mirror of
https://github.com/k2-fsa/icefall.git
synced 2025-08-09 01:52:41 +00:00
Merge branch 'k2-fsa:master' into Blank-regularized-CTC
This commit is contained in:
Yifan Yang
GitHub
commit
28c3d43d74
@ -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/
|
||||
|
||||
@ -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 $repo/exp
|
||||
|
||||
@ -18,7 +18,6 @@ repo=$(basename $repo_url)
|
||||
|
||||
log "Display test files"
|
||||
tree $repo/
|
||||
soxi $repo/test_wavs/*.wav
|
||||
ls -lh $repo/test_wavs/*.wav
|
||||
|
||||
pushd $repo/exp
|
||||
|
||||
@ -20,7 +20,6 @@ abs_repo=$(realpath $repo)
|
||||
|
||||
log "Display test files"
|
||||
tree $repo/
|
||||
soxi $repo/test_wavs/*.wav
|
||||
ls -lh $repo/test_wavs/*.wav
|
||||
|
||||
pushd $repo/exp
|
||||
|
||||
@ -19,7 +19,6 @@ repo=$(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
|
||||
|
||||
@ -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
|
||||
|
||||
@ -22,7 +22,6 @@ popd
|
||||
|
||||
log "Display test files"
|
||||
tree $repo/
|
||||
soxi $repo/test_wavs/*.wav
|
||||
ls -lh $repo/test_wavs/*.wav
|
||||
|
||||
pushd $repo/exp
|
||||
|
||||
@ -19,7 +19,6 @@ repo=$(basename $repo_url)
|
||||
|
||||
log "Display test files"
|
||||
tree $repo/
|
||||
soxi $repo/test_wavs/*.wav
|
||||
ls -lh $repo/test_wavs/*.wav
|
||||
|
||||
pushd $repo/exp
|
||||
|
||||
@ -19,7 +19,6 @@ repo=$(basename $repo_url)
|
||||
|
||||
log "Display test files"
|
||||
tree $repo/
|
||||
soxi $repo/test_wavs/*.wav
|
||||
ls -lh $repo/test_wavs/*.wav
|
||||
|
||||
pushd $repo/exp
|
||||
|
||||
@ -19,7 +19,6 @@ repo=$(basename $repo_url)
|
||||
|
||||
log "Display test files"
|
||||
tree $repo/
|
||||
soxi $repo/test_wavs/*.wav
|
||||
ls -lh $repo/test_wavs/*.wav
|
||||
|
||||
pushd $repo/exp
|
||||
|
||||
@ -18,7 +18,6 @@ repo=$(basename $repo_url)
|
||||
|
||||
log "Display test files"
|
||||
tree $repo/
|
||||
soxi $repo/test_wavs/*.wav
|
||||
ls -lh $repo/test_wavs/*.wav
|
||||
|
||||
pushd $repo/exp
|
||||
|
||||
@ -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 $repo/exp
|
||||
|
||||
@ -19,7 +19,6 @@ repo=$(basename $repo_url)
|
||||
|
||||
log "Display test files"
|
||||
tree $repo/
|
||||
soxi $repo/test_wavs/*.wav
|
||||
ls -lh $repo/test_wavs/*.wav
|
||||
|
||||
pushd $repo
|
||||
|
||||
@ -19,7 +19,6 @@ repo=$(basename $repo_url)
|
||||
|
||||
log "Display test files"
|
||||
tree $repo/
|
||||
soxi $repo/test_wavs/*.wav
|
||||
ls -lh $repo/test_wavs/*.wav
|
||||
|
||||
pushd $repo/exp
|
||||
|
||||
@ -19,7 +19,6 @@ repo=$(basename $repo_url)
|
||||
|
||||
log "Display test files"
|
||||
tree $repo/
|
||||
soxi $repo/test_wavs/*.wav
|
||||
ls -lh $repo/test_wavs/*.wav
|
||||
|
||||
pushd $repo/exp
|
||||
|
||||
@ -19,7 +19,6 @@ repo=$(basename $repo_url)
|
||||
|
||||
log "Display test files"
|
||||
tree $repo/
|
||||
soxi $repo/test_wavs/*.wav
|
||||
ls -lh $repo/test_wavs/*.wav
|
||||
|
||||
for sym in 1 2 3; do
|
||||
|
||||
@ -18,7 +18,6 @@ repo=$(basename $repo_url)
|
||||
|
||||
log "Display test files"
|
||||
tree $repo/
|
||||
