Spaces:
Running
Running
File size: 1,721 Bytes
79616b0 ee42cb0 f96ce38 ee42cb0 f96ce38 ee42cb0 e544543 9d49358 ee42cb0 79616b0 ee42cb0 e544543 ee42cb0 |
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 |
from fastapi import FastAPI, File, UploadFile
import librosa
import numpy as np
import shutil
import uvicorn
import os
from funasr import AutoModel
# Khởi tạo FastAPI
app = FastAPI()
# Tạo thư mục temp nếu chưa có
if not os.path.exists("temp"):
os.makedirs("temp")
# Load mô hình SenseVoiceSmall từ Hugging Face
model_dir = "FunAudioLLM/SenseVoiceSmall"
model = AutoModel(
model=model_dir,
vad_model="fsmn-vad",
vad_kwargs={"max_single_segment_time": 30000},
device="cuda:0",
hub="hf",
)
# Hàm tính RMS energy
def calculate_rms_energy(audio_path):
y, sr = librosa.load(audio_path)
rms = librosa.feature.rms(y=y)[0]
return np.mean(rms)
# Hàm phát hiện tiếng ồn
def detect_noise(audio_path):
rms_energy = calculate_rms_energy(audio_path)
res = model.generate(input=audio_path, language="auto", audio_event_detection=True)
audio_events = res[0].get("audio_event_detection", {})
if rms_energy > 0.02:
return "ồn ào"
elif rms_energy > 0.01:
for event_label, event_score in audio_events.items():
if event_score > 0.7 and event_label in ["laughter", "applause", "crying", "coughing"]:
return f"ồn ào ({event_label})"
return "yên tĩnh"
# API nhận file âm thanh từ Flutter
@app.post("/detect-noise/")
async def detect_noise_api(file: UploadFile = File(...)):
file_path = f"temp/{file.filename}"
with open(file_path, "wb") as buffer:
shutil.copyfileobj(file.file, buffer)
result = detect_noise(file_path)
return {"noise_level": result}
# Chạy FastAPI trên Hugging Face Spaces
if __name__ == "__main__":
uvicorn.run(app, host="0.0.0.0", port=7860)
|