File size: 2,387 Bytes
5b74a4b fcc244c 83e3ccb 5b74a4b fcc244c 425531b fcc244c 5b74a4b fcc244c 83e3ccb fcc244c 83e3ccb fcc244c a48f8e0 fcc244c a48f8e0 5b74a4b 72632b9 fcc244c c58bd88 fcc244c 17cfe18 83e3ccb a5ec736 b2c7d3a 5b74a4b 8fe6fd5 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 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60 61 62 63 64 |
import gradio as gr
import requests
import soundfile as sf
import numpy as np
import tempfile
# Define the Hugging Face Inference API URLs and headers
ASR_API_URL = "https://api-inference.huggingface.co/models/Baghdad99/saad-speech-recognition-hausa-audio-to-text"
TTS_API_URL = "https://api-inference.huggingface.co/models/Baghdad99/english_voice_tts"
TRANSLATION_API_URL = "https://api-inference.huggingface.co/models/Baghdad99/saad-hausa-text-to-english-text"
headers = {"Authorization": "Bearer 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"}
# Define the function to query the Hugging Face Inference API
def query(api_url, payload):
response = requests.post(api_url, headers=headers, json=payload)
return response.json()
# Define the function to translate speech
def translate_speech(audio):
# audio is a tuple (np.ndarray, int), we need to save it as a file
audio_data, sample_rate = audio
with tempfile.NamedTemporaryFile(suffix=".wav", delete=False) as f:
sf.write(f, audio_data, sample_rate)
audio_file = f.name
# Use the ASR pipeline to transcribe the audio
with open(audio_file, "rb") as f:
data = f.read()
response = requests.post(ASR_API_URL, headers=headers, data=data)
output = response.json()
# Check if the output contains 'text'
if 'text' in output:
transcription = output["text"]
else:
print("The output does not contain 'text'")
return
# Use the translation pipeline to translate the transcription
translated_text = query(TRANSLATION_API_URL, {"inputs": transcription})
# Use the TTS pipeline to synthesize the translated text
response = requests.post(TTS_API_URL, headers=headers, json={"inputs": translated_text})
audio_bytes = response.content
# Convert the audio bytes to numpy array
with tempfile.NamedTemporaryFile(suffix=".wav", delete=False) as f:
f.write(audio_bytes)
audio_file = f.name
audio_data, _ = sf.read(audio_file)
return audio_data
# 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()
|