soxi $repo/test_wavs/*.wav
|
||||
ls -lh $repo/test_wavs/*.wav
|
||||
|
||||
pushd $repo/exp
|
||||
|
||||
@ -19,7 +19,6 @@ repo=$(basename $repo_url)
|
||||
|
||||
log "Display test files"
|
||||
tree $repo/
|
||||
soxi $repo/test_wavs/*.flac
|
||||
ls -lh $repo/test_wavs/*.flac
|
||||
|
||||
log "CTC decoding"
|
||||
|
||||
@ -19,7 +19,6 @@ repo=$(basename $repo_url)
|
||||
|
||||
log "Display test files"
|
||||
tree $repo/
|
||||
soxi $repo/test_wavs/*.wav
|
||||
ls -lh $repo/test_wavs/*.wav
|
||||
|
||||
for sym in 1 2 3; do
|
||||
|
||||
@ -19,7 +19,6 @@ repo=$(basename $repo_url)
|
||||
|
||||
log "Display test files"
|
||||
tree $repo/
|
||||
soxi $repo/test_wavs/*.wav
|
||||
ls -lh $repo/test_wavs/*.wav
|
||||
|
||||
for sym in 1 2 3; do
|
||||
|
||||
@ -19,7 +19,6 @@ repo=$(basename $repo_url)
|
||||
|
||||
log "Display test files"
|
||||
tree $repo/
|
||||
soxi $repo/test_wavs/*.wav
|
||||
ls -lh $repo/test_wavs/*.wav
|
||||
|
||||
for sym in 1 2 3; do
|
||||
|
||||
@ -19,7 +19,6 @@ repo=$(basename $repo_url)
|
||||
|
||||
log "Display test files"
|
||||
tree $repo/
|
||||
soxi $repo/test_wavs/*.wav
|
||||
ls -lh $repo/test_wavs/*.wav
|
||||
|
||||
for sym in 1 2 3; do
|
||||
|
||||
@ -19,7 +19,6 @@ repo=$(basename $repo_url)
|
||||
|
||||
log "Display test files"
|
||||
tree $repo/
|
||||
soxi $repo/test_wavs/*.wav
|
||||
ls -lh $repo/test_wavs/*.wav
|
||||
|
||||
for sym in 1 2 3; do
|
||||
|
||||
@ -19,7 +19,6 @@ repo=$(basename $repo_url)
|
||||
|
||||
log "Display test files"
|
||||
tree $repo/
|
||||
soxi $repo/test_wavs/*.wav
|
||||
ls -lh $repo/test_wavs/*.wav
|
||||
|
||||
log "Beam search decoding"
|
||||
|
||||
@ -20,7 +20,6 @@ repo=$(basename $repo_url)
|
||||
|
||||
log "Display test files"
|
||||
tree $repo/
|
||||
soxi $repo/test_wavs/*.wav
|
||||
ls -lh $repo/test_wavs/*.wav
|
||||
|
||||
pushd $repo/exp
|
||||
|
||||
67
.github/scripts/test-ncnn-export.sh
vendored
67
.github/scripts/test-ncnn-export.sh
vendored
@ -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 "--------------------------------------------------------------------------"
|
||||
|
||||
4
.github/workflows/run-aishell-2022-06-20.yml
vendored
4
.github/workflows/run-aishell-2022-06-20.yml
vendored
@ -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:$PYTHONPATH
|
||||
|
||||
@ -64,7 +64,7 @@ jobs:
|
||||
run: |
|
||||
grep -v '^#' ./requirements-ci.txt | xargs -n 1 -L 1 pip install
|
||||
pip uninstall -y protobuf
|
||||
pip install --no-binary protobuf protobuf
|
||||
pip install --no-binary protobuf protobuf==3.20.*
|
||||
|
||||
- name: Cache kaldifeat
|
||||
id: my-cache
|
||||
|
||||
@ -64,7 +64,7 @@ jobs:
|
||||
run: |
|
||||
grep -v '^#' ./requirements-ci.txt | xargs -n 1 -L 1 pip install
|
||||
pip uninstall -y protobuf
|
||||
pip install --no-binary protobuf protobuf
|
||||
pip install --no-binary protobuf protobuf==3.20.*
|
||||
|
||||
- name: Cache kaldifeat
|
||||
id: my-cache
|
||||
@ -123,7 +123,7 @@ jobs:
|
||||
ln -sfv ~/tmp/fbank-libri egs/librispeech/ASR/data/fbank
|
||||
ls -lh egs/librispeech/ASR/data/*
|
||||
