语音识别代码linux,语音识别LinuxSDK（C++）正式发布

最新推荐文章于 2021-05-13 11:24:36 发布

转载最新推荐文章于 2021-05-13 11:24:36 发布 · 306 阅读

文章标签：

#语音识别代码linux

这篇博客记录了一次使用Baidu Speech SDK进行语音识别的过程，开始时遇到VAD启动错误，随后修复并成功识别出'小猪佩奇全集'，展示了配置参数和识别结果。

部署运行你感兴趣的模型镜像

[root@iz2ze3skq324qf4hkeg1isz ~]# pwd

/root

[root@iz2ze3skq324qf4hkeg1isz ~]# /tmp/xxj/sample/asrDemo2/run/asrDemo /tmp/xxj/filetmp/a.pcm

-------------==========----------

BEGIN!

0x7ffe69142bf8

------------------------------

[GlobalThreadPool::GlobalThreadPool] alloc thread num:30

FOR Feedback : Config is filled :

name: asr.config

string: asr.param.secret_key.string=037dc446820ec143d1628c20146b9d34; asr_param_key_app_id=10455099; asr_param_key_chunk_key.string=rKCHBLmYiFPuCQTS0HttLbUD; asr_param_key_mfe_cmvn_dat_file.string=../../../resources/asr_resource/bds_easr_mfe_cmvn.dat; asr_param_key_mfe_dnn_dat_file.string=../../../resources/asr_resource/bds_easr_mfe_dnn.dat; asr_param_key_product_id.string=15361;

int: asr_param_key_chunk_enable.bool=1; asr_param_key_compression_type.int=1; asr_param_key_enable_long_speech.bool=1; common_param_key_debug_log_level.int=0;

float: asr_param_key_max_speech_pause.float=30;

ERROR:[/home/hanyong01/work/work/sdk_3_0/baidu/speech-arch/baidu-speech-sdk/core/VAD/MFE/dnn/vad_nn.cpp:30]failed to open dnn model file[../../../resources/asr_resource/bds_easr_mfe_dnn.dat]

ERROR:[/home/hanyong01/work/work/sdk_3_0/baidu/speech-arch/baidu-speech-sdk/core/VAD/MFE/dnn/dnn_vad.cpp:347]DNN Init Failed

[2018-07-17 15:45:59.519][/tmp/xxj/filetmp/a.pcm] :识别错误， err_domain=20 ;err_code=1 ;err_desc=VAD start: start error. ;sn= 6B31949E-5D5D-4D1D-AF21-EC3A6EF98AFD

stop audio data error: sdk is finished and released : 0

[root@iz2ze3skq324qf4hkeg1isz ~]#

[root@iz2ze3skq324qf4hkeg1isz ~]# cd /tmp/xxj/sample/asrDemo2/run/

[root@iz2ze3skq324qf4hkeg1isz run]# ./asrDemo /tmp/xxj/filetmp/a.pcm

-------------==========----------

BEGIN!

0x7ffd175c6398

------------------------------

[GlobalThreadPool::GlobalThreadPool] alloc thread num:30

FOR Feedback : Config is filled :

name: asr.config

int: asr_param_key_chunk_enable.bool=1; asr_param_key_compression_type.int=1; asr_param_key_enable_long_speech.bool=1; common_param_key_debug_log_level.int=0;

float: asr_param_key_max_speech_pause.float=30;

ASR MULTIFILES BEGIN TO SLEEP

[2018-07-17 15:46:26.268][/tmp/xxj/filetmp/a.pcm] :识别结果：{"results_recognition":["小猪佩奇全集"],"origin_result":{"corpus_no":6579089251098034286,"err_no":0,"result":{"word":["小猪佩奇全集"]},"sn":"8059C319-7194-497C-B139-C5CB4CB29C15","voice_energy":24449.4394531250},"sn_start_time":"00:00.160","sn_end_time":"00:01.951"}

[2018-07-17 15:46:26.469][/tmp/xxj/filetmp/a.pcm] :语义理解结果：{"merged_res":{"semantic_form":{"appid":15361,"err_no":0,"parsed_text":"小猪佩奇全集 ","raw_text":"小猪佩奇全集","results":[]}}}

ASR MULTIFILES FINISHED

[root@iz2ze3skq324qf4hkeg1isz run]# pwd

/tmp/xxj/sample/asrDemo2/run

[root@iz2ze3skq324qf4hkeg1isz run]#

您可能感兴趣的与本文相关的镜像