Fix for workflow (#934)

2025-08-08 09:32:20 +00:00 · 2023-03-09 17:38:15 +08:00 · 2023-03-09 17:38:15 +08:00 · 28af269e5e
commit 28af269e5e
parent f5de2e90c6
56 changed files with 82 additions and 159 deletions
--- a/.github/scripts/compute-fbank-librispeech-test-clean-and-test-other.sh
+++ b/.github/scripts/compute-fbank-librispeech-test-clean-and-test-other.sh
@ -15,5 +15,5 @@ mkdir -p data
 cd data
 [ ! -e fbank ] && ln -s ~/tmp/fbank-libri fbank
 cd ..
-./local/compute_fbank_librispeech.py
+./local/compute_fbank_librispeech.py --dataset 'test-clean test-other'
 ls -lh data/fbank/
--- a/.github/scripts/run-aishell-pruned-transducer-stateless3-2022-06-20.sh
+++ b/.github/scripts/run-aishell-pruned-transducer-stateless3-2022-06-20.sh
@ -25,7 +25,6 @@ repo=$(basename $repo_url)

 log "Display test files"
 tree $repo/
-soxi $repo/test_wavs/*.wav
 ls -lh $repo/test_wavs/*.wav

 pushd $repo/exp
--- a/.github/scripts/run-librispeech-conformer-ctc3-2022-11-28.sh
+++ b/.github/scripts/run-librispeech-conformer-ctc3-2022-11-28.sh
@ -18,7 +18,6 @@ repo=$(basename $repo_url)

 log "Display test files"
 tree $repo/
-soxi $repo/test_wavs/*.wav
 ls -lh $repo/test_wavs/*.wav

 pushd $repo/exp
--- a/.github/scripts/run-librispeech-lstm-transducer-stateless2-2022-09-03.sh
+++ b/.github/scripts/run-librispeech-lstm-transducer-stateless2-2022-09-03.sh
@ -20,7 +20,6 @@ abs_repo=$(realpath $repo)

 log "Display test files"
 tree $repo/
-soxi $repo/test_wavs/*.wav
 ls -lh $repo/test_wavs/*.wav

 pushd $repo/exp
--- a/.github/scripts/run-librispeech-pruned-transducer-stateless-2022-03-12.sh
+++ b/.github/scripts/run-librispeech-pruned-transducer-stateless-2022-03-12.sh
@ -19,7 +19,6 @@ repo=$(basename $repo_url)

 log "Display test files"
 tree $repo/
-soxi $repo/test_wavs/*.wav
 ls -lh $repo/test_wavs/*.wav

 for sym in 1 2 3; do
--- a/.github/scripts/run-librispeech-pruned-transducer-stateless2-2022-04-29.sh
+++ b/.github/scripts/run-librispeech-pruned-transducer-stateless2-2022-04-29.sh
@ -23,7 +23,6 @@ popd

 log "Display test files"
 tree $repo/
-soxi $repo/test_wavs/*.wav
 ls -lh $repo/test_wavs/*.wav

 pushd $repo/exp
--- a/.github/scripts/run-librispeech-pruned-transducer-stateless3-2022-04-29.sh
+++ b/.github/scripts/run-librispeech-pruned-transducer-stateless3-2022-04-29.sh
@ -22,7 +22,6 @@ popd

 log "Display test files"
 tree $repo/
-soxi $repo/test_wavs/*.wav
 ls -lh $repo/test_wavs/*.wav

 pushd $repo/exp
--- a/.github/scripts/run-librispeech-pruned-transducer-stateless3-2022-05-13.sh
+++ b/.github/scripts/run-librispeech-pruned-transducer-stateless3-2022-05-13.sh
@ -19,7 +19,6 @@ repo=$(basename $repo_url)

 log "Display test files"
 tree $repo/
-soxi $repo/test_wavs/*.wav
 ls -lh $repo/test_wavs/*.wav

 pushd $repo/exp
--- a/.github/scripts/run-librispeech-pruned-transducer-stateless5-2022-05-13.sh
+++ b/.github/scripts/run-librispeech-pruned-transducer-stateless5-2022-05-13.sh
@ -19,7 +19,6 @@ repo=$(basename $repo_url)

 log "Display test files"
 tree $repo/
-soxi $repo/test_wavs/*.wav
 ls -lh $repo/test_wavs/*.wav

 pushd $repo/exp
--- a/.github/scripts/run-librispeech-pruned-transducer-stateless7-2022-11-11.sh
+++ b/.github/scripts/run-librispeech-pruned-transducer-stateless7-2022-11-11.sh
@ -19,7 +19,6 @@ repo=$(basename $repo_url)

 log "Display test files"
 tree $repo/
-soxi $repo/test_wavs/*.wav
 ls -lh $repo/test_wavs/*.wav

 pushd $repo/exp
--- a/.github/scripts/run-librispeech-pruned-transducer-stateless7-ctc-2022-12-01.sh
+++ b/.github/scripts/run-librispeech-pruned-transducer-stateless7-ctc-2022-12-01.sh
@ -18,7 +18,6 @@ repo=$(basename $repo_url)

 log "Display test files"
 tree $repo/
-soxi $repo/test_wavs/*.wav
 ls -lh $repo/test_wavs/*.wav

 pushd $repo/exp
@ -148,4 +147,4 @@ if [[ x"${GITHUB_EVENT_NAME}" == x"schedule" || x"${GITHUB_EVENT_LABEL_NAME}" ==
  done

  rm pruned_transducer_stateless7_ctc/exp/*.pt
-fi
+fi
--- a/.github/scripts/run-librispeech-pruned-transducer-stateless7-ctc-bs-2022-12-15.sh
+++ b/.github/scripts/run-librispeech-pruned-transducer-stateless7-ctc-bs-2022-12-15.sh
@ -10,7 +10,7 @@ log() {