|
||||
sudo apt-get -qq install git-lfs tree sox
|
||||
sudo apt-get -qq install git-lfs tree
|
||||
export PYTHONPATH=$PWD:$PYTHONPATH
|
||||
export PYTHONPATH=~/tmp/kaldifeat/kaldifeat/python:$PYTHONPATH
|
||||
export PYTHONPATH=~/tmp/kaldifeat/build/lib:$PYTHONPATH
|
||||
|
||||
@ -64,7 +64,7 @@ jobs:
|
||||
run: |
|
||||
grep -v '^#' ./requirements-ci.txt | xargs -n 1 -L 1 pip install
|
||||
pip uninstall -y protobuf
|
||||
pip install --no-binary protobuf protobuf
|
||||
pip install --no-binary protobuf protobuf==3.20.*
|
||||
|
||||
- name: Cache kaldifeat
|
||||
id: my-cache
|
||||
@ -123,7 +123,7 @@ jobs:
|
||||
ln -sfv ~/tmp/fbank-libri egs/librispeech/ASR/data/fbank
|
||||
ls -lh egs/librispeech/ASR/data/*
|
||||
|
||||
sudo apt-get -qq install git-lfs tree sox
|
||||
sudo apt-get -qq install git-lfs tree
|
||||
export PYTHONPATH=$PWD:$PYTHONPATH
|
||||
export PYTHONPATH=~/tmp/kaldifeat/kaldifeat/python:$PYTHONPATH
|
||||
export PYTHONPATH=~/tmp/kaldifeat/build/lib:$PYTHONPATH
|
||||
|
||||
@ -64,7 +64,7 @@ jobs:
|
||||
run: |
|
||||
grep -v '^#' ./requirements-ci.txt | xargs -n 1 -L 1 pip install
|
||||
pip uninstall -y protobuf
|
||||
pip install --no-binary protobuf protobuf
|
||||
pip install --no-binary protobuf protobuf==3.20.*
|
||||
|
||||
- name: Cache kaldifeat
|
||||
id: my-cache
|
||||
@ -123,7 +123,7 @@ jobs:
|
||||
ln -sfv ~/tmp/fbank-libri egs/librispeech/ASR/data/fbank
|
||||
ls -lh egs/librispeech/ASR/data/*
|
||||
|
||||
sudo apt-get -qq install git-lfs tree sox
|
||||
sudo apt-get -qq install git-lfs tree
|
||||
export PYTHONPATH=$PWD:$PYTHONPATH
|
||||
export PYTHONPATH=~/tmp/kaldifeat/kaldifeat/python:$PYTHONPATH
|
||||
export PYTHONPATH=~/tmp/kaldifeat/build/lib:$PYTHONPATH
|
||||
|
||||
@ -64,7 +64,7 @@ jobs:
|
||||
run: |
|
||||
grep -v '^#' ./requirements-ci.txt | xargs -n 1 -L 1 pip install
|
||||
pip uninstall -y protobuf
|
||||
pip install --no-binary protobuf protobuf
|
||||
pip install --no-binary protobuf protobuf==3.20.*
|
||||
|
||||
- name: Cache kaldifeat
|
||||
id: my-cache
|
||||
@ -123,7 +123,7 @@ jobs:
|
||||
ln -sfv ~/tmp/fbank-libri egs/librispeech/ASR/data/fbank
|
||||
ls -lh egs/librispeech/ASR/data/*
|
||||
|
||||
sudo apt-get -qq install git-lfs tree sox
|
||||
sudo apt-get -qq install git-lfs tree
|
||||
export PYTHONPATH=$PWD:$PYTHONPATH
|
||||
export PYTHONPATH=~/tmp/kaldifeat/kaldifeat/python:$PYTHONPATH
|
||||
export PYTHONPATH=~/tmp/kaldifeat/build/lib:$PYTHONPATH
|
||||
|
||||
@ -64,7 +64,7 @@ jobs:
|
||||
run: |
|
||||
grep -v '^#' ./requirements-ci.txt | xargs -n 1 -L 1 pip install
|
||||
pip uninstall -y protobuf
|
||||
pip install --no-binary protobuf protobuf
|
||||
pip install --no-binary protobuf protobuf==3.20.*
|
||||
|
||||
- name: Cache kaldifeat
|
||||
id: my-cache
|
||||
@ -123,7 +123,7 @@ jobs:
|
||||
ln -sfv ~/tmp/fbank-libri egs/librispeech/ASR/data/fbank
|
||||
ls -lh egs/librispeech/ASR/data/*
|
||||
|
||||
sudo apt-get -qq install git-lfs tree sox
|
||||
sudo apt-get -qq install git-lfs tree
|
||||
export PYTHONPATH=$PWD:$PYTHONPATH
|
