File size: 7,461 Bytes
b4d8745
951b505
b4d8745
 
5351689
b4d8745
5e70a25
5351689
1dc3846
 
 
5e70a25
4e7d1a7
 
b4d8745
 
 
 
 
 
4e7d1a7
 
b4d8745
 
 
 
5e70a25
b4d8745
 
 
 
 
4e7d1a7
b4d8745
 
 
 
 
 
5351689
 
b4d8745
 
 
5e70a25
b4d8745
 
 
5e70a25
 
b4d8745
 
5e70a25
 
 
b4d8745
 
 
5e70a25
b4d8745
 
 
1dc3846
b4d8745
5e70a25
b4d8745
 
 
5e70a25
b4d8745
 
 
 
 
 
5e70a25
b4d8745
5e70a25
 
b4d8745
4e7d1a7
 
 
 
 
5e70a25
4e7d1a7
 
 
 
 
 
 
 
 
b4d8745
1dc3846
 
 
 
5e70a25
1dc3846
 
 
 
951b505
b4d8745
 
951b505
5e70a25
951b505
99cdbb0
951b505
 
 
 
 
 
5e70a25
951b505
 
 
 
 
 
 
 
 
 
 
 
 
5e70a25
951b505
 
 
 
 
 
 
5e70a25
951b505
 
 
5e70a25
951b505
 
 
 
 
 
5e70a25
951b505
 
 
 
 
 
5e70a25
951b505
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
b4d8745
5e70a25
b4d8745
951b505
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
83
84
85
86
87
88
89
90
91
92
93
94
95
96
97
98
99
100
101
102
103
104
105
106
107
108
109
110
111
112
113
114
115
116
117
118
119
120
121
122
123
124
125
126
127
128
129
130
131
132
133
134
135
136
137
138
139
140
141
142
143
144
145
146
147
148
149
150
151
152
153
154
155
156
157
158
159
160
161
162
163
164
165
166
167
168
169
170
171
172
173
174
175
176
177
178
179
180
"""
Speech Translation Demo with Automatic TTS, Restart Option, and About Tab

This demo performs the following:
  1. Accepts up to 15 seconds of audio recording from the microphone.
  2. Uses OpenAI’s Whisper model to transcribe the speech.
  3. Splits the transcription into segments and translates each segment on-the-fly using Facebook’s M2M100 model.
  4. Streams the cumulative translation output to the user.
  5. Automatically converts the final translated text to speech using gTTS.
  6. Provides a "Restart Recording" button (located just below the recording section)
     to reset the audio input, translated text, and TTS output.
     
Note: True real-time translation (i.e. while speaking) requires a continuous streaming
solution which is not provided by the standard browser microphone input.
"""

import gradio as gr
import whisper
import torch
from transformers import M2M100ForConditionalGeneration, M2M100Tokenizer
from gtts import gTTS
import uuid

# -----------------------------------------------------------------------------
# Global Model Loading
# -----------------------------------------------------------------------------
whisper_model = whisper.load_model("base")  # Using "base" for a balance between speed and accuracy

tokenizer = M2M100Tokenizer.from_pretrained("facebook/m2m100_418M")
m2m100_model = M2M100ForConditionalGeneration.from_pretrained("facebook/m2m100_418M")

# -----------------------------------------------------------------------------
# Define Supported Languages (including Polish)
# -----------------------------------------------------------------------------
LANGUAGES = {
    "English": "en",
    "Spanish": "es",
    "French": "fr",
    "German": "de",
    "Chinese": "zh",
    "Polish": "pl"
}

# -----------------------------------------------------------------------------
# Main Processing Function: Translation
# -----------------------------------------------------------------------------
def translate_audio(audio, target_language):
    """
    Transcribes the input audio using Whisper and translates the text into the target language.
    Returns the cumulative translated text.
    """
    if audio is None:
        return "No audio provided."
    
