mirrors/icefall
1
0
Fork 0
mirror of https://github.com/k2-fsa/icefall.git synced 2025-08-09 10:02:22 +00:00
Code Issues Packages Projects Releases Wiki Activity

new version of multi_ja_en prepare.sh script which swaps Librispeech for MLS English

Browse Source
This commit is contained in:
Kinan Martin 2025-05-09 10:57:41 +09:00
parent 547f5c5cfb
commit 21d1bf73bb
1 changed files with 44 additions and 40 deletions
Show Stats Download Patch File Download Diff File Expand all files Collapse all files

84
egs/multi_ja_en/ASR/prepare.sh
View File

@ -19,6 +19,8 @@ vocab_sizes=(
# All files generated by this script are saved in "data". # All files generated by this script are saved in "data".
# You can safely remove "data" and rerun this script to regenerate it. # You can safely remove "data" and rerun this script to regenerate it.
mkdir -p data mkdir -p data
mkdir -p data/lang
mkdir -p data/manifests
log() { log() {
# This function is from espnet # This function is from espnet
@ -31,55 +33,54 @@ log "dl_dir: $dl_dir"
log "Dataset: musan" log "Dataset: musan"
if [ $stage -le 1 ] && [ $stop_stage -ge 1 ]; then if [ $stage -le 1 ] && [ $stop_stage -ge 1 ]; then
log "Stage 1: Soft link fbank of musan" log "Stage 1: Soft link fbank of musan"
mkdir -p data/fbank
if [ -e ../../librispeech/ASR/data/fbank/.musan.done ]; then if [ -e ../../librispeech/ASR/data/fbank/.musan.done ]; then
cd data/fbank cd data/manifests
ln -svf $(realpath ../../../../librispeech/ASR/data/fbank/musan_feats) . mkdir -p musan
ln -svf $(realpath ../../../../librispeech/ASR/data/fbank/musan_cuts.jsonl.gz) . cd musan
cd ../.. ln -svf $(realpath ../../../../../librispeech/ASR/data/fbank/musan_feats) .
ln -svf $(realpath ../../../../../librispeech/ASR/data/fbank/musan_cuts.jsonl.gz) .
cd ../../..
else else
log "Abort! Please run ../../librispeech/ASR/prepare.sh --stage 4 --stop-stage 4" log "Abort! Please run ../../librispeech/ASR/prepare.sh --stage 4 --stop-stage 4"
exit 1 exit 1
fi fi
fi fi
log "Dataset: LibriSpeech" log "Dataset: MLS English"
if [ $stage -le 2 ] && [ $stop_stage -ge 2 ]; then if [ $stage -le 2 ] && [ $stop_stage -ge 2 ]; then
log "Stage 1: Soft link fbank of LibriSpeech" log "Stage 2: Soft link manifests (including fbank) of MLS English"
mkdir -p data/fbank if [ -e ../../mls_english/ASR/data/manifests/.mls_english-validated.done ]; then
if [ -e ../../librispeech/ASR/data/fbank/.librispeech.done ]; then cd data/manifests
cd data/fbank mkdir -p mls_english
ln -svf $(realpath ../../../../librispeech/ASR/data/fbank/librispeech_cuts*) . cd mls_english
ln -svf $(realpath ../../../../librispeech/ASR/data/fbank/librispeech_feats*) . ln -svf $(realpath ../../../../../mls_english/ASR/data/manifests/mls_eng_cuts*) .
cd ../.. ln -svf $(realpath ../../../../../mls_english/ASR/data/manifests/feats*) .
cd ../../..
else else
log "Abort! Please run ../../librispeech/ASR/prepare.sh --stage 1 --stop-stage 1 and ../../librispeech/ASR/prepare.sh --stage 3 --stop-stage 3" log "Abort! Please run ../../mls_english/ASR/prepare.sh --stage 1 --stop-stage 1"
exit 1 exit 1
fi fi
fi fi
log "Dataset: ReazonSpeech" log "Dataset: ReazonSpeech"
if [ $stage -le 3 ] && [ $stop_stage -ge 3 ]; then if [ $stage -le 3 ] && [ $stop_stage -ge 3 ]; then
log "Stage 2: Soft link fbank of ReazonSpeech" log "Stage 3: Soft link fbank of ReazonSpeech"
mkdir -p data/fbank
if [ -e ../../reazonspeech/ASR/data/manifests/.reazonspeech.done ]; then if [ -e ../../reazonspeech/ASR/data/manifests/.reazonspeech.done ]; then
cd data/fbank cd data/manifests
ln -svf $(realpath ../../../../reazonspeech/ASR/data/manifests/reazonspeech_cuts*) . mkdir -p reazonspeech
cd .. cd reazonspeech
mkdir -p manifests ln -svf $(realpath ../../../../../reazonspeech/ASR/data/manifests/reazonspeech_cuts*) .
cd manifests ln -svf $(realpath ../../../../../reazonspeech/ASR/data/manifests/feats*) .