||||
export PYTHONPATH=~/tmp/kaldifeat/kaldifeat/python:$PYTHONPATH
|
||||
export PYTHONPATH=~/tmp/kaldifeat/build/lib:$PYTHONPATH
|
||||
|
||||
@ -60,7 +60,7 @@ jobs:
|
||||
run: |
|
||||
grep -v '^#' ./requirements-ci.txt | xargs -n 1 -L 1 pip install
|
||||
pip uninstall -y protobuf
|
||||
pip install --no-binary protobuf protobuf
|
||||
pip install --no-binary protobuf protobuf==3.20.*
|
||||
|
||||
- name: Cache kaldifeat
|
||||
id: my-cache
|
||||
@ -119,7 +119,7 @@ jobs:
|
||||
ln -sfv ~/tmp/fbank-libri egs/librispeech/ASR/data/fbank
|
||||
ls -lh egs/librispeech/ASR/data/*
|
||||
|
||||
sudo apt-get -qq install git-lfs tree sox
|
||||
sudo apt-get -qq install git-lfs tree
|
||||
export PYTHONPATH=$PWD:$PYTHONPATH
|
||||
export PYTHONPATH=~/tmp/kaldifeat/kaldifeat/python:$PYTHONPATH
|
||||
export PYTHONPATH=~/tmp/kaldifeat/build/lib:$PYTHONPATH
|
||||
|
||||
@ -64,7 +64,7 @@ jobs:
|
||||
run: |
|
||||
grep -v '^#' ./requirements-ci.txt | xargs -n 1 -L 1 pip install
|
||||
pip uninstall -y protobuf
|
||||
pip install --no-binary protobuf protobuf
|
||||
pip install --no-binary protobuf protobuf==3.20.*
|
||||
|
||||
- name: Cache kaldifeat
|
||||
id: my-cache
|
||||
@ -123,7 +123,7 @@ jobs:
|
||||
ln -sfv ~/tmp/fbank-libri egs/librispeech/ASR/data/fbank
|
||||
ls -lh egs/librispeech/ASR/data/*
|
||||
|
||||
sudo apt-get -qq install git-lfs tree sox
|
||||
sudo apt-get -qq install git-lfs tree
|
||||
export PYTHONPATH=$PWD:$PYTHONPATH
|
||||
export PYTHONPATH=~/tmp/kaldifeat/kaldifeat/python:$PYTHONPATH
|
||||
export PYTHONPATH=~/tmp/kaldifeat/build/lib:$PYTHONPATH
|
||||
|
||||
@ -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:$PYTHONPATH
|
||||
|
||||
@ -64,7 +64,7 @@ jobs:
|
||||
run: |
|
||||
grep -v '^#' ./requirements-ci.txt | xargs -n 1 -L 1 pip install
|
||||
pip uninstall -y protobuf
|
||||
pip install --no-binary protobuf protobuf
|
||||
pip install --no-binary protobuf protobuf==3.20.*
|
||||
|
||||
- name: Cache kaldifeat
|
||||
id: my-cache
|
||||
@ -123,7 +123,7 @@ jobs:
|
||||
ln -sfv ~/tmp/fbank-libri egs/librispeech/ASR/data/fbank
|
||||
ls -lh egs/librispeech/ASR/data/*
|
||||
|
||||
sudo apt-get -qq install git-lfs tree sox
|
||||
sudo apt-get -qq install git-lfs tree
|
||||
export PYTHONPATH=$PWD:$PYTHONPATH
|
||||
export PYTHONPATH=~/tmp/kaldifeat/kaldifeat/python:$PYTHONPATH
|
||||
export PYTHONPATH=~/tmp/kaldifeat/build/lib:$PYTHONPATH
|
||||
|
||||
@ -64,7 +64,7 @@ jobs:
|
||||
run: |
|
||||
grep -v '^#' ./requirements-ci.txt | xargs -n 1 -L 1 pip install
|
||||
pip uninstall -y protobuf
|
||||
pip install --no-binary protobuf protobuf
|
||||
pip install --no-binary protobuf protobuf==3.20.*
|
||||
|
||||
- name: Cache kaldifeat
|
||||
id: my-cache
|
||||
@ -123,7 +123,7 @@ jobs:
|
||||
ln -sfv ~/tmp/fbank-libri egs/librispeech/ASR/data/fbank
|
||||
ls -lh egs/librispeech/ASR/data/*
|
||||
|
||||
sudo apt-get -qq install git-lfs tree sox
|
||||
sudo apt-get -qq install git-lfs tree
|
||||
export PYTHONPATH=$PWD:$PYTHONPATH
|
||||
export PYTHONPATH=~/tmp/kaldifeat/kaldifeat/python:$PYTHONPATH
