|
Download README.md from ml-resources/FunASR-bmodel: direct link, hf CLI and curl.
- Browser
- Download file 2.73 kB
-
https://huggingface.co/ml-resources/FunASR-bmodel/resolve/main/README.md
- Command line
-
hf download hf://ml-resources/FunASR-bmodel/README.md
-
curl -L -o README.md https://huggingface.co/ml-resources/FunASR-bmodel/resolve/main/README.md
2.73 kB
metadata
license: unknown
language:
- zh
pipeline_tag: automatic-speech-recognition
tags:
- funasr
- asr
- speech-recognition
- sophgo
- bmodel
- tpu
- edge
Mirror of wlc952/FunASR-bmodel from ModelScope. All credit to the original authors. Not an official release from them. No license was declared at the source; all rights remain with the original authors.
FunASR-bmodel for monitor
快速开始
git lfs install
git clone https://www.modelscope.cn/wlc952/FunASR-bmodel.git
cd FunASR-bmodel
uv sync
source .venv/bin/activate
cd bmodel # 注意工作目录是 ./bmodel
python monitor.py --rtsp rtsp://127.0.0.1:8554/stream1 --system-audio ../system_ref.wav --system-phrases ../system_phrases.txt --chunk-seconds 10
说明
- 本python项目使用uv管理,请先安装uv。参考安装方式:
# 推荐方式
curl -LsSf https://astral.sh/uv/install.sh | sh
# pipx 安装
sudo apt install pipx
pipx install uv
pipx ensurepath
虚拟环境创建/同步
uv sync source .venv/bin/activate关键文件
- 模型测试脚本
bmodel/offline.py,直接运行即可。
cd bmodel python offline.py- 对话监测脚本
bmodel/monitor.py,运行方式如下
cd bmodel python monitor.py --rtsp rtsp://172.24.64.225:8554/stream1 --system-audio ../system_ref.wav --system-phrases ../system_phrases.txt --chunk-seconds 10相关参数说明:
parser = argparse.ArgumentParser(description="监控场景下非系统人声检测与上报") parser.add_argument("--rtsp", required=True, help="RTSP 地址,例如 rtsp://user:pass@ip:554/Streaming/Channels/101") parser.add_argument( "--system-audio", default="../system_ref.wav", help="系统参考音频,作为前置拼接,系统说话人 spk=0", ) parser.add_argument("--out-dir", default="/data/alarms", help="告警本地保存目录 (已由 nginx 映射到 /static)") parser.add_argument("--chunk-seconds", type=int, default=10, help="检测窗口秒数,默认 10s") parser.add_argument( "--system-phrases", default="../system_phrases.txt", help="系统提示词文本库(每行一条),用于过滤系统播报文本。如果系统声音错误上报,请将对应的识别文本加入到该文件。", )还有一些参数需要直接修改代码:
- ip修改:搜索并修改
get_local_ip_fallback() - POST信息修改:搜索并修改
def post_alert(safety_url: str, when: str, timeout: int = 2)
- 模型测试脚本