Spaces:
Running
Running
| import logging | |
| import os | |
| import time | |
| import uuid | |
| import gradio as gr | |
| import soundfile as sf | |
| from model import get_pretrained_model, language_to_models | |
| title = "# Next-gen Kaldi: Text-to-speech (TTS)" | |
| description = """ | |
| This space shows how to convert text to speech with Next-gen Kaldi. | |
| It is running on CPU within a docker container provided by Hugging Face. | |
| See more information by visiting the following links: | |
| - <https://github.com/k2-fsa/sherpa-onnx> | |
| If you want to deploy it locally, please see | |
| <https://k2-fsa.github.io/sherpa/> | |
| If you want to use Android APKs, please see | |
| <https://k2-fsa.github.io/sherpa/onnx/tts/apk.html> | |
| If you want to use Android text-to-speech engine APKs, please see | |
| <https://k2-fsa.github.io/sherpa/onnx/tts/apk-engine.html> | |
| If you want to download an all-in-one exe for Windows, please see | |
| <https://github.com/k2-fsa/sherpa-onnx/releases/tag/tts-models> | |
| """ | |
| css = """ | |
| .result {display:flex;flex-direction:column} | |
| .result_item {padding:15px;margin-bottom:8px;border-radius:15px;width:100%} | |
| .result_item_success {background-color:mediumaquamarine;color:white;align-self:start} | |
| .result_item_error {background-color:#ff7070;color:white;align-self:start} | |
| """ | |
| # Simplified examples for Portuguese only | |
| examples = [ | |
| ["Portuguese", "csukuangfj/vits-mms-por", "Eu desejo uma versão simplificada para português.", 0, 1.0], | |
| ] | |
| # Use only Portuguese as a language choice | |
| language_choices = ["Portuguese"] | |
| demo = gr.Blocks(css=css) | |
| with demo: | |
| gr.Markdown(title) | |
| # Use Radio instead of Dropdown for language choice | |
| language_radio = gr.Radio( | |
| label="Language", | |
| choices=language_choices, | |
| value=language_choices[0], | |
| ) | |
| # Initialize model_dropdown with Portuguese models | |
| model_dropdown = gr.Dropdown( | |
| choices=language_to_models["Portuguese"], | |
| label="Select a model", | |
| value=language_to_models["Portuguese"][0], | |
| ) | |
| # No need to update model_dropdown for a single language | |
| with gr.Tabs(): | |
| with gr.TabItem("Please input your text"): | |
| input_text = gr.Textbox( | |
| label="Input text", | |
| info="Your text", | |
| lines=3, | |
| placeholder="Please input your text here", | |
| ) | |
| input_sid = gr.Textbox( | |
| label="Speaker ID", | |
| info="Speaker ID", | |
| lines=1, | |
| max_lines=1, | |
| value="0", | |
| placeholder="Speaker ID. Valid only for mult-speaker model", | |
| ) | |
| input_speed = gr.Slider( | |
| minimum=0.1, | |
| maximum=10, | |
| value=1, | |
| step=0.1, | |
| label="Speed (larger->faster; smaller->slower)", | |
| ) | |
| input_button = gr.Button("Submit") | |
| output_audio = gr.Audio(label="Output") | |
| output_info = gr.HTML(label="Info") | |
| gr.Examples( | |
| examples=examples, | |
| fn=process, | |
| inputs=[ | |
| language_radio, | |
| model_dropdown, | |
| input_text, | |
| input_sid, | |
| input_speed, | |
| ], | |
| outputs=[ | |
| output_audio, | |
| output_info, | |
| ], | |
| ) | |
| input_button.click( | |
| process, | |
| inputs=[ | |
| language_radio, | |
| model_dropdown, | |
| input_text, | |
| input_sid, | |
| input_speed, | |
| ], | |
| outputs=[ | |
| output_audio, | |
| output_info, | |
| ], | |
| ) | |
| gr.Markdown(description) | |
| def download_espeak_ng_data(): | |
| os.system( | |
| """ | |
| cd /tmp | |
| wget -qq https://github.com/k2-fsa/sherpa-onnx/releases/download/tts-models/espeak-ng-data.tar.bz2 | |
| tar xf espeak-ng-data.tar.bz2 | |
| """ | |
| ) | |
| if __name__ == "__main__": | |
| download_espeak_ng_data() | |
| formatter = "%(asctime)s %(levelname)s [%(filename)s:%(lineno)d] %(message)s" | |
| logging.basicConfig(format=formatter, level=logging.INFO) | |
| demo.launch() | |