|
||||
export PYTHONPATH=~/tmp/kaldifeat/build/lib:$PYTHONPATH
|
||||
|
||||
@ -47,7 +47,7 @@ jobs:
|
||||
run: |
|
||||
grep -v '^#' ./requirements-ci.txt | xargs -n 1 -L 1 pip install
|
||||
pip uninstall -y protobuf
|
||||
pip install --no-binary protobuf protobuf
|
||||
pip install --no-binary protobuf protobuf==3.20.*
|
||||
|
||||
- name: Cache kaldifeat
|
||||
id: my-cache
|
||||
@ -106,7 +106,7 @@ jobs:
|
||||
ln -sfv ~/tmp/fbank-libri egs/librispeech/ASR/data/fbank
|
||||
ls -lh egs/librispeech/ASR/data/*
|
||||
|
||||
sudo apt-get -qq install git-lfs tree sox
|
||||
sudo apt-get -qq install git-lfs tree
|
||||
export PYTHONPATH=$PWD:$PYTHONPATH
|
||||
export PYTHONPATH=~/tmp/kaldifeat/kaldifeat/python:$PYTHONPATH
|
||||
export PYTHONPATH=~/tmp/kaldifeat/build/lib:$PYTHONPATH
|
||||
|
||||
@ -64,7 +64,7 @@ jobs:
|
||||
run: |
|
||||
grep -v '^#' ./requirements-ci.txt | xargs -n 1 -L 1 pip install
|
||||
pip uninstall -y protobuf
|
||||
pip install --no-binary protobuf protobuf
|
||||
pip install --no-binary protobuf protobuf==3.20.*
|
||||
|
||||
- name: Cache kaldifeat
|
||||
id: my-cache
|
||||
@ -123,7 +123,7 @@ jobs:
|
||||
ln -sfv ~/tmp/fbank-libri egs/librispeech/ASR/data/fbank
|
||||
ls -lh egs/librispeech/ASR/data/*
|
||||
|
||||
sudo apt-get -qq install git-lfs tree sox
|
||||
sudo apt-get -qq install git-lfs tree
|
||||
export PYTHONPATH=$PWD:$PYTHONPATH
|
||||
export PYTHONPATH=~/tmp/kaldifeat/kaldifeat/python:$PYTHONPATH
|
||||
export PYTHONPATH=~/tmp/kaldifeat/build/lib:$PYTHONPATH
|
||||
|
||||
@ -64,7 +64,7 @@ jobs:
|
||||
run: |
|
||||
grep -v '^#' ./requirements-ci.txt | xargs -n 1 -L 1 pip install
|
||||
pip uninstall -y protobuf
|
||||
pip install --no-binary protobuf protobuf
|
||||
pip install --no-binary protobuf protobuf==3.20.*
|
||||
|
||||
- name: Cache kaldifeat
|
||||
id: my-cache
|
||||
@ -123,7 +123,7 @@ jobs:
|
||||
ln -sfv ~/tmp/fbank-libri egs/librispeech/ASR/data/fbank
|
||||
ls -lh egs/librispeech/ASR/data/*
|
||||
|
||||
sudo apt-get -qq install git-lfs tree sox
|
||||
sudo apt-get -qq install git-lfs tree
|
||||
export PYTHONPATH=$PWD:$PYTHONPATH
|
||||
export PYTHONPATH=~/tmp/kaldifeat/kaldifeat/python:$PYTHONPATH
|
||||
export PYTHONPATH=~/tmp/kaldifeat/build/lib:$PYTHONPATH
|
||||
|
||||
@ -64,7 +64,7 @@ jobs:
|
||||
run: |
|
||||
grep -v '^#' ./requirements-ci.txt | xargs -n 1 -L 1 pip install
|
||||
pip uninstall -y protobuf
|
||||
pip install --no-binary protobuf protobuf
|
||||
pip install --no-binary protobuf protobuf==3.20.*
|
||||
|
||||
- name: Cache kaldifeat
|
||||
id: my-cache
|
||||
@ -123,7 +123,7 @@ jobs:
|
||||
ln -sfv ~/tmp/fbank-libri egs/librispeech/ASR/data/fbank
|
||||
ls -lh egs/librispeech/ASR/data/*
|
||||
|
||||
sudo apt-get -qq install git-lfs tree sox
|
||||
sudo apt-get -qq install git-lfs tree
|
||||
export PYTHONPATH=$PWD:$PYTHONPATH
|
||||
export PYTHONPATH=~/tmp/kaldifeat/kaldifeat/python:$PYTHONPATH
|
||||
export PYTHONPATH=~/tmp/kaldifeat/build/lib:$PYTHONPATH
|
||||
|
||||
@ -54,7 +54,7 @@ jobs:
|
||||
run: |
|
||||
