File size: 1,665 Bytes
5b74a4b
 
5c4fa2e
5b74a4b
 
 
 
 
 
 
 
 
 
 
 
425531b
 
 
5b74a4b
425531b
5b74a4b
 
 
 
 
 
 
e3a6dbd
 
df27a26
5b74a4b
df27a26
17cfe18
a5ec736
b2c7d3a
5b74a4b
 
eaff29b
5b74a4b
 
 
 
 
b2c7d3a
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
import gradio as gr
from transformers import pipeline
import numpy as np

# Load the pipeline for speech recognition and translation
pipe = pipeline(
    "automatic-speech-recognition",
    model="Baghdad99/saad-speech-recognition-hausa-audio-to-text",
    tokenizer="Baghdad99/saad-speech-recognition-hausa-audio-to-text"
)
translator = pipeline("text2text-generation", model="Baghdad99/saad-hausa-text-to-english-text")
tts = pipeline("text-to-speech", model="Baghdad99/english_voice_tts")

# Define the function to translate speech
def translate_speech(audio):
    # Separate the sample rate and the audio data
    sample_rate, audio_data = audio

    # Use the speech recognition pipeline to transcribe the audio
    transcription = pipe(audio_data)["transcription"]

    # Use the translation pipeline to translate the transcription
    translated_text = translator(transcription, return_tensors="pt", padding=True)

    # Use the text-to-speech pipeline to synthesize the translated text
    synthesised_speech = tts(translated_text, return_tensors='pt')

    # Define the max_range variable
    max_range = 32767  # You can adjust this value based on your requirements
    synthesised_speech = (synthesised_speech.numpy() * max_range).astype(np.int16)

    return 16000, synthesised_speech


# Define the Gradio interface
iface = gr.Interface(
    fn=translate_speech, 
    inputs=gr.inputs.Audio(source="microphone", type="numpy"), 
    outputs=gr.outputs.Audio(type="numpy"),
    title="Hausa to English Translation",
    description="Realtime demo for Hausa to English translation using speech recognition and text-to-speech synthesis."
)

iface.launch()