[egs][aishell][v1] wav name is not align with its path.

[vzxxbacq]

Thanks everyone's great work and I had a trouble when I run aishell/v1 project.
In aishell/v1/local/aishell_data_prep.sh script, the author write this code:

# get all wavs' path
find $audio_dir -iname "*.wav" | grep -i "wav/train" > $train_dir/wav.flist || exit 1;

# After some steps.....
    # get utt.list
    sed -e 's/\.wav//' $dir/wav.flist | awk -F '/' '{print( $NF)}' | sort > $dir/utt.list
    # then few steps later, get wav.scp_all
    paste -d' ' $dir/utt.list $dir/wav.flist > $dir/wav.scp_all

The wav.flist isn't sorted and utt.list is sorted, so I got a scp that wav isn't align with its path.

In wav.scp_all:

....
BAC009S0002W0240 /media/fhq/common/1data//data_aishell/wav/train/S0002/BAC009S0002W0225.wav
BAC009S0002W0241 /media/fhq/common/1data//data_aishell/wav/train/S0002/BAC009S0002W0226.wav
BAC009S0002W0242 /media/fhq/common/1data//data_aishell/wav/train/S0002/BAC009S0002W0227.wav
BAC009S0002W0243 /media/fhq/common/1data//data_aishell/wav/train/S0002/BAC009S0002W0228.wav
.....

And after I remove the sort step, I got a aligned scp file.
Is it a bug or I missed something? Thanks.

[danpovey]

@naxingyu can you please look at this?

[naxingyu]

on it.

…

On Fri, Sep 14, 2018 at 12:17 AM Daniel Povey ***@***.***> wrote: @naxingyu <https://github.com/naxingyu> can you please look at this? — You are receiving this because you were mentioned. Reply to this email directly, view it on GitHub <#2700 (comment)>, or mute the thread <https://github.com/notifications/unsubscribe-auth/ADKpxH0V7OcTUtMOZslNDP1wcS1gonQ-ks5uaoUWgaJpZM4Wnaid> .

[vzxxbacq]

issue fixed, thanks.