grep -v '^#' ./requirements-ci.txt | xargs -n 1 -L 1 pip install
|
||||
pip uninstall -y protobuf
|
||||
pip install --no-binary protobuf protobuf
|
||||
pip install --no-binary protobuf protobuf==3.20.*
|
||||
|
||||
- name: Cache kaldifeat
|
||||
id: my-cache
|
||||
@ -73,7 +73,7 @@ jobs:
|
||||
- name: Inference with pre-trained model
|
||||
shell: bash
|
||||
run: |
|
||||
sudo apt-get -qq install git-lfs tree sox
|
||||
sudo apt-get -qq install git-lfs tree
|
||||
export PYTHONPATH=$PWD:$PYTHONPATH
|
||||
export PYTHONPATH=~/tmp/kaldifeat/kaldifeat/python:$PYTHONPATH
|
||||
export PYTHONPATH=~/tmp/kaldifeat/build/lib:$PYTHONPATH
|
||||
|
||||
@ -63,7 +63,7 @@ jobs:
|
||||
run: |
|
||||
grep -v '^#' ./requirements-ci.txt | xargs -n 1 -L 1 pip install
|
||||
pip uninstall -y protobuf
|
||||
pip install --no-binary protobuf protobuf
|
||||
pip install --no-binary protobuf protobuf==3.20.*
|
||||
|
||||
- name: Cache kaldifeat
|
||||
id: my-cache
|
||||
@ -122,7 +122,7 @@ jobs:
|
||||
ln -sfv ~/tmp/fbank-libri egs/librispeech/ASR/data/fbank
|
||||
ls -lh egs/librispeech/ASR/data/*
|
||||
|
||||
sudo apt-get -qq install git-lfs tree sox
|
||||
sudo apt-get -qq install git-lfs tree
|
||||
export PYTHONPATH=$PWD:$PYTHONPATH
|
||||
export PYTHONPATH=~/tmp/kaldifeat/kaldifeat/python:$PYTHONPATH
|
||||
export PYTHONPATH=~/tmp/kaldifeat/build/lib:$PYTHONPATH
|
||||
|
||||
@ -63,7 +63,7 @@ jobs:
|
||||
run: |
|
||||
grep -v '^#' ./requirements-ci.txt | xargs -n 1 -L 1 pip install
|
||||
pip uninstall -y protobuf
|
||||
pip install --no-binary protobuf protobuf
|
||||
pip install --no-binary protobuf protobuf==3.20.*
|
||||
|
||||
- name: Cache kaldifeat
|
||||
id: my-cache
|
||||
@ -122,7 +122,7 @@ jobs:
|
||||
ln -sfv ~/tmp/fbank-libri egs/librispeech/ASR/data/fbank
|
||||
ls -lh egs/librispeech/ASR/data/*
|
||||
|
||||
sudo apt-get -qq install git-lfs tree sox
|
||||
sudo apt-get -qq install git-lfs tree
|
||||
export PYTHONPATH=$PWD:$PYTHONPATH
|
||||
export PYTHONPATH=~/tmp/kaldifeat/kaldifeat/python:$PYTHONPATH
|
||||
export PYTHONPATH=~/tmp/kaldifeat/build/lib:$PYTHONPATH
|
||||
|
||||
@ -54,7 +54,7 @@ jobs:
|
||||
run: |
|
||||
grep -v '^#' ./requirements-ci.txt | xargs -n 1 -L 1 pip install
|
||||
pip uninstall -y protobuf
|
||||
pip install --no-binary protobuf protobuf
|
||||
pip install --no-binary protobuf protobuf==3.20.*
|
||||
|
||||
- name: Cache kaldifeat
|
||||
id: my-cache
|
||||
@ -73,7 +73,7 @@ jobs:
|
||||
- name: Inference with pre-trained model
|
||||
shell: bash
|
||||
run: |
|
||||
sudo apt-get -qq install git-lfs tree sox
|
||||
sudo apt-get -qq install git-lfs tree
|
||||
export PYTHONPATH=$PWD:$PYTHONPATH
|
||||
export PYTHONPATH=~/tmp/kaldifeat/kaldifeat/python:$PYTHONPATH
|
||||
export PYTHONPATH=~/tmp/kaldifeat/build/lib:$PYTHONPATH
|
||||
|
||||
@ -54,7 +54,7 @@ jobs:
|
||||
run: |
|
||||
grep -v '^#' ./requirements-ci.txt | xargs -n 1 -L 1 pip install
|
||||
pip uninstall -y protobuf
|
||||
pip install --no-binary protobuf protobuf
|
||||
pip install --no-binary protobuf protobuf==3.20.*
|
||||
|
||||
