Hugging Face's logo

Spaces:
Tonic
/
SeamlessOnDevice
Sleeping

App Files Community
SeamlessOnDevice
File size: 4,169 Bytes 
e667211
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
61
62
63
64
65
66
67
68
69
70
71
72
73
74
75
76
77
78
79
80
81
82
 
import gradio as gr
import torchaudio
import torch
import os
import time
import soundfile as sf

welcome_message = """
# Welcome to Tonic's Unity On Device!

Tonic's Unity On Device uses [facebook/seamless-m4t-unity-small](https://huggingface.co/facebook/seamless-m4t-unity-small) for audio translation & accessibility.
Tonic's Unity On Device!🚀 on your own data & in your own way by cloning this space. Simply click here: <a style="display:inline-block" href="https://huggingface.co/spaces/TeamTonic/SeamlessOnDevice?duplicate=true"><img src="https://img.shields.io/badge/-Duplicate%20Space-blue?labelColor=white&style=flat&logo=data:image/png;base64,iVBORw0KGgoAAAANSUhEUgAAABAAAAAQCAYAAAAf8/9hAAAAAXNSR0IArs4c6QAAAP5JREFUOE+lk7FqAkEURY+ltunEgFXS2sZGIbXfEPdLlnxJyDdYB62sbbUKpLbVNhyYFzbrrA74YJlh9r079973psed0cvUD4A+4HoCjsA85X0Dfn/RBLBgBDxnQPfAEJgBY+A9gALA4tcbamSzS4xq4FOQAJgCDwV2CPKV8tZAJcAjMMkUe1vX+U+SMhfAJEHasQIWmXNN3abzDwHUrgcRGmYcgKe0bxrblHEB4E/pndMazNpSZGcsZdBlYJcEL9Afo75molJyM2FxmPgmgPqlWNLGfwZGG6UiyEvLzHYDmoPkDDiNm9JR9uboiONcBXrpY1qmgs21x1QwyZcpvxt9NS09PlsPAAAAAElFTkSuQmCC&logoWidth=14" alt="Duplicate Space"></a></h3>
### Join us : 
TeamTonic is always making cool demos! Join our active builder's community on Discord: [Discord](https://discord.gg/GWpVpekp) On 🤗Huggingface: [TeamTonic](https://huggingface.co/TeamTonic) & [MultiTransformer](https://huggingface.co/MultiTransformer) On 🌐Github: [Polytonic](https://github.com/tonic-ai) & contribute to 🌟 [PolyGPT](https://github.com/tonic-ai/polygpt-alpha)"             
"""

# Define the list of target languages
languages = {
    "English": "eng",
    "Hindi": "hin",
    "Portuguese": "por",
    "Russian": "rus",
    "Spanish": "spa"
}

def save_audio(audio_input, output_dir="saved_audio"):
    if not os.path.exists(output_dir):
        os.makedirs(output_dir)

    # Extract sample rate and audio data
    sample_rate, audio_data = audio_input

    # Generate a unique file name
    file_name = f"audio_{int(time.time())}.wav"
    file_path = os.path.join(output_dir, file_name)

    # Save the audio file
    sf.write(file_path, audio_data, sample_rate)

    return file_path

def speech_to_text(audio_data, tgt_lang):
    file_path = save_audio(audio_data)
    audio_input, _ = torchaudio.load(file_path)
    s2t_model = torch.jit.load("unity_on_device_s2t.ptl")
    with torch.no_grad():
        text = s2t_model(audio_input, tgt_lang=languages[tgt_lang])
    return text

def speech_to_speech_translation(audio_data, tgt_lang):
    file_path = save_audio(audio_data)
    audio_input, _ = torchaudio.load(file_path)
    s2st_model = torch.jit.load("unity_on_device.ptl")
    with torch.no_grad():
        text, units, waveform = s2st_model(audio_input, tgt_lang=languages[tgt_lang])
    output_file = "/tmp/result.wav"
    torchaudio.save(output_file, waveform.unsqueeze(0), sample_rate=16000)
    return text, output_file

def create_interface():
    with gr.Markdown(welcome_message)
    with gr.Blocks(theme='ParityError/Anime') as interface:
        # Dropdown for language selection
        input_language = gr.Dropdown(list(languages.keys()), label="Select Target Language", value="English")

        with gr.Accordion("Speech to Text", open=False) as stt_accordion:
            audio_input_stt = gr.Audio(label="Upload or Record Audio")
            text_output_stt = gr.Text(label="Transcribed Text")
            stt_button = gr.Button("Transcribe")
            stt_button.click(speech_to_text, inputs=[audio_input_stt, input_language], outputs=text_output_stt)

        with gr.Accordion("Speech to Speech Translation", open=False) as s2st_accordion:
            audio_input_s2st = gr.Audio(label="Upload or Record Audio")
            text_output_s2st = gr.Text(label="Translated Text")
            audio_output_s2st = gr.Audio(label="Translated Audio", type="filepath")
            s2st_button = gr.Button("Translate")
            s2st_button.click(speech_to_speech_translation, inputs=[audio_input_s2st, input_language], outputs=[text_output_s2st, audio_output_s2st])

    return interface

app = create_interface()
app.launch(show_error=True, debug=True)