app.py

import os
import gradio as gr
import numpy as np
from fairseq.checkpoint_utils import load_model_ensemble_and_task_from_hf_hub
from fairseq.models.speech_to_speech.hub_interface import S2SHubInterface
from fairseq.models.speech_to_text.hub_interface import S2THubInterface
from audio_pipe import SpeechToSpeechPipeline

io1 = gr.Interface.load("huggingface/facebook/xm_transformer_s2ut_en-hk", api_key=os.environ['api_key'])
io2 = gr.Interface.load("huggingface/facebook/xm_transformer_s2ut_hk-en", api_key=os.environ['api_key'])
io3 = gr.Interface.load("huggingface/facebook/xm_transformer_unity_en-hk", api_key=os.environ['api_key'])
io4 = gr.Interface.load("huggingface/facebook/xm_transformer_unity_hk-en", api_key=os.environ['api_key'])
pipe = SpeechToSpeechPipeline("facebook/xm_transformer_unity_hk-en")

def call_model(audio, model):
    # pipe = SpeechToSpeechPipeline("facebook/xm_transformer_unity_hk-en")
    # wav, sr, text = pipe(audio)
    temp_file = pipe(audio)
    return gr.Audio(temp_file)

def inference(audio, model):
    if model == "xm_transformer_s2ut_en-hk":
        out_audio = io1(audio)
    elif model == "xm_transformer_s2ut_hk-en":
        out_audio = io2(audio)
    elif model == "xm_transformer_unity_en-hk":
        out_audio = io3(audio)
    elif model == "xm_transformer_unity_hk-en_gpu":
        out_audio = call_model(audio, model)
    else:
        out_audio = io4(audio)
    return out_audio 

    
css = """
        .gradio-container {
            font-family: 'IBM Plex Sans', sans-serif;
        }
        .gr-button {
            color: black;
            border-color: grey;
            background: white;
        }
        input[type='range'] {
            accent-color: black;
        }
        .dark input[type='range'] {
            accent-color: #dfdfdf;
        }
        .container {
            max-width: 730px;
            margin: auto;
            padding-top: 1.5rem;
        }
     
        .details:hover {
            text-decoration: underline;
        }
        .gr-button {
            white-space: nowrap;
        }
        .gr-button:focus {
            border-color: rgb(147 197 253 / var(--tw-border-opacity));
            outline: none;
            box-shadow: var(--tw-ring-offset-shadow), var(--tw-ring-shadow), var(--tw-shadow, 0 0 #0000);
            --tw-border-opacity: 1;
            --tw-ring-offset-shadow: var(--tw-ring-inset) 0 0 0 var(--tw-ring-offset-width) var(--tw-ring-offset-color);
            --tw-ring-shadow: var(--tw-ring-inset) 0 0 0 calc(3px var(--tw-ring-offset-width)) var(--tw-ring-color);
            --tw-ring-color: rgb(191 219 254 / var(--tw-ring-opacity));
            --tw-ring-opacity: .5;
        }
        .footer {
            margin-bottom: 45px;
            margin-top: 35px;
            text-align: center;
            border-bottom: 1px solid #e5e5e5;
        }
        .footer>p {
            font-size: .8rem;
            display: inline-block;
            padding: 0 10px;
            transform: translateY(10px);
            background: white;
        }
        .dark .footer {
            border-color: #303030;
        }
        .dark .footer>p {
            background: #0b0f19;
        }
        .prompt h4{
            margin: 1.25em 0 .25em 0;
            font-weight: bold;
            font-size: 115%;
        }
        .animate-spin {
            animation: spin 1s linear infinite;
        }
        @keyframes spin {
            from {
                transform: rotate(0deg);
            }
            to {
                transform: rotate(360deg);
            }
        }
"""

block = gr.Blocks(css=css)



with block:
    gr.HTML(
        """
            <div style="text-align: center; max-width: 700px; margin: 0 auto;">
              <div
                style="
                  display: inline-flex;
                  align-items: center;
                  gap: 0.8rem;
                  font-size: 1.75rem;
                "
              >
                <h1 style="font-weight: 900; margin-bottom: 7px;">
                  Hokkien Translation
                </h1>
              </div>
              <p style="margin-bottom: 10px; font-size: 94%">
                A demo for fairseq speech-to-speech translation models. It supports S2UT and UnitY models for bidirectional Hokkien and English translation. Please select the model and record the input to submit.
              </p>
            </div>
        """
    )
    with gr.Group():
        with gr.Box():
            with gr.Row().style(mobile_collapse=False, equal_height=True):
                audio = gr.Audio(
                   source="microphone", type="filepath", label="Input"
                )
                
                btn = gr.Button("Submit")
        model = gr.Dropdown(choices=["xm_transformer_unity_en-hk", "xm_transformer_unity_hk-en", "xm_transformer_unity_hk-en_gpu", "xm_transformer_s2ut_en-hk", "xm_transformer_s2ut_hk-en"], value="xm_transformer_unity_en-hk",type="value", label="Model")
        out = gr.Audio(label="Output")
        
        btn.click(inference, inputs=[audio, model], outputs=out) 
        gr.HTML('''
        <div class="footer">
                    <p>Model by <a href="https://ai.facebook.com/" style="text-decoration: underline;" target="_blank">Meta AI</a>
                    </p>
        </div>
        ''')

block.launch()