- name: Cache kaldifeat
|
||||
id: my-cache
|
||||
@ -73,7 +73,7 @@ jobs:
|
||||
- name: Inference with pre-trained model
|
||||
shell: bash
|
||||
run: |
|
||||
sudo apt-get -qq install git-lfs tree sox
|
||||
sudo apt-get -qq install git-lfs tree
|
||||
export PYTHONPATH=$PWD:$PYTHONPATH
|
||||
export PYTHONPATH=~/tmp/kaldifeat/kaldifeat/python:$PYTHONPATH
|
||||
export PYTHONPATH=~/tmp/kaldifeat/build/lib:$PYTHONPATH
|
||||
|
||||
@ -63,7 +63,7 @@ jobs:
|
||||
run: |
|
||||
grep -v '^#' ./requirements-ci.txt | xargs -n 1 -L 1 pip install
|
||||
pip uninstall -y protobuf
|
||||
pip install --no-binary protobuf protobuf
|
||||
pip install --no-binary protobuf protobuf==3.20.*
|
||||
|
||||
- name: Cache kaldifeat
|
||||
id: my-cache
|
||||
@ -122,7 +122,7 @@ jobs:
|
||||
ln -sfv ~/tmp/fbank-libri egs/librispeech/ASR/data/fbank
|
||||
ls -lh egs/librispeech/ASR/data/*
|
||||
|
||||
sudo apt-get -qq install git-lfs tree sox
|
||||
sudo apt-get -qq install git-lfs tree
|
||||
export PYTHONPATH=$PWD:$PYTHONPATH
|
||||
export PYTHONPATH=~/tmp/kaldifeat/kaldifeat/python:$PYTHONPATH
|
||||
export PYTHONPATH=~/tmp/kaldifeat/build/lib:$PYTHONPATH
|
||||
|
||||
@ -54,7 +54,7 @@ jobs:
|
||||
run: |
|
||||
grep -v '^#' ./requirements-ci.txt | xargs -n 1 -L 1 pip install
|
||||
pip uninstall -y protobuf
|
||||
pip install --no-binary protobuf protobuf
|
||||
pip install --no-binary protobuf protobuf==3.20.*
|
||||
|
||||
- name: Cache kaldifeat
|
||||
id: my-cache
|
||||
@ -73,7 +73,7 @@ jobs:
|
||||
- name: Inference with pre-trained model
|
||||
shell: bash
|
||||
run: |
|
||||
sudo apt-get -qq install git-lfs tree sox
|
||||
sudo apt-get -qq install git-lfs tree
|
||||
export PYTHONPATH=$PWD:$PYTHONPATH
|
||||
export PYTHONPATH=~/tmp/kaldifeat/kaldifeat/python:$PYTHONPATH
|
||||
export PYTHONPATH=~/tmp/kaldifeat/build/lib:$PYTHONPATH
|
||||
|
||||
2
.github/workflows/run-ptb-rnn-lm.yml
vendored
2
.github/workflows/run-ptb-rnn-lm.yml
vendored
@ -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
|
||||
|
||||
@ -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
|
||||
|
||||
2
.github/workflows/test-ncnn-export.yml
vendored
2
.github/workflows/test-ncnn-export.yml
vendored
@ -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
|
||||
|
||||
2
.github/workflows/test-onnx-export.yml
vendored
2
.github/workflows/test-onnx-export.yml
vendored
@ -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
|
||||
|
||||
4
.github/workflows/test.yml
vendored
4
.github/workflows/test.yml
vendored
@ -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
|
||||
|
||||
@ -391,18 +391,14 @@ def save_results(
|
||||
):
|
||||
test_set_wers = dict()
|
||||
for key, results in results_dict.items():
|
||||
recog_path = (
|
||||
params.res_dir / f"recogs-{test_set_name}-{key}-{params.suffix}.txt"
|
||||
)
|
||||
recog_path = params.res_dir / f"recogs-{test_set_name}-{params.suffix}.txt"
|
||||
results = sorted(results)
|
||||
store_transcripts(filename=recog_path, texts=results)
|
||||
logging.info(f"The transcripts are stored in {recog_path}")
