dawjur/dawjur-mlworkshops
1
0
Fork 0
Code Issues Pull Requests Releases Wiki Activity

fix srr score

Browse Source
This commit is contained in:
Dawid Jurkiewicz 2020-04-03 15:38:02 +02:00
parent 2a89af348b
commit 2d97b3f37c
1 changed files with 1 additions and 3 deletions
Show Stats Download Patch File Download Diff File Expand all files Collapse all files

4
calc_metrics.sh
View File

@ -1,13 +1,11 @@
#!/usr/bin/env bash #!/usr/bin/env bash
cut -f2 wikiniews_results.tsv | awk 'BEGIN{FS=OFS="\t"}{print $0,"(sp1_"NR")"}' > hyp.trn cut -f2 wikiniews_results.tsv | awk 'BEGIN{FS=OFS="\t"}{print $0,"(sp1_"NR")"}' > hyp.trn
cut -f3 wikiniews_results.tsv | awk 'BEGIN{FS=OFS="\t"}{print $0,"(sp1_"NR")"}' > ref.trn cut -f3 wikiniews_results.tsv | awk 'BEGIN{FS=OFS="\t"}{print $0,"(sp1_"NR")"}' > ref.trn
sclite -f 0 -r ref.trn trn -h hyp.trn trn -e utf-8 -i rm -o sum stdout | grep "Sum/Avg" | python -c "import sys; import re; print('\t'.join(re.findall(r'(\d+\.\d+)', sys.stdin.read())[-2:]))" > wer_srr.txt
cut -f1 wer_srr.txt > wer_total.txt
cut -f2 wer_srr.txt >> srr.txt
sclite -f 0 -r ref.trn trn -h hyp.trn trn -e utf-8 -i rm -o pra stdout | grep "Scores:" | sed 's/Scores: (#C #S #D #I) // ' | awk '{ c=$1; s=$2; d=$3; i=$4; print ($s+$d+$i)/($s+$d+$c)}' > wer_all.txt sclite -f 0 -r ref.trn trn -h hyp.trn trn -e utf-8 -i rm -o pra stdout | grep "Scores:" | sed 's/Scores: (#C #S #D #I) // ' | awk '{ c=$1; s=$2; d=$3; i=$4; print ($s+$d+$i)/($s+$d+$c)}' > wer_all.txt
awk '{ sum += $1; n++ } END { print sum / n; }' < wer_all.txt >> wer.txt awk '{ sum += $1; n++ } END { print sum / n; }' < wer_all.txt >> wer.txt
python -c "print(`grep -c '^0$' wer_all.txt`/`cat wer_all.txt | wc -l`)" >> srr.txt
paste wikiniews_results.tsv wer.txt > wikinews_results.tsv paste wikiniews_results.tsv wer.txt > wikinews_results.tsv
rm wikiniews_results.tsv rm wikiniews_results.tsv