Faster-Whisper 本地实时语音识别部署实战

前言

要实现类似微信或豆包的语音输入功能，通常有云端 API 和本地模型两种方案。云端准确度高但涉及隐私和费用，本地模型则免费且数据不出域。Faster-Whisper 作为 Whisper 的高效实现，非常适合本地实时语音识别场景。

环境准备

在虚拟环境中安装依赖。核心库是 faster-whisper，录音部分推荐使用 pyaudiowpatch（Windows 下音频兼容性更好）或标准 pyaudio。

pip install faster-whisper
pip install pyaudiowpatch

模型下载

支持多种模型尺寸，从 Tiny 到 Large-v3。如果服务器无法联网，可手动从 Hugging Face 下载模型文件（config.json, model.bin, tokenizer.json 等），放入指定文件夹后设置 local_files_only=True。

核心实现

以下是完整的实时录音转文本脚本。代码采用了多线程机制，录音与转录并行进行，同时集成了 VAD（语音活动检测）过滤静音片段。

# -*- coding: utf-8 -*-
import os
import sys
import time
import wave
import tempfile
import threading
import torch
import pyaudiowpatch as pyaudio
from faster_whisper import WhisperModel

AUDIO_BUFFER = 5

def record_audio(p, device):
    with tempfile.NamedTemporaryFile(suffix=".wav", delete=False) as f:
        filename = f.name
        wave_file = wave.open(filename, "wb")
        wave_file.setnchannels(int(device["maxInputChannels"]))
        wave_file.setsampwidth(p.get_sample_size(pyaudio.paInt16))
        wave_file.setframerate(int(device["defaultSampleRate"]))

        def callback(in_data, frame_count, time_info, status):
            wave_file.writeframes(in_data)
             (in_data, pyaudio.paContinue)

        :
            stream = p.(=pyaudio.paInt16, channels=(device[]),
                            rate=(device[]), frames_per_buffer=,
                            =, input_device_index=device[], stream_callback=callback)
            time.sleep(AUDIO_BUFFER)
         Exception  e:
            ()
        :
               ():
                stream.stop_stream()
                stream.close()
            wave_file.close()
         filename

 ():
    :
        segments, info = model.transcribe(
            filename, beam_size=, language=, vad_filter=,
            vad_parameters=(min_silence_duration_ms=)
        )
         segment  segments:
            ( % (segment.start, segment.end, segment.text))
     Exception  e:
        ()
    :
         os.path.exists(filename):
            os.remove(filename)

 ():
    ()
     torch.cuda.is_available():
        device = 
        compute_type = 
        ()
    :
        device = 
        compute_type = 
        ()

    model_path = 
    :
        model = WhisperModel(model_path, device=device, compute_type=compute_type, local_files_only=)
        ()
     Exception  e:
        ()
        

     pyaudio.PyAudio()  p:
        :
            default_mic = p.get_default_input_device_info()
            ()
            ( * )
            ()
             :
                filename = record_audio(p, default_mic)
                thread = threading.Thread(target=whisper_audio, args=(filename, model))
                thread.start()
         OSError:
            ()
         KeyboardInterrupt:
            ()
         Exception  e:
            ()

 __name__ == :
    main()