| import gradio as gr | |
| from gradio_client import Client, handle_file | |
| client = Client("Ghana-NLP/Southern-Ghana-ASR") | |
| demo = gr.Blocks() | |
| title = "# Khaya AI: Speech Recognition for Languages of Southern Ghana" | |
| description = """ | |
| <b>How to use:</b> <br> | |
| SPECIFY INPUT LANGUAGE & Speak through mic (or upload a file of same).<br> | |
| MAXIMUM 30 seconds please.<br> | |
| This model does not execute capitalization and punctuation.<br> | |
| """ | |
| def transcribe_wrapper_fn(audio, LANG): | |
| result = client.predict( | |
| audio=handle_file(audio), | |
| LANG=LANG, | |
| api_name="/predict" | |
| ) | |
| return result | |
| mic_transcribe = gr.Interface( | |
| fn=transcribe_wrapper_fn, | |
| inputs=[gr.Audio(sources=["microphone"], type="filepath"), gr.Dropdown(["Asante Twi","Adangme","Ewe","Fante","French","Ga","Nzema","African English"], value="Asante Twi", multiselect=False, label="Language", info="Select Lamguage")], | |
| outputs=gr.Textbox(), | |
| ) | |
| file_transcribe = gr.Interface( | |
| fn=transcribe_wrapper_fn, | |
| inputs=[gr.Audio(sources=["upload"], type="filepath"), gr.Dropdown(["Asante Twi","Adangme","Ewe","Fante","French","Ga","Nzema","African English"], value="Asante Twi", multiselect=False, label="Language", info="Select Lamguage")], | |
| outputs=gr.Textbox(), | |
| ) | |
| with demo: | |
| gr.Markdown(title) | |
| gr.Markdown(description) | |
| gr.TabbedInterface( | |
| [mic_transcribe, file_transcribe], | |
| ["Transcribe Microphone", "Transcribe Audio File"], | |
| ) | |
| if __name__ == "__main__": | |
| demo.queue(max_size=20).launch() | |