    # Transcribe the audio (using fp16=False for CPU compatibility)
    result = whisper_model.transcribe(audio, fp16=False)
    source_lang = result.get("language", "en")
    target_lang_code = LANGUAGES.get(target_language, "en")
    
    cumulative_translation = ""
    for segment in result.get("segments", []):
        segment_text = segment.get("text", "").strip()
        if not segment_text:
            continue
        
        if source_lang == target_lang_code:
            translated_segment = segment_text
        else:
            tokenizer.src_lang = source_lang  # Set source language for proper translation.
            encoded = tokenizer(segment_text, return_tensors="pt")
            generated_tokens = m2m100_model.generate(
                **encoded,
                forced_bos_token_id=tokenizer.get_lang_id(target_lang_code)
            )
            translated_segment = tokenizer.batch_decode(generated_tokens, skip_special_tokens=True)[0]
        
        cumulative_translation += translated_segment + " "
    
    return cumulative_translation.strip()

# -----------------------------------------------------------------------------
# TTS Generation Function
# -----------------------------------------------------------------------------
def generate_tts(text, target_language):
    """
    Converts the given text to speech using gTTS and returns the filename of the generated audio.
    """
    lang_code = LANGUAGES.get(target_language, "en")
    if not text or not text.strip():
        return None
    filename = f"tts_{uuid.uuid4().hex}.mp3"
    tts = gTTS(text=text, lang=lang_code)
    tts.save(filename)
    return filename

# -----------------------------------------------------------------------------
# Restart Function
# -----------------------------------------------------------------------------
def restart_recording():
    """
    Clears the audio input, translated text, and TTS output.
    """
    return None, "", None

# -----------------------------------------------------------------------------
# Gradio Interface Definition with Tabs
# -----------------------------------------------------------------------------
with gr.Blocks() as demo:
    with gr.Tabs():
        # Demo Tab
        with gr.TabItem("Demo"):
            gr.Markdown("# Speech Translation Demo")
            gr.Markdown(
                "Speak into the microphone and your speech will be transcribed and translated "
                "segment-by-segment. (Recording is limited to 15 seconds.)\n\n"
                "**Note:** The translation and speech synthesis occur automatically after recording."
            )
            
            # Row for audio input and language selection.
            with gr.Row():
                audio_input = gr.Audio(
                    sources=["microphone"],
                    type="filepath",
                    label="Record your speech (max 15 seconds)",
                    elem_id="audio_input"
                )
                target_lang_dropdown = gr.Dropdown(
                    choices=list(LANGUAGES.keys()),
                    value="English",
                    label="Select Target Language"
                )
            
            # Restart Recording button placed just below the recording section.
            with gr.Row():
                restart_button = gr.Button("Restart Recording")
            
            # Output components: Translated text and TTS audio.
            output_text = gr.Textbox(label="Translated Text", lines=10)
            tts_audio = gr.Audio(label="Translated Speech", type="filepath")
            
            # When audio is recorded, process translation and then generate TTS.
            audio_input.change(
                fn=translate_audio,
                inputs=[audio_input, target_lang_dropdown],
                outputs=output_text
            ).then(
                fn=generate_tts,
                inputs=[output_text, target_lang_dropdown],
                outputs=tts_audio
            )
            
            # Restart button clears all outputs.
            restart_button.click(
                fn=restart_recording,
                inputs=[],
                outputs=[audio_input, output_text, tts_audio]
            )
        
        # About Tab
        with gr.TabItem("About"):
            gr.Markdown(
                """
**Speech Translation Demo with Automatic TTS and Restart Option**

This demo performs the following:
  1. Accepts up to 15 seconds of audio recording from the microphone.
  2. Uses OpenAI’s Whisper model to transcribe the speech.
  3. Splits the transcription into segments and translates each segment on-the-fly using Facebook’s M2M100 model.
  4. Streams the cumulative translation output to the user.
  5. Automatically converts the final translated text to speech using gTTS.
  6. Provides a "Restart Recording" button (located just below the recording section) to reset the audio input, translated text, and TTS output.

**Note:** True real-time translation (i.e. while speaking) requires a continuous streaming solution which is not provided by the standard browser microphone input.
                """
            )

# Launch the Gradio app.
demo.launch()