speech_recognition/FunASR
1
0
Fork 0
mirror of https://github.com/modelscope/FunASR synced 2025-09-15 14:48:36 +08:00
Code Issues Actions Packages Projects Releases Wiki Activity

update

Browse Source
This commit is contained in:
嘉渊 2023-07-21 11:24:31 +08:00
parent 311894a7aa
commit 5f3f194ffd
2 changed files with 42 additions and 9 deletions
Show Stats Download Patch File Download Diff File Expand all files Collapse all files

27
egs/callhome/eend_ola/local/gen_feats_scp.py Normal file
View File

@ -0,0 +1,27 @@
import os
import argparse
if __name__ == '__main__':
parser = argparse.ArgumentParser()
parser.add_argument("--root_path", type=str)
parser.add_argument("--out_path", type=str)
parser.add_argument("--split_num", type=int, default=64)
args = parser.parse_args()
root_path = args.root_path
out_path = args.out_path
datasets = ["train", "dev"]
split_num = args.split_num
for dataset in datasets:
with open(os.path.join(out_path, dataset, "feats.scp"), "w") as out_f:
for i in range(split_num):
idx = str(i + 1)
feature_file = os.path.join(root_path, dataset, "feature.scp.{}".format(idx))
label_file = os.path.join(root_path, dataset, "label.scp.{}".format(idx))
with open(feature_file) as ff, open(label_file) as fl:
ff_lines = ff.readlines()
fl_lines = fl.readlines()
for ff_line, fl_line in zip(ff_lines, fl_lines):
sample_name, f_path = ff_line.strip().split()
_, l_path = fl_line.strip().split()
out_f.write("{} {} {}\n".format(sample_name, f_path, l_path))

24
egs/callhome/eend_ola/run_test.sh
View File

@ -12,7 +12,7 @@ dump_cmd=utils/run.pl
nj=64
# feature configuration
data_dir="/nfs/wangjiaming.wjm/EEND_DATA_sad30_snr10n15n20/convert_chunk2000/data"
data_dir="/nfs/wangjiaming.wjm/EEND_DATA_sad30_snr10n15n20/convert_test/data"
simu_feats_dir="/nfs/wangjiaming.wjm/EEND_ARK_DATA/dump/simu_data/data"
simu_feats_dir_chunk2000="/nfs/wangjiaming.wjm/EEND_ARK_DATA/dump/simu_data_chunk2000/data"
callhome_feats_dir_chunk2000="/nfs/wangjiaming.wjm/EEND_ARK_DATA/dump/callhome_chunk2000/data"
@ -74,7 +74,7 @@ if [ ${stage} -le 0 ] && [ ${stop_stage} -ge 0 ]; then
simu_opts_sil_scale_array=(2 2 5 9)
simu_opts_num_train=100000
# for simulated data of chunk500
# for simulated data of chunk500 and chunk2000
for dset in swb_sre_tr swb_sre_cv; do
if [ "$dset" == "swb_sre_tr" ]; then
n_mixtures=${simu_opts_num_train}
@ -91,13 +91,19 @@ if [ ${stage} -le 0 ] && [ ${stop_stage} -ge 0 ]; then
# done
# utils/split_scp.pl "${data_dir}/simu/data/${simu_data_dir}/wav.scp" $split_scps || exit 1
# python local/split.py ${data_dir}/simu/data/${simu_data_dir}
output_dir=${data_dir}/ark_data/dump/simu_data/$dataset
mkdir -p $output_dir/.logs
$dump_cmd --max-jobs-run $nj JOB=1:$nj $output_dir/.logs/dump.JOB.log \
python local/dump_feature.py \
--data_dir ${data_dir}/simu/data/${simu_data_dir}/.work \
--output_dir ${data_dir}/ark_data/dump/simu_data/$dataset \
--index JOB
# # for chunk_size=500
# output_dir=${data_dir}/ark_data/dump/simu_data/$dataset
# mkdir -p $output_dir/.logs
# $dump_cmd --max-jobs-run $nj JOB=1:$nj $output_dir/.logs/dump.JOB.log \
# python local/dump_feature.py \
# --data_dir ${data_dir}/simu/data/${simu_data_dir}/.work \
# --output_dir ${data_dir}/ark_data/dump/simu_data/$dataset \
# --index JOB
mkdir -p ${data_dir}/ark_data/dump/simu_data/data/$dataset
python local_rank/gen_feats_scp.py \
--root_path ${data_dir}/ark_data/dump/simu_data \
--out_path ${data_dir}/ark_data/dump/simu_data/data/$dataset \
--split_num $nj
done
fi