 cd egs/librispeech/ASR

-repo_url=https://huggingface.co/yfyeung/icefall-asr-librispeech-pruned_transducer_stateless7_ctc_bs-2022-12-14
+repo_url=https://huggingface.co/yfyeung/icefall-asr-librispeech-pruned_transducer_stateless7_ctc_bs-2023-01-29

 log "Downloading pre-trained model from $repo_url"
 GIT_LFS_SKIP_SMUDGE=1 git clone $repo_url
@ -18,7 +18,6 @@ repo=$(basename $repo_url)

 log "Display test files"
 tree $repo/
-soxi $repo/test_wavs/*.wav
 ls -lh $repo/test_wavs/*.wav

 pushd $repo/exp
--- a/.github/scripts/run-librispeech-pruned-transducer-stateless7-streaming-2022-12-29.sh
+++ b/.github/scripts/run-librispeech-pruned-transducer-stateless7-streaming-2022-12-29.sh
@ -19,7 +19,6 @@ repo=$(basename $repo_url)

 log "Display test files"
 tree $repo/
-soxi $repo/test_wavs/*.wav
 ls -lh $repo/test_wavs/*.wav

 pushd $repo
--- a/.github/scripts/run-librispeech-pruned-transducer-stateless8-2022-11-14.sh
+++ b/.github/scripts/run-librispeech-pruned-transducer-stateless8-2022-11-14.sh
@ -19,7 +19,6 @@ repo=$(basename $repo_url)

 log "Display test files"
 tree $repo/
-soxi $repo/test_wavs/*.wav
 ls -lh $repo/test_wavs/*.wav

 pushd $repo/exp
--- a/.github/scripts/run-librispeech-streaming-pruned-transducer-stateless2-2022-06-26.sh
+++ b/.github/scripts/run-librispeech-streaming-pruned-transducer-stateless2-2022-06-26.sh
@ -19,7 +19,6 @@ repo=$(basename $repo_url)

 log "Display test files"
 tree $repo/
-soxi $repo/test_wavs/*.wav
 ls -lh $repo/test_wavs/*.wav

 pushd $repo/exp
--- a/.github/scripts/run-librispeech-transducer-stateless2-2022-04-19.sh
+++ b/.github/scripts/run-librispeech-transducer-stateless2-2022-04-19.sh
@ -19,7 +19,6 @@ repo=$(basename $repo_url)

 log "Display test files"
 tree $repo/
-soxi $repo/test_wavs/*.wav
 ls -lh $repo/test_wavs/*.wav

 for sym in 1 2 3; do
--- a/.github/scripts/run-librispeech-zipformer-mmi-2022-12-08.sh
+++ b/.github/scripts/run-librispeech-zipformer-mmi-2022-12-08.sh
@ -18,7 +18,6 @@ repo=$(basename $repo_url)

 log "Display test files"
 tree $repo/
-soxi $repo/test_wavs/*.wav
 ls -lh $repo/test_wavs/*.wav

 pushd $repo/exp
--- a/.github/scripts/run-pre-trained-conformer-ctc.sh
+++ b/.github/scripts/run-pre-trained-conformer-ctc.sh
@ -19,7 +19,6 @@ repo=$(basename $repo_url)

 log "Display test files"
 tree $repo/
-soxi $repo/test_wavs/*.flac
 ls -lh $repo/test_wavs/*.flac

 log "CTC decoding"
--- a/.github/scripts/run-pre-trained-transducer-stateless-librispeech-100h.sh
+++ b/.github/scripts/run-pre-trained-transducer-stateless-librispeech-100h.sh
@ -19,7 +19,6 @@ repo=$(basename $repo_url)

 log "Display test files"
 tree $repo/
-soxi $repo/test_wavs/*.wav
 ls -lh $repo/test_wavs/*.wav

 for sym in 1 2 3; do
--- a/.github/scripts/run-pre-trained-transducer-stateless-librispeech-960h.sh
+++ b/.github/scripts/run-pre-trained-transducer-stateless-librispeech-960h.sh
@ -19,7 +19,6 @@ repo=$(basename $repo_url)

 log "Display test files"
 tree $repo/
-soxi $repo/test_wavs/*.wav
 ls -lh $repo/test_wavs/*.wav

 for sym in 1 2 3; do
--- a/.github/scripts/run-pre-trained-transducer-stateless-modified-2-aishell.sh
+++ b/.github/scripts/run-pre-trained-transducer-stateless-modified-2-aishell.sh
@ -19,7 +19,6 @@ repo=$(basename $repo_url)

 log "Display test files"
 tree $repo/
-soxi $repo/test_wavs/*.wav
 ls -lh $repo/test_wavs/*.wav

 for sym in 1 2 3; do
--- a/.github/scripts/run-pre-trained-transducer-stateless-modified-aishell.sh
+++ b/.github/scripts/run-pre-trained-transducer-stateless-modified-aishell.sh
@ -19,7 +19,6 @@ repo=$(basename $repo_url)

 log "Display test files"
 tree $repo/
-soxi $repo/test_wavs/*.wav
 ls -lh $repo/test_wavs/*.wav

 for sym in 1 2 3; do
--- a/.github/scripts/run-pre-trained-transducer-stateless.sh
+++ b/.github/scripts/run-pre-trained-transducer-stateless.sh
@ -19,7 +19,6 @@ repo=$(basename $repo_url)

 log "Display test files"
 tree $repo/
-soxi $repo/test_wavs/*.wav
 ls -lh $repo/test_wavs/*.wav

 for sym in 1 2 3; do
--- a/.github/scripts/run-pre-trained-transducer.sh
+++ b/.github/scripts/run-pre-trained-transducer.sh
@ -19,7 +19,6 @@ repo=$(basename $repo_url)

 log "Display test files"
 tree $repo/
-soxi $repo/test_wavs/*.wav
 ls -lh $repo/test_wavs/*.wav