ln -svf $(realpath ../../../../reazonspeech/ASR/data/manifests/feats_*) . cd ../../..
cd ../..
else else
log "Abort! Please run ../../reazonspeech/ASR/prepare.sh --stage 0 --stop-stage 2" log "Abort! Please run ../../reazonspeech/ASR/prepare.sh --stage 0 --stop-stage 2"
exit 1 exit 1
fi fi
fi fi
# New Stage 3: Prepare char based lang for ReazonSpeech if [ $stage -le 4 ] && [ $stop_stage -ge 4 ]; then
if [ $stage -le 3 ] && [ $stop_stage -ge 3 ]; then
lang_char_dir=data/lang_char lang_char_dir=data/lang_char
log "Stage 3: Prepare char based lang for ReazonSpeech" log "Stage 4: Prepare char-based lang for ReazonSpeech"
mkdir -p $lang_char_dir mkdir -p $lang_char_dir
# Prepare text # Prepare text
@ -89,7 +90,7 @@ if [ $stage -le 3 ] && [ $stop_stage -ge 3 ]; then
| ./local/text2token.py -t "char" > $lang_char_dir/text | ./local/text2token.py -t "char" > $lang_char_dir/text
fi fi
# jp word segmentation for text # Japanese word segmentation
if [ ! -f $lang_char_dir/text_words_segmentation ]; then if [ ! -f $lang_char_dir/text_words_segmentation ]; then
python3 ./local/text2segments.py \ python3 ./local/text2segments.py \
--input-file $lang_char_dir/text \ --input-file $lang_char_dir/text \
@ -106,29 +107,31 @@ if [ $stage -le 3 ] && [ $stop_stage -ge 3 ]; then
fi fi
if [ ! -f $lang_char_dir/L_disambig.pt ]; then if [ ! -f $lang_char_dir/L_disambig.pt ]; then
python3 ./local/prepare_char.py --lang-dir data/lang_char python3 ./local/prepare_char.py --lang-dir $lang_char_dir
fi fi
fi fi
if [ $stage -le 4 ] && [ $stop_stage -ge 4 ]; then
log "Stage 4: Prepare Byte BPE based lang"
mkdir -p data/fbank if [ $stage -le 5 ] && [ $stop_stage -ge 5 ]; then
log "Stage 5: Prepare Byte BPE based lang"
# Check if required char-based lang data exists
if [ ! -d ../../reazonspeech/ASR/data/lang_char ] && [ ! -d ./data/lang_char ]; then if [ ! -d ../../reazonspeech/ASR/data/lang_char ] && [ ! -d ./data/lang_char ]; then
log "Abort! Please run ../../reazonspeech/ASR/prepare.sh --stage 3 --stop-stage 3" log "Abort! Please run ../../reazonspeech/ASR/prepare.sh --stage 3 --stop-stage 3"
exit 1 exit 1
fi fi
if [ ! -d ../../librispeech/ASR/data/lang_bpe_500 ] && [ ! -d ./data/lang_bpe_500 ]; then # Check if BPE data from MLS English exists
log "Abort! Please run ../../librispeech/ASR/prepare.sh --stage 5 --stop-stage 5" if [ ! -d ../../mls_english/ASR/data/lang ] ; then
log "Abort! Please run ../../mls_english/ASR/prepare.sh --stage 3 --stop-stage 3"
exit 1 exit 1
fi fi
cd data/ cd data/
# if [ ! -d ./lang_char ]; then # Symlink the full lang directory from MLS English (which includes bpe_2000/ and transcript.txt)
# ln -svf $(realpath ../../../reazonspeech/ASR/data/lang_char) . if [ ! -d ./lang_bpe_2000 ]; then
# fi ln -svf $(realpath ../../../mls_english/ASR/data/lang) lang_bpe_2000
if [ ! -d ./lang_bpe_500 ]; then
ln -svf $(realpath ../../../librispeech/ASR/data/lang_bpe_500) .
fi fi
cd ../ cd ../
@ -136,7 +139,8 @@ if [ $stage -le 4 ] && [ $stop_stage -ge 4 ]; then
lang_dir=data/lang_bbpe_${vocab_size} lang_dir=data/lang_bbpe_${vocab_size}
mkdir -p $lang_dir mkdir -p $lang_dir
cat data/lang_char/text data/lang_bpe_500/transcript_words.txt \ # Combine Japanese char-level text and English BPE transcript
cat data/lang_char/text data/lang_bpe_2000/transcript.txt \
> $lang_dir/text > $lang_dir/text
if [ ! -f $lang_dir/transcript_chars.txt ]; then if [ ! -f $lang_dir/transcript_chars.txt ]; then
@ -150,7 +154,7 @@ if [ $stage -le 4 ] && [ $stop_stage -ge 4 ]; then
--input-file ./data/lang_char/text \ --input-file ./data/lang_char/text \
--output-file $lang_dir/text_words_segmentation --output-file $lang_dir/text_words_segmentation
cat ./data/lang_bpe_500/transcript_words.txt \ cat ./data/lang_bpe_2000/transcript.txt \
>> $lang_dir/text_words_segmentation >> $lang_dir/text_words_segmentation
fi fi