wenet-e2e/WenetSpeech

Segment fault when downloading

duj12 opened this issue · 3 comments

duj12 commented

segment fault happened for 3 times, I can't figure out the reason,
I can only comment the lines that have been downloaded in metadata/v1.list, to resume downloading

the following is the 3rd logging of segment fault error.

utils/download_wenetspeech.sh: Downloading audio/train/podcast/B00100.aes.tgz remote_md5=20b73cab2f0448a4b05baefab4174827
--2023-07-03 13:35:02--  http://wenet.meeting.tencent.com/WenetSpeech/audio/train/podcast/B00100.aes.tgz
正在解析主机 wenet.meeting.tencent.com (wenet.meeting.tencent.com)... 61.170.69.115, 61.170.72.232, 61.170.72.248
正在连接 wenet.meeting.tencent.com (wenet.meeting.tencent.com)|61.170.69.115|:80... 已连接。
已发出 HTTP 请求,正在等待回应... 200 OK
长度: 1457464864 (1.4G) [application/x-compressed]
正在保存至: “/data/megastore/Datasets/ASR/WenetSpeech/WenetSpeech/tar/audio/train/podcast/B00100.aes.tgz”

B00100.aes.tgz                             99%[====================================================================================> ]   1.36G  1.52MB/s    剩余 1s    utils/download_wenetspeech.sh:行 75: 95160 段错误               (核心已转储) wget -t 20 -T 90 -P $location $remote_obj

Is there any thing wrong?

me too

maybe network problem,try twice

可以试试从 modelscope 下载,详见 https://github.com/wenet-e2e/WenetSpeech#from-modelscope