 log "Beam search decoding"
--- a/.github/scripts/run-wenetspeech-pruned-transducer-stateless2.sh
+++ b/.github/scripts/run-wenetspeech-pruned-transducer-stateless2.sh
@ -20,7 +20,6 @@ repo=$(basename $repo_url)

 log "Display test files"
 tree $repo/
-soxi $repo/test_wavs/*.wav
 ls -lh $repo/test_wavs/*.wav

 pushd $repo/exp
--- a/.github/scripts/test-ncnn-export.sh
+++ b/.github/scripts/test-ncnn-export.sh
@ -232,70 +232,3 @@ python3 ./pruned_transducer_stateless7_streaming/streaming-ncnn-decode.py \

 rm -rf $repo
 log "--------------------------------------------------------------------------"
-
-# Go back to the root directory of icefall repo
-popd
-
-pushd egs/csj/ASR
-
-log "=========================================================================="
-repo_url=https://huggingface.co/TeoWenShen/icefall-asr-csj-pruned-transducer-stateless7-streaming-230208
-GIT_LFS_SKIP_SMUDGE=1 git clone $repo_url
-repo=$(basename $repo_url)
-
-pushd $repo
-git lfs pull --include "exp_fluent/pretrained.pt"
-git lfs pull --include "exp_disfluent/pretrained.pt"
-
-cd exp_fluent
-ln -s pretrained.pt epoch-99.pt
-
-cd ../exp_disfluent
-ln -s pretrained.pt epoch-99.pt
-
-cd ../test_wavs
-git lfs pull --include "*.wav"
-popd
-
-log "Export via torch.jit.trace()"
-
-for exp in exp_fluent exp_disfluent; do
-  ./pruned_transducer_stateless7_streaming/export-for-ncnn.py \
-    --exp-dir $repo/$exp/ \
-    --lang $repo/data/lang_char \
-    --epoch 99 \
-    --avg 1 \
-    --use-averaged-model 0 \
-    \
-    --decode-chunk-len 32 \
-    --num-left-chunks 4 \
-    --num-encoder-layers "2,4,3,2,4" \
-    --feedforward-dims  "1024,1024,2048,2048,1024" \
-    --nhead "8,8,8,8,8" \
-    --encoder-dims "384,384,384,384,384" \
-    --attention-dims "192,192,192,192,192" \
-    --encoder-unmasked-dims "256,256,256,256,256" \
-    --zipformer-downsampling-factors "1,2,4,8,2" \
-    --cnn-module-kernels "31,31,31,31,31" \
-    --decoder-dim 512 \
-    --joiner-dim 512
-
-  pnnx $repo/$exp/encoder_jit_trace-pnnx.pt
-  pnnx $repo/$exp/decoder_jit_trace-pnnx.pt
-  pnnx $repo/$exp/joiner_jit_trace-pnnx.pt
-
-  for wav in aps-smp.wav interview_aps-smp.wav reproduction-smp.wav sps-smp.wav; do
-    python3 ./pruned_transducer_stateless7_streaming/streaming-ncnn-decode.py \
-      --tokens $repo/data/lang_char/tokens.txt \
-      --encoder-param-filename $repo/$exp/encoder_jit_trace-pnnx.ncnn.param \
-      --encoder-bin-filename $repo/$exp/encoder_jit_trace-pnnx.ncnn.bin \
-      --decoder-param-filename $repo/$exp/decoder_jit_trace-pnnx.ncnn.param \
-      --decoder-bin-filename $repo/$exp/decoder_jit_trace-pnnx.ncnn.bin \
-      --joiner-param-filename $repo/$exp/joiner_jit_trace-pnnx.ncnn.param \
-      --joiner-bin-filename $repo/$exp/joiner_jit_trace-pnnx.ncnn.bin \
-      $repo/test_wavs/$wav
-  done
-done
-
-rm -rf $repo
-log "--------------------------------------------------------------------------"
--- a/.github/workflows/run-aishell-2022-06-20.yml
+++ b/.github/workflows/run-aishell-2022-06-20.yml
@ -65,7 +65,7 @@ jobs:
        run: |
          grep -v '^#' ./requirements-ci.txt  | xargs -n 1 -L 1 pip install
          pip uninstall -y protobuf
-          pip install --no-binary protobuf protobuf
+          pip install --no-binary protobuf protobuf==3.20.*

      - name: Cache kaldifeat
        id: my-cache
@ -87,7 +87,7 @@ jobs:
          GITHUB_EVENT_NAME: ${{ github.event_name }}
          GITHUB_EVENT_LABEL_NAME: ${{ github.event.label.name }}
        run: |
-          sudo apt-get -qq install git-lfs tree sox
+          sudo apt-get -qq install git-lfs tree
          export PYTHONPATH=$PWD:$PYTHONPATH
          export PYTHONPATH=~/tmp/kaldifeat/kaldifeat/python:$PYTHONPATH
          export PYTHONPATH=~/tmp/kaldifeat/build/lib:$PYTHONPATH
--- a/.github/workflows/run-gigaspeech-2022-05-13.yml
+++ b/.github/workflows/run-gigaspeech-2022-05-13.yml
@ -64,7 +64,7 @@ jobs:
        run: |
          grep -v '^#' ./requirements-ci.txt  | xargs -n 1 -L 1 pip install
          pip uninstall -y protobuf
-          pip install --no-binary protobuf protobuf
+          pip install --no-binary protobuf protobuf==3.20.*

      - name: Cache kaldifeat
        id: my-cache
--- a/.github/workflows/run-librispeech-2022-03-12.yml
+++ b/.github/workflows/run-librispeech-2022-03-12.yml
@ -64,7 +64,7 @@ jobs:
        run: |
          grep -v '^#' ./requirements-ci.txt  | xargs -n 1 -L 1 pip install
          pip uninstall -y protobuf
-          pip install --no-binary protobuf protobuf
+          pip install --no-binary protobuf protobuf==3.20.*

