support BPE based lang

egs/gigaspeech/ASR/local/convert_transcript_words_to_tokens.py

@@ -0,0 +1 @@
+../../../librispeech/ASR/local/convert_transcript_words_to_tokens.py

egs/gigaspeech/ASR/local/prepare_lang.py

@@ -0,0 +1 @@
+../../../librispeech/ASR/local/prepare_lang.py

egs/gigaspeech/ASR/local/prepare_lang_bpe.py

@@ -0,0 +1 @@
+../../../librispeech/ASR/local/prepare_lang_bpe.py

egs/gigaspeech/ASR/local/train_bpe_model.py

@@ -0,0 +1 @@
+../../../librispeech/ASR/local/train_bpe_model.py

egs/gigaspeech/ASR/prepare.sh

@@ -31,9 +31,9 @@ dl_dir=$PWD/download
 # data/lang_bpe_yyy if the array contains xxx, yyy
 vocab_sizes=(
   5000
-  2000
-  1000
-  500
+  # 2000
+  # 1000
+  # 500
 )
 
 # All files generated by this script are saved in "data".
@@ -125,15 +125,61 @@ if [ $stage -le 5 ] && [ $stop_stage -ge 5 ]; then
   lang_dir=data/lang_phone
   mkdir -p $lang_dir
 
-  (echo '!SIL SIL'; echo '<SPOKEN_NOISE> SPN'; echo '<UNK> SPN'; ) |
-    cat - $dl_dir/lm/librispeech-lexicon.txt |
-    sort | uniq > $lang_dir/lexicon.txt
+  # (echo '!SIL SIL'; echo '<SPOKEN_NOISE> SPN'; echo '<UNK> SPN'; ) |
+  #   cat - $dl_dir/lm/librispeech-lexicon.txt |
+  #   sort | uniq > $lang_dir/lexicon.txt
 
-  if [ ! -f $lang_dir/L_disambig.pt ]; then
-    ./local/prepare_lang.py --lang-dir $lang_dir
+  # if [ ! -f $lang_dir/L_disambig.pt ]; then
+  #   ./local/prepare_lang.py --lang-dir $lang_dir
+  # fi
+  if [ ! -f $lang_dir/transcript_words.txt ]; then
+    gunzip -c "data/manifests/gigaspeech_supervisions_XL.jsonl.gz" \
+      | jq '.text' \
+      | sed 's/"//g' \
+      > $lang_dir/transcript_words.txt
+
+    # Delete utterances with garbage meta tags
+    garbage_utterance_tags="<SIL> <MUSIC> <NOISE> <OTHER>"
+    for tag in $garbage_utterance_tags; do
+      sed -i "/${tag}/d" $lang_dir/transcript_words.txt
+    done
+
+    # Delete punctuations in utterances
+    punctuation_tags="<COMMA> <EXCLAMATIONPOINT> <PERIOD> <QUESTIONMARK>"
+    for tag in $punctuation_tags; do
+      sed -i "s/${tag}//g" $lang_dir/transcript_words.txt
+    done
+
+    # Ensure space only appears once
+    sed -i 's/\t/ /g' $lang_dir/transcript_words.txt
+    sed -i 's/[ ][ ]*/ /g' $lang_dir/transcript_words.txt
   fi
-fi
 
+  cat $lang_dir/transcript_words.txt | sed 's| |\n|g' \
+    | sort -u | sed '/^$/d' > $lang_dir/words.txt
+  (echo '!SIL'; echo '<SPOKEN_NOISE>'; echo '<UNK>'; ) |
+    cat - $lang_dir/words.txt | sort | uniq | awk '
+    BEGIN {
+      print "<eps> 0";
+    }
+    {
+      if ($1 == "<s>") {
+        print "<s> is in the vocabulary!" | "cat 1>&2"
+        exit 1;
+      }
+      if ($1 == "</s>") {
+        print "</s> is in the vocabulary!" | "cat 1>&2"
+        exit 1;
+      }
+      printf("%s %d\n", $1, NR);
+    }
+    END {
+      printf("#0 %d\n", NR+1);
+      printf("<s> %d\n", NR+2);
+      printf("</s> %d\n", NR+3);
+    }' > $lang_dir/words || exit 1;
+  mv $lang_dir/words $lang_dir/words.txt
+fi
 
 if [ $stage -le 6 ] && [ $stop_stage -ge 6 ]; then
   log "Stage 6: Prepare BPE based lang"
@@ -141,26 +187,24 @@ if [ $stage -le 6 ] && [ $stop_stage -ge 6 ]; then
   for vocab_size in ${vocab_sizes[@]}; do
     lang_dir=data/lang_bpe_${vocab_size}
     mkdir -p $lang_dir
-    # We reuse words.txt from phone based lexicon
-    # so that the two can share G.pt later.
-    cp data/lang_phone/words.txt $lang_dir
+    # # We reuse words.txt from phone based lexicon
+    # # so that the two can share G.pt later.
+    cp data/lang_phone/{words.txt,transcript_words.txt} $lang_dir
 
     if [ ! -f $lang_dir/transcript_words.txt ]; then
       log "Generate data for BPE training"
-      files=$(
-        find "$dl_dir/LibriSpeech/train-clean-100" -name "*.trans.txt"
-        find "$dl_dir/LibriSpeech/train-clean-360" -name "*.trans.txt"
-        find "$dl_dir/LibriSpeech/train-other-500" -name "*.trans.txt"
-      )
-      for f in ${files[@]}; do
-        cat $f | cut -d " " -f 2-
-      done > $lang_dir/transcript_words.txt
+      gunzip -c "data/manifests/gigaspeech_supervisions_XL.jsonl.gz" \
+        | jq '.text' \
+        | sed 's/"//g' \
+        > $lang_dir/transcript_words.txt
     fi
 
-    ./local/train_bpe_model.py \
-      --lang-dir $lang_dir \
-      --vocab-size $vocab_size \
-      --transcript $lang_dir/transcript_words.txt
+    if [ ! -f $lang_dir/bpe.model ]; then
+      ./local/train_bpe_model.py \
+        --lang-dir $lang_dir \
+        --vocab-size $vocab_size \
+        --transcript $lang_dir/transcript_words.txt
+    fi
 
     if [ ! -f $lang_dir/L_disambig.pt ]; then
       ./local/prepare_lang_bpe.py --lang-dir $lang_dir