Paraformer语音识别-中文-通用-16k-离线-轻量

Highlights

模型为Paraformer语音识别-中文-通用-16k-离线的ox量化导出版本，可以直接用来做生产部署，一键部署教程（点击此处）

模型转换及测试脚本

测试数据：https://isv-data.oss-c-hagzhou.aliyucs.com/ics/MaaS/ASR/testaudio/asrexample_zh.pcm

from fuasr_ox import Paraformer
from pathlib import Path

model_dir = "damo/speech_paraformer_asr_at-zh-c-16k-commo-vocab8358-tesorflow1"
model = Paraformer(model_dir, batch_size=1, quatize=True)

wav_path = path_to_asr_example_zh

result = model(wav_path)
prit(result)

• model_dir: model_ame i modelscope or local path dowloaded from modelscope. If the local path is set, it should cotai model.ox, cofig.yaml, am.mv
• batch_size: 1 (Default), the batch size duratio iferece
• device_id: -1 (Default), ifer o CPU. If you wat to ifer with GPU, set it to gpu_id (Please make sure that you have istall the oxrutime-gpu)
• quatize: False (Default), load the model of model.ox i model_dir. If set True, load the model of model_quat.ox i model_dir
• itra_op_um_threads: 4 (Default), sets the umber of threads used for itraop parallelism o CPU

参考教程：https://alibaba-damo-academy.github.io/FuASR/e/rutime/pytho/oxrutime/README.html

相关论文以及引用信息

@iproceedigs{gao2022paraformer,
  title={Paraformer: Fast ad Accurate Parallel Trasformer for No-autoregressive Ed-to-Ed Speech Recogitio},
  author={Gao, Zhifu ad Zhag, Shiliag ad McLoughli, Ia ad Ya, Zhijie},
  booktitle={INTERSPEECH},
  year={2022}
}

Highlights 模型为Paraformer语音识别-中文-通用-16k-离线的onnx量化导出版本，可以直接用来做生产部署，一键部署教程（点击此处） 模型转换及测试脚本 测试数据：https:/