      - name: Cache kaldifeat
        id: my-cache
@ -123,7 +123,7 @@ jobs:
          ln -sfv ~/tmp/fbank-libri egs/librispeech/ASR/data/fbank
          ls -lh egs/librispeech/ASR/data/*

-          sudo apt-get -qq install git-lfs tree sox
+          sudo apt-get -qq install git-lfs tree
          export PYTHONPATH=$PWD:$PYTHONPATH
          export PYTHONPATH=~/tmp/kaldifeat/kaldifeat/python:$PYTHONPATH
          export PYTHONPATH=~/tmp/kaldifeat/build/lib:$PYTHONPATH
--- a/.github/workflows/run-librispeech-2022-04-29.yml
+++ b/.github/workflows/run-librispeech-2022-04-29.yml
@ -64,7 +64,7 @@ jobs:
        run: |
          grep -v '^#' ./requirements-ci.txt  | xargs -n 1 -L 1 pip install
          pip uninstall -y protobuf
-          pip install --no-binary protobuf protobuf
+          pip install --no-binary protobuf protobuf==3.20.*

      - name: Cache kaldifeat
        id: my-cache
@ -123,7 +123,7 @@ jobs:
          ln -sfv ~/tmp/fbank-libri egs/librispeech/ASR/data/fbank
          ls -lh egs/librispeech/ASR/data/*

-          sudo apt-get -qq install git-lfs tree sox
+          sudo apt-get -qq install git-lfs tree
          export PYTHONPATH=$PWD:$PYTHONPATH
          export PYTHONPATH=~/tmp/kaldifeat/kaldifeat/python:$PYTHONPATH
          export PYTHONPATH=~/tmp/kaldifeat/build/lib:$PYTHONPATH
--- a/.github/workflows/run-librispeech-2022-05-13.yml
+++ b/.github/workflows/run-librispeech-2022-05-13.yml
@ -64,7 +64,7 @@ jobs:
        run: |
          grep -v '^#' ./requirements-ci.txt  | xargs -n 1 -L 1 pip install
          pip uninstall -y protobuf
-          pip install --no-binary protobuf protobuf
+          pip install --no-binary protobuf protobuf==3.20.*

      - name: Cache kaldifeat
        id: my-cache
@ -123,7 +123,7 @@ jobs:
          ln -sfv ~/tmp/fbank-libri egs/librispeech/ASR/data/fbank
          ls -lh egs/librispeech/ASR/data/*

-          sudo apt-get -qq install git-lfs tree sox
+          sudo apt-get -qq install git-lfs tree
          export PYTHONPATH=$PWD:$PYTHONPATH
          export PYTHONPATH=~/tmp/kaldifeat/kaldifeat/python:$PYTHONPATH
          export PYTHONPATH=~/tmp/kaldifeat/build/lib:$PYTHONPATH
--- a/.github/workflows/run-librispeech-2022-11-11-stateless7.yml
+++ b/.github/workflows/run-librispeech-2022-11-11-stateless7.yml
@ -64,7 +64,7 @@ jobs:
        run: |
          grep -v '^#' ./requirements-ci.txt  | xargs -n 1 -L 1 pip install
          pip uninstall -y protobuf
-          pip install --no-binary protobuf protobuf
+          pip install --no-binary protobuf protobuf==3.20.*

      - name: Cache kaldifeat
        id: my-cache
@ -123,7 +123,7 @@ jobs:
          ln -sfv ~/tmp/fbank-libri egs/librispeech/ASR/data/fbank
          ls -lh egs/librispeech/ASR/data/*

-          sudo apt-get -qq install git-lfs tree sox
+          sudo apt-get -qq install git-lfs tree
          export PYTHONPATH=$PWD:$PYTHONPATH
          export PYTHONPATH=~/tmp/kaldifeat/kaldifeat/python:$PYTHONPATH
          export PYTHONPATH=~/tmp/kaldifeat/build/lib:$PYTHONPATH
--- a/.github/workflows/run-librispeech-2022-11-14-stateless8.yml
+++ b/.github/workflows/run-librispeech-2022-11-14-stateless8.yml
@ -64,7 +64,7 @@ jobs:
        run: |
          grep -v '^#' ./requirements-ci.txt  | xargs -n 1 -L 1 pip install
          pip uninstall -y protobuf
-          pip install --no-binary protobuf protobuf
+          pip install --no-binary protobuf protobuf==3.20.*

      - name: Cache kaldifeat
        id: my-cache
@ -123,7 +123,7 @@ jobs:
          ln -sfv ~/tmp/fbank-libri egs/librispeech/ASR/data/fbank
          ls -lh egs/librispeech/ASR/data/*

-          sudo apt-get -qq install git-lfs tree sox
+          sudo apt-get -qq install git-lfs tree
          export PYTHONPATH=$PWD:$PYTHONPATH
          export PYTHONPATH=~/tmp/kaldifeat/kaldifeat/python:$PYTHONPATH
          export PYTHONPATH=~/tmp/kaldifeat/build/lib:$PYTHONPATH
--- a/.github/workflows/run-librispeech-2022-12-01-stateless7-ctc.yml
+++ b/.github/workflows/run-librispeech-2022-12-01-stateless7-ctc.yml
@ -60,7 +60,7 @@ jobs:
        run: |
          grep -v '^#' ./requirements-ci.txt  | xargs -n 1 -L 1 pip install
          pip uninstall -y protobuf
-          pip install --no-binary protobuf protobuf
+          pip install --no-binary protobuf protobuf==3.20.*

      - name: Cache kaldifeat
        id: my-cache
@ -119,7 +119,7 @@ jobs:
          ln -sfv ~/tmp/fbank-libri egs/librispeech/ASR/data/fbank
          ls -lh egs/librispeech/ASR/data/*

