add lang selection

app.py

@@ -16,6 +16,34 @@ description = "Cross Language Speech to Speech (Indonesian/English to 25 other l
 info = "This application uses [Indonesian Whisperer Medium](https://huggingface.co/cahya/whisper-medium-id) model"
 badge = "https://img.shields.io/badge/Powered%20by-Indonesian%20Whisperer-red"
 
+languages = {
+    'English': 'en',
+    'German': 'de',
+    'Spanish': 'es',
+    'French': 'fr',
+    'Portuguese': 'pt',
+    'Polish': 'pl',
+    'Dutch': 'nl',
+    'Swedish': 'sv',
+    'Italian': 'it',
+    'Finnish': 'fi',
+    'Ukrainian': 'uk',
+    'Greek': 'el',
+    'Czech': 'cs',
+    'Romanian': 'ro',
+    'Danish': 'da',
+    'Hungarian': 'hu',
+    'Croatian': 'hr',
+    'Bulgarian': 'bg',
+    'Lithuanian': 'lt',
+    'Slovak': 'sk',
+    'Latvian': 'lv',
+    'Slovenian': 'sl',
+    'Estonian': 'et',
+    'Maltese': 'mt'
+}
+
+
 device = 0 if torch.cuda.is_available() else "cpu"
 
 pipe = pipeline(
@@ -27,6 +55,7 @@ pipe = pipeline(
 
 pipe.model.config.forced_decoder_ids = pipe.tokenizer.get_decoder_prompt_ids(language=lang, task="transcribe")
 
+
 def transcribe(microphone, file_upload):
     warn_output = ""
     if (microphone is not None) and (file_upload is not None):
@@ -78,6 +107,8 @@ with gr.Blocks() as blocks:
                 choices=LANGUAGES,
                 value=default_lang
             )
+            language = gr.Dropdown([lang for lang in languages.keys()], label="Target Language", value="English")
+            language = languages[language]
             with gr.Row(): # mobile_collapse=False
                 submit = gr.Button("Submit", variant="primary")
             examples = gr.Examples(examples=["data/Jokowi - 2022.mp3", "data/Soekarno - 1963.mp3", "data/JFK.mp3"],
@@ -100,7 +131,7 @@ with gr.Blocks() as blocks:
     # actions
     submit.click(
         tts,
-        [radio, audio_microphone, audio_upload],
+        [language, audio_microphone, audio_upload],
         [text_source, text_target, audio],
     )
     radio.change(lambda lang: CoquiTTS.langs[lang]["sentence"], radio)