voice of the Speaker")],
                      outputs=gr.outputs.Audio(type="numpy", label=None),
                      layout="vertical",
                      title="IMS Toucan Multilingual Multispeaker Demo",