-          sudo apt-get -qq install git-lfs tree sox
+          sudo apt-get -qq install git-lfs tree
          export PYTHONPATH=$PWD:$PYTHONPATH
          export PYTHONPATH=~/tmp/kaldifeat/kaldifeat/python:$PYTHONPATH
          export PYTHONPATH=~/tmp/kaldifeat/build/lib:$PYTHONPATH
--- a/.github/workflows/run-librispeech-2022-12-08-zipformer-mmi.yml
+++ b/.github/workflows/run-librispeech-2022-12-08-zipformer-mmi.yml
@ -64,7 +64,7 @@ jobs:
        run: |
          grep -v '^#' ./requirements-ci.txt  | xargs -n 1 -L 1 pip install
          pip uninstall -y protobuf
-          pip install --no-binary protobuf protobuf
+          pip install --no-binary protobuf protobuf==3.20.*

      - name: Cache kaldifeat
        id: my-cache
@ -123,7 +123,7 @@ jobs:
          ln -sfv ~/tmp/fbank-libri egs/librispeech/ASR/data/fbank
          ls -lh egs/librispeech/ASR/data/*

-          sudo apt-get -qq install git-lfs tree sox
+          sudo apt-get -qq install git-lfs tree
          export PYTHONPATH=$PWD:$PYTHONPATH
          export PYTHONPATH=~/tmp/kaldifeat/kaldifeat/python:$PYTHONPATH
          export PYTHONPATH=~/tmp/kaldifeat/build/lib:$PYTHONPATH
--- a/.github/workflows/run-librispeech-2022-12-15-stateless7-ctc-bs.yml
+++ b/.github/workflows/run-librispeech-2022-12-15-stateless7-ctc-bs.yml
@ -35,7 +35,7 @@ on:

 jobs:
  run_librispeech_2022_12_15_zipformer_ctc_bs:
-    if: github.event.label.name == 'ready' || github.event.label.name == 'run-decode' || github.event.label.name == 'blank-skip' || github.event_name == 'push' || github.event_name == 'schedule'
+    if: github.event.label.name == 'run-decode' || github.event.label.name == 'blank-skip' || github.event_name == 'push' || github.event_name == 'schedule'
    runs-on: ${{ matrix.os }}
    strategy:
      matrix:
@ -60,7 +60,7 @@ jobs:
        run: |
          grep -v '^#' ./requirements-ci.txt  | xargs -n 1 -L 1 pip install
          pip uninstall -y protobuf
-          pip install --no-binary protobuf protobuf
+          pip install --no-binary protobuf protobuf==3.20.*

      - name: Cache kaldifeat
        id: my-cache
@ -119,7 +119,7 @@ jobs:
          ln -sfv ~/tmp/fbank-libri egs/librispeech/ASR/data/fbank
          ls -lh egs/librispeech/ASR/data/*

-          sudo apt-get -qq install git-lfs tree sox
+          sudo apt-get -qq install git-lfs tree
          export PYTHONPATH=$PWD:$PYTHONPATH
          export PYTHONPATH=~/tmp/kaldifeat/kaldifeat/python:$PYTHONPATH
          export PYTHONPATH=~/tmp/kaldifeat/build/lib:$PYTHONPATH
--- a/.github/workflows/run-librispeech-2022-12-29-stateless7-streaming.yml
+++ b/.github/workflows/run-librispeech-2022-12-29-stateless7-streaming.yml
@ -64,7 +64,7 @@ jobs:
        run: |
          grep -v '^#' ./requirements-ci.txt  | xargs -n 1 -L 1 pip install
          pip uninstall -y protobuf
-          pip install --no-binary protobuf protobuf
+          pip install --no-binary protobuf protobuf==3.20.*

      - name: Cache kaldifeat
        id: my-cache
@ -123,7 +123,7 @@ jobs:
          ln -sfv ~/tmp/fbank-libri egs/librispeech/ASR/data/fbank
          ls -lh egs/librispeech/ASR/data/*

-          sudo apt-get -qq install git-lfs tree sox
+          sudo apt-get -qq install git-lfs tree
          export PYTHONPATH=$PWD:$PYTHONPATH
          export PYTHONPATH=~/tmp/kaldifeat/kaldifeat/python:$PYTHONPATH
          export PYTHONPATH=~/tmp/kaldifeat/build/lib:$PYTHONPATH
--- a/.github/workflows/run-librispeech-conformer-ctc3-2022-11-28.yml
+++ b/.github/workflows/run-librispeech-conformer-ctc3-2022-11-28.yml
@ -64,7 +64,7 @@ jobs:
        run: |
          grep -v '^#' ./requirements-ci.txt  | xargs -n 1 -L 1 pip install
          pip uninstall -y protobuf
-          pip install --no-binary protobuf protobuf
+          pip install --no-binary protobuf protobuf==3.20.*

      - name: Cache kaldifeat
        id: my-cache
@ -123,7 +123,7 @@ jobs:
          ln -sfv ~/tmp/fbank-libri egs/librispeech/ASR/data/fbank
          ls -lh egs/librispeech/ASR/data/*

-          sudo apt-get -qq install git-lfs tree sox
+          sudo apt-get -qq install git-lfs tree
          export PYTHONPATH=$PWD:$PYTHONPATH
          export PYTHONPATH=~/tmp/kaldifeat/kaldifeat/python:$PYTHONPATH
          export PYTHONPATH=~/tmp/kaldifeat/build/lib:$PYTHONPATH
--- a/.github/workflows/run-librispeech-lstm-transducer-stateless2-2022-09-03.yml
+++ b/.github/workflows/run-librispeech-lstm-transducer-stateless2-2022-09-03.yml
@ -47,7 +47,7 @@ jobs:
        run: |
          grep -v '^#' ./requirements-ci.txt  | xargs -n 1 -L 1 pip install
          pip uninstall -y protobuf
-          pip install --no-binary protobuf protobuf
+          pip install --no-binary protobuf protobuf==3.20.*

