- Firstly clone the project
git clone [email protected]:AInelo/hackathon_langage.gitor
git clone https://github.com/AInelo/hackathon_langage.git- Enter in ./Code/ repository Run in the project Racine
npm install- After run
npm start- We just deploy the Speech voice to Fon texte model of transformers IA and we've runned it on some link and PORT Download all dependency "transformer", "flask", "numpy", "PyTorch"
pip install <Dependency>OUR CODE fort API is :
from flask import Flask, request, jsonify
from transformers import pipeline
import requests
import os
import librosa
app = Flask(__name__)
pipe = pipeline("automatic-speech-recognition", model="chrisjay/fonxlsr")
@app.route('/transcribe', methods=['POST'])
def transcribe_audio():
audio_link = request.json.get('audio_link')
if not audio_link:
return jsonify({'error': 'No audio link provided'}), 400
# Check if the audio_link is a URL or a local file path
if audio_link.startswith('http://') or audio_link.startswith('https://'):
# Handle URL case
try:
audio_response = requests.get(audio_link)
audio_response.raise_for_status()
temp_audio_path = 'temp_audio.webm'
with open(temp_audio_path, 'wb') as f:
f.write(audio_response.content)
except requests.exceptions.RequestException as e:
return jsonify({'error': 'Failed to download audio file: ' + str(e)}), 400
else:
# Handle local file path case
if not os.path.exists(audio_link):
return jsonify({'error': 'File does not exist: ' + audio_link}), 400
temp_audio_path = audio_link
# Load the audio file with librosa to get a NumPy array
try:
audio_data, _ = librosa.load(temp_audio_path, sr=16000)
except Exception as e:
return jsonify({'error': 'Failed to load audio file: ' + str(e)}), 400
# Perform transcription on the audio data
try:
transcription = pipe(audio_data)
return jsonify({'transcription': transcription['text']}), 200
except Exception as e:
return jsonify({'error': 'Transcription failed: ' + str(e)}), 500
finally:
if audio_link.startswith('http://') or audio_link.startswith('https://'):
if os.path.exists(temp_audio_path):
os.remove(temp_audio_path)
if __name__ == '__main__':
# app.run(debug=True)
app.run(debug=True, host='XXX.XXX.XXX.XXX', port=5000)
- Last STEP of building creat a ".env" file and add
TRANSFORMER_URL_API_SPEECHTOTEXT = http://XXX.XXX.XXX.XXX:5000/transcribe- Thanks you go and visit this.
http://XXX.XXX.XXX.XXX:PORT/
http://XXX.XXX.XXX.XXX:PORT/voiceartisans.html
http://XXX.XXX.XXX.XXX:PORT/imagenerate.html
http://XXX.XXX.XXX.XXX:PORT/bracelet.html