import gradio as gr from TTS.api import TTS css = """ #warning {background-color: #FFCCCB !important} .feedback label textarea {height: auto !important; font-size: 22px !important; font-weight: 800 !important; text-align: center !important; color: #801313 !important; padding: 0px !important} #alert {background-color: #fff !important} """ # Init TTS tts = TTS(model_name="tts_models/multilingual/multi-dataset/your_tts", progress_bar=False, gpu=False) zh_tts = TTS(model_name="tts_models/zh-CN/baker/tacotron2-DDC-GST", progress_bar=False, gpu=False) de_tts = TTS(model_name="tts_models/de/thorsten/vits", gpu=False) es_tts = TTS(model_name="tts_models/es/mai/tacotron2-DDC", progress_bar=False, gpu=False) def text_to_speech(text: str, speaker_wav, speaker_wav_file): if len(text) > 0: return change_aud(text, speaker_wav, speaker_wav_file) else: return (None) def change_aud(text: str, speaker_wav, speaker_wav_file): if speaker_wav_file and not speaker_wav: speaker_wav = speaker_wav_file file_path = "output.wav" if speaker_wav is not None: tts.tts_to_file(text, speaker_wav=speaker_wav, language="en", file_path=file_path) else: tts.tts_to_file(text, speaker=tts.speakers[0], language="en", file_path=file_path) return file_path def show_error(text): if text == "": return gr.update(visible=True, elem_id="warning", elem_classes="feedback"), gr.update(visible=False) else: return gr.update(visible=False), gr.update(visible=True) # def download_file(): # return file_path title = "Voice-Cloning-Demo" def toggle(choice): if choice == "mic": return gr.update(visible=True, value=None), gr.update(visible=False, value=None) else: return gr.update(visible=False, value=None), gr.update(visible=True, value=None) def change_color(text_input): if len(text_input) == 0: return gr.update(elem_id="warning", autofocus=True) else: return gr.update(elem_id="alert", autofocus=False) def clear_color(text_input, radio,error_box): return gr.update(elem_id="alert"), gr.update(value="mic"), gr.update(visible=False) with gr.Blocks(css="footer {visibility: hidden}") as demo: with gr.Row(): with gr.Column(): text_input = gr.Textbox(label="Input the text", value="", max_lines=4, lines=4) radio = gr.Radio(["mic", "file"], value="mic", label="How would you like to upload your audio?") audio_input_mic = gr.Audio(label="Voice to clone", sources="microphone", type="filepath", visible=True) audio_input_file = gr.Audio(label="Voice to clone", type="filepath", visible=False) with gr.Row(): with gr.Column(): btn_clear = gr.ClearButton([text_input, radio, audio_input_file]) with gr.Column(): btn = gr.Button("Generate", variant="primary") with gr.Column(): audio_output = gr.Audio(label="Output", visible=True, autoplay=True, show_share_button=False) # download_button = gr.DownloadButton(label="Download Audio", value=None, visible=True) error_box = gr.Textbox(label="WARNING", value="Input box cannot be blank!!", visible=False, container=True) # download_button.click(download_file, outputs=download_button) btn_clear.add(audio_output) btn.click(text_to_speech, inputs=[text_input, audio_input_mic, audio_input_file], outputs=audio_output) btn.click(show_error, text_input, [error_box, audio_output]) radio.change(toggle, radio, [audio_input_mic, audio_input_file]) btn_clear.click(clear_color, [text_input, radio, error_box], [text_input, radio, error_box]) btn.click(change_color, text_input, text_input) demo.launch()