      - name: Cache kaldifeat
        id: my-cache
@ -106,7 +106,7 @@ jobs:
          ln -sfv ~/tmp/fbank-libri egs/librispeech/ASR/data/fbank
          ls -lh egs/librispeech/ASR/data/*

-          sudo apt-get -qq install git-lfs tree sox
+          sudo apt-get -qq install git-lfs tree
          export PYTHONPATH=$PWD:$PYTHONPATH
          export PYTHONPATH=~/tmp/kaldifeat/kaldifeat/python:$PYTHONPATH
          export PYTHONPATH=~/tmp/kaldifeat/build/lib:$PYTHONPATH
--- a/.github/workflows/run-librispeech-pruned-transducer-stateless3-2022-05-13.yml
+++ b/.github/workflows/run-librispeech-pruned-transducer-stateless3-2022-05-13.yml
@ -64,7 +64,7 @@ jobs:
        run: |
          grep -v '^#' ./requirements-ci.txt  | xargs -n 1 -L 1 pip install
          pip uninstall -y protobuf
-          pip install --no-binary protobuf protobuf
+          pip install --no-binary protobuf protobuf==3.20.*

      - name: Cache kaldifeat
        id: my-cache
@ -123,7 +123,7 @@ jobs:
          ln -sfv ~/tmp/fbank-libri egs/librispeech/ASR/data/fbank
          ls -lh egs/librispeech/ASR/data/*

-          sudo apt-get -qq install git-lfs tree sox
+          sudo apt-get -qq install git-lfs tree
          export PYTHONPATH=$PWD:$PYTHONPATH
          export PYTHONPATH=~/tmp/kaldifeat/kaldifeat/python:$PYTHONPATH
          export PYTHONPATH=~/tmp/kaldifeat/build/lib:$PYTHONPATH
--- a/.github/workflows/run-librispeech-streaming-transducer-stateless2-2022-06-26.yml
+++ b/.github/workflows/run-librispeech-streaming-transducer-stateless2-2022-06-26.yml
@ -64,7 +64,7 @@ jobs:
        run: |
          grep -v '^#' ./requirements-ci.txt  | xargs -n 1 -L 1 pip install
          pip uninstall -y protobuf
-          pip install --no-binary protobuf protobuf
+          pip install --no-binary protobuf protobuf==3.20.*

      - name: Cache kaldifeat
        id: my-cache
@ -123,7 +123,7 @@ jobs:
          ln -sfv ~/tmp/fbank-libri egs/librispeech/ASR/data/fbank
          ls -lh egs/librispeech/ASR/data/*

-          sudo apt-get -qq install git-lfs tree sox
+          sudo apt-get -qq install git-lfs tree
          export PYTHONPATH=$PWD:$PYTHONPATH
          export PYTHONPATH=~/tmp/kaldifeat/kaldifeat/python:$PYTHONPATH
          export PYTHONPATH=~/tmp/kaldifeat/build/lib:$PYTHONPATH
--- a/.github/workflows/run-librispeech-transducer-stateless2-2022-04-19.yml
+++ b/.github/workflows/run-librispeech-transducer-stateless2-2022-04-19.yml
@ -64,7 +64,7 @@ jobs:
        run: |
          grep -v '^#' ./requirements-ci.txt  | xargs -n 1 -L 1 pip install
          pip uninstall -y protobuf
-          pip install --no-binary protobuf protobuf
+          pip install --no-binary protobuf protobuf==3.20.*

      - name: Cache kaldifeat
        id: my-cache
@ -123,7 +123,7 @@ jobs:
          ln -sfv ~/tmp/fbank-libri egs/librispeech/ASR/data/fbank
          ls -lh egs/librispeech/ASR/data/*

-          sudo apt-get -qq install git-lfs tree sox
+          sudo apt-get -qq install git-lfs tree
          export PYTHONPATH=$PWD:$PYTHONPATH
          export PYTHONPATH=~/tmp/kaldifeat/kaldifeat/python:$PYTHONPATH
          export PYTHONPATH=~/tmp/kaldifeat/build/lib:$PYTHONPATH
--- a/.github/workflows/run-pretrained-conformer-ctc.yml
+++ b/.github/workflows/run-pretrained-conformer-ctc.yml
@ -54,7 +54,7 @@ jobs:
        run: |
          grep -v '^#' ./requirements-ci.txt  | xargs -n 1 -L 1 pip install
          pip uninstall -y protobuf
-          pip install --no-binary protobuf protobuf
+          pip install --no-binary protobuf protobuf==3.20.*

      - name: Cache kaldifeat
        id: my-cache
@ -73,7 +73,7 @@ jobs:
      - name: Inference with pre-trained model
        shell: bash
        run: |
-          sudo apt-get -qq install git-lfs tree sox
+          sudo apt-get -qq install git-lfs tree
          export PYTHONPATH=$PWD:$PYTHONPATH
          export PYTHONPATH=~/tmp/kaldifeat/kaldifeat/python:$PYTHONPATH
          export PYTHONPATH=~/tmp/kaldifeat/build/lib:$PYTHONPATH
--- a/.github/workflows/run-pretrained-transducer-stateless-librispeech-100h.yml
+++ b/.github/workflows/run-pretrained-transducer-stateless-librispeech-100h.yml
@ -63,7 +63,7 @@ jobs:
        run: |
          grep -v '^#' ./requirements-ci.txt  | xargs -n 1 -L 1 pip install
          pip uninstall -y protobuf
-          pip install --no-binary protobuf protobuf
+          pip install --no-binary protobuf protobuf==3.20.*

      - name: Cache kaldifeat
        id: my-cache
@ -122,7 +122,7 @@ jobs:
          ln -sfv ~/tmp/fbank-libri egs/librispeech/ASR/data/fbank
          ls -lh egs/librispeech/ASR/data/*

-          sudo apt-get -qq install git-lfs tree sox
+          sudo apt-get -qq install git-lfs tree
          export PYTHONPATH=$PWD:$PYTHONPATH
          export PYTHONPATH=~/tmp/kaldifeat/kaldifeat/python:$PYTHONPATH
          export PYTHONPATH=~/tmp/kaldifeat/build/lib:$PYTHONPATH
--- a/.github/workflows/run-pretrained-transducer-stateless-librispeech-multi-datasets.yml
+++ b/.github/workflows/run-pretrained-transducer-stateless-librispeech-multi-datasets.yml
@ -63,7 +63,7 @@ jobs:
        run: |
          grep -v '^#' ./requirements-ci.txt  | xargs -n 1 -L 1 pip install
          pip uninstall -y protobuf
-          pip install --no-binary protobuf protobuf
+          pip install --no-binary protobuf protobuf==3.20.*

      - name: Cache kaldifeat
        id: my-cache
@ -122,7 +122,7 @@ jobs:
          ln -sfv ~/tmp/fbank-libri egs/librispeech/ASR/data/fbank
          ls -lh egs/librispeech/ASR/data/*

-          sudo apt-get -qq install git-lfs tree sox
+          sudo apt-get -qq install git-lfs tree
          export PYTHONPATH=$PWD:$PYTHONPATH
          export PYTHONPATH=~/tmp/kaldifeat/kaldifeat/python:$PYTHONPATH
          export PYTHONPATH=~/tmp/kaldifeat/build/lib:$PYTHONPATH
--- a/.github/workflows/run-pretrained-transducer-stateless-modified-2-aishell.yml
+++ b/.github/workflows/run-pretrained-transducer-stateless-modified-2-aishell.yml
@ -54,7 +54,7 @@ jobs:
        run: |
          grep -v '^#' ./requirements-ci.txt  | xargs -n 1 -L 1 pip install
          pip uninstall -y protobuf
-          pip install --no-binary protobuf protobuf
+          pip install --no-binary protobuf protobuf==3.20.*

      - name: Cache kaldifeat
        id: my-cache
@ -73,7 +73,7 @@ jobs:
      - name: Inference with pre-trained model
        shell: bash
        run: |
-          sudo apt-get -qq install git-lfs tree sox
+          sudo apt-get -qq install git-lfs tree
          export PYTHONPATH=$PWD:$PYTHONPATH
          export PYTHONPATH=~/tmp/kaldifeat/kaldifeat/python:$PYTHONPATH
          export PYTHONPATH=~/tmp/kaldifeat/build/lib:$PYTHONPATH
--- a/.github/workflows/run-pretrained-transducer-stateless-modified-aishell.yml
+++ b/.github/workflows/run-pretrained-transducer-stateless-modified-aishell.yml
@ -54,7 +54,7 @@ jobs:
        run: |
          grep -v '^#' ./requirements-ci.txt  | xargs -n 1 -L 1 pip install
          pip uninstall -y protobuf
-          pip install --no-binary protobuf protobuf
+          pip install --no-binary protobuf protobuf==3.20.*

      - name: Cache kaldifeat
        id: my-cache
@ -73,7 +73,7 @@ jobs:
      - name: Inference with pre-trained model
        shell: bash
        run: |
-          sudo apt-get -qq install git-lfs tree sox
+          sudo apt-get -qq install git-lfs tree
          export PYTHONPATH=$PWD:$PYTHONPATH
          export PYTHONPATH=~/tmp/kaldifeat/kaldifeat/python:$PYTHONPATH
          export PYTHONPATH=~/tmp/kaldifeat/build/lib:$PYTHONPATH
--- a/.github/workflows/run-pretrained-transducer-stateless.yml
+++ b/.github/workflows/run-pretrained-transducer-stateless.yml
@ -63,7 +63,7 @@ jobs:
        run: |
          grep -v '^#' ./requirements-ci.txt  | xargs -n 1 -L 1 pip install
          pip uninstall -y protobuf
-          pip install --no-binary protobuf protobuf
+          pip install --no-binary protobuf protobuf==3.20.*

      - name: Cache kaldifeat
        id: my-cache
@ -122,7 +122,7 @@ jobs:
          ln -sfv ~/tmp/fbank-libri egs/librispeech/ASR/data/fbank
          ls -lh egs/librispeech/ASR/data/*

-          sudo apt-get -qq install git-lfs tree sox
+          sudo apt-get -qq install git-lfs tree
          export PYTHONPATH=$PWD:$PYTHONPATH
          export PYTHONPATH=~/tmp/kaldifeat/kaldifeat/python:$PYTHONPATH
          export PYTHONPATH=~/tmp/kaldifeat/build/lib:$PYTHONPATH
--- a/.github/workflows/run-pretrained-transducer.yml
+++ b/.github/workflows/run-pretrained-transducer.yml
@ -54,7 +54,7 @@ jobs:
        run: |
          grep -v '^#' ./requirements-ci.txt  | xargs -n 1 -L 1 pip install
          pip uninstall -y protobuf
-          pip install --no-binary protobuf protobuf
+          pip install --no-binary protobuf protobuf==3.20.*

      - name: Cache kaldifeat
        id: my-cache
@ -73,7 +73,7 @@ jobs:
      - name: Inference with pre-trained model
        shell: bash
        run: |
-          sudo apt-get -qq install git-lfs tree sox
+          sudo apt-get -qq install git-lfs tree
          export PYTHONPATH=$PWD:$PYTHONPATH
          export PYTHONPATH=~/tmp/kaldifeat/kaldifeat/python:$PYTHONPATH
          export PYTHONPATH=~/tmp/kaldifeat/build/lib:$PYTHONPATH
--- a/.github/workflows/run-ptb-rnn-lm.yml
+++ b/.github/workflows/run-ptb-rnn-lm.yml
@ -47,7 +47,7 @@ jobs:
        run: |
          grep -v '^#' ./requirements-ci.txt  | grep -v kaldifst | xargs -n 1 -L 1 pip install
          pip uninstall -y protobuf
-          pip install --no-binary protobuf protobuf
+          pip install --no-binary protobuf protobuf==3.20.*

      - name: Prepare data
        shell: bash
--- a/.github/workflows/run-wenetspeech-pruned-transducer-stateless2.yml
+++ b/.github/workflows/run-wenetspeech-pruned-transducer-stateless2.yml
@ -54,7 +54,7 @@ jobs:
        run: |
          grep -v '^#' ./requirements-ci.txt  | xargs -n 1 -L 1 pip install
          pip uninstall -y protobuf
-          pip install --no-binary protobuf protobuf
+          pip install --no-binary protobuf protobuf==3.20.*

      - name: Cache kaldifeat
        id: my-cache
@ -76,7 +76,7 @@ jobs:
          GITHUB_EVENT_NAME: ${{ github.event_name }}
          GITHUB_EVENT_LABEL_NAME: ${{ github.event.label.name }}
        run: |
-          sudo apt-get -qq install git-lfs tree sox
+          sudo apt-get -qq install git-lfs tree
          export PYTHONPATH=$PWD:$PYTHONPATH
          export PYTHONPATH=~/tmp/kaldifeat/kaldifeat/python:$PYTHONPATH
          export PYTHONPATH=~/tmp/kaldifeat/build/lib:$PYTHONPATH
--- a/.github/workflows/run-yesno-recipe.yml
+++ b/.github/workflows/run-yesno-recipe.yml
@ -67,7 +67,7 @@ jobs:
        run: |
          grep -v '^#' ./requirements-ci.txt  | grep -v kaldifst | xargs -n 1 -L 1 pip install
          pip uninstall -y protobuf
-          pip install --no-binary protobuf protobuf
+          pip install --no-binary protobuf protobuf==3.20.*

      - name: Run yesno recipe
        shell: bash
--- a/.github/workflows/test-ncnn-export.yml
+++ b/.github/workflows/test-ncnn-export.yml
@ -46,7 +46,7 @@ jobs:
        run: |
          grep -v '^#' ./requirements-ci.txt  | xargs -n 1 -L 1 pip install
          pip uninstall -y protobuf
-          pip install --no-binary protobuf protobuf
+          pip install --no-binary protobuf protobuf==3.20.*

      - name: Cache kaldifeat
        id: my-cache
--- a/.github/workflows/test-onnx-export.yml
+++ b/.github/workflows/test-onnx-export.yml
@ -46,7 +46,7 @@ jobs:
        run: |
          grep -v '^#' ./requirements-ci.txt  | xargs -n 1 -L 1 pip install
          pip uninstall -y protobuf
-          pip install --no-binary protobuf protobuf
+          pip install --no-binary protobuf protobuf==3.20.*

      - name: Cache kaldifeat
        id: my-cache
--- a/.github/workflows/test.yml
+++ b/.github/workflows/test.yml
@ -56,7 +56,7 @@ jobs:
        run: |
          sudo apt update
          sudo apt install -q -y libsndfile1-dev libsndfile1 ffmpeg
-          sudo apt install -q -y --fix-missing sox libsox-dev libsox-fmt-all
+          sudo apt install -q -y --fix-missing libsox-dev libsox-fmt-all

      - name: Install Python dependencies
        run: |
@ -70,7 +70,7 @@ jobs:
          pip install git+https://github.com/lhotse-speech/lhotse
          # icefall requirements
          pip uninstall -y protobuf
-          pip install --no-binary protobuf protobuf
+          pip install --no-binary protobuf protobuf==3.20.*

          pip install kaldifst
          pip install onnxruntime
--- a/egs/librispeech/ASR/local/compute_fbank_librispeech.py
+++ b/egs/librispeech/ASR/local/compute_fbank_librispeech.py
@ -54,10 +54,20 @@ def get_args():
        help="""Path to the bpe.model. If not None, we will remove short and
        long utterances before extracting features""",
    )
+
+    parser.add_argument(
+        "--dataset",
+        type=str,
+        help="""Dataset parts to compute fbank. If None, we will use all""",
+    )
+
    return parser.parse_args()


-def compute_fbank_librispeech(bpe_model: Optional[str] = None):
+def compute_fbank_librispeech(
+    bpe_model: Optional[str] = None,
+    dataset: Optional[str] = None,
+):
    src_dir = Path("data/manifests")
    output_dir = Path("data/fbank")
    num_jobs = min(15, os.cpu_count())
@ -68,15 +78,19 @@ def compute_fbank_librispeech(bpe_model: Optional[str] = None):
        sp = spm.SentencePieceProcessor()
        sp.load(bpe_model)

-    dataset_parts = (
-        "dev-clean",
-        "dev-other",
-        "test-clean",
-        "test-other",
-        "train-clean-100",
-        "train-clean-360",
-        "train-other-500",
-    )
+    if dataset is None:
+        dataset_parts = (
+            "dev-clean",
+            "dev-other",
+            "test-clean",
+            "test-other",
+            "train-clean-100",
+            "train-clean-360",
+            "train-other-500",
+        )
+    else:
+        dataset_parts = dataset.split(" ", -1)
+
    prefix = "librispeech"
    suffix = "jsonl.gz"
    manifests = read_manifests_if_cached(
@ -131,4 +145,4 @@ if __name__ == "__main__":
    logging.basicConfig(format=formatter, level=logging.INFO)
    args = get_args()
    logging.info(vars(args))
-    compute_fbank_librispeech(bpe_model=args.bpe_model)
+    compute_fbank_librispeech(bpe_model=args.bpe_model, dataset=args.dataset)