|
||||
|
||||
# The following prints out WERs, per-word error statistics and aligned
|
||||
# ref/hyp pairs.
|
||||
errs_filename = (
|
||||
params.res_dir / f"errs-{test_set_name}-{key}-{params.suffix}.txt"
|
||||
)
|
||||
errs_filename = params.res_dir / f"errs-{test_set_name}-{params.suffix}.txt"
|
||||
with open(errs_filename, "w") as f:
|
||||
wer = write_error_stats(
|
||||
f, f"{test_set_name}-{key}", results, enable_log=True
|
||||
@ -412,9 +408,7 @@ def save_results(
|
||||
logging.info("Wrote detailed error stats to {}".format(errs_filename))
|
||||
|
||||
test_set_wers = sorted(test_set_wers.items(), key=lambda x: x[1])
|
||||
errs_info = (
|
||||
params.res_dir / f"wer-summary-{test_set_name}-{key}-{params.suffix}.txt"
|
||||
)
|
||||
errs_info = params.res_dir / f"wer-summary-{test_set_name}-{params.suffix}.txt"
|
||||
with open(errs_info, "w") as f:
|
||||
print("settings\tWER", file=f)
|
||||
for key, val in test_set_wers:
|
||||
|
||||
@ -388,18 +388,14 @@ def save_results(
|
||||
):
|
||||
test_set_wers = dict()
|
||||
for key, results in results_dict.items():
|
||||
recog_path = (
|
||||
params.res_dir / f"recogs-{test_set_name}-{key}-{params.suffix}.txt"
|
||||
)
|
||||
recog_path = params.res_dir / f"recogs-{test_set_name}-{params.suffix}.txt"
|
||||
results = sorted(results)
|
||||
store_transcripts(filename=recog_path, texts=results)
|
||||
logging.info(f"The transcripts are stored in {recog_path}")
|
||||
|
||||
# The following prints out WERs, per-word error statistics and aligned
|
||||
# ref/hyp pairs.
|
||||
errs_filename = (
|
||||
params.res_dir / f"errs-{test_set_name}-{key}-{params.suffix}.txt"
|
||||
)
|
||||
errs_filename = params.res_dir / f"errs-{test_set_name}-{params.suffix}.txt"
|
||||
# we compute CER for aishell dataset.
|
||||
results_char = []
|
||||
for res in results:
|
||||
@ -413,9 +409,7 @@ def save_results(
|
||||
logging.info("Wrote detailed error stats to {}".format(errs_filename))
|
||||
|
||||
test_set_wers = sorted(test_set_wers.items(), key=lambda x: x[1])
|
||||
errs_info = (
|
||||
params.res_dir / f"wer-summary-{test_set_name}-{key}-{params.suffix}.txt"
|
||||
)
|
||||
errs_info = params.res_dir / f"wer-summary-{test_set_name}-{params.suffix}.txt"
|
||||
with open(errs_info, "w") as f:
|
||||
print("settings\tWER", file=f)
|
||||
for key, val in test_set_wers:
|
||||
|
||||
@ -406,18 +406,14 @@ def save_results(
|
||||
):
|
||||
test_set_wers = dict()
|
||||
for key, results in results_dict.items():
|
||||
recog_path = (
|
||||
params.res_dir / f"recogs-{test_set_name}-{key}-{params.suffix}.txt"
|
||||
)
|
||||
recog_path = params.res_dir / f"recogs-{test_set_name}-{params.suffix}.txt"
|
||||
results = sorted(results)
|
||||
store_transcripts(filename=recog_path, texts=results)
|
||||
logging.info(f"The transcripts are stored in {recog_path}")
|
||||
|
||||
# The following prints out WERs, per-word error statistics and aligned
|
||||
# ref/hyp pairs.
|
||||
errs_filename = (
|
||||
params.res_dir / f"errs-{test_set_name}-{key}-{params.suffix}.txt"
|
||||
)
|
||||
errs_filename = params.res_dir / f"errs-{test_set_name}-{params.suffix}.txt"
|
||||
# we compute CER for aishell dataset.
|
||||
results_char = []
|
||||
for res in results:
|
||||
@ -431,9 +427,7 @@ def save_results(
|
||||
logging.info("Wrote detailed error stats to {}".format(errs_filename))
|
||||
|
||||
test_set_wers = sorted(test_set_wers.items(), key=lambda x: x[1])
|
||||
errs_info = (
|
||||
params.res_dir / f"wer-summary-{test_set_name}-{key}-{params.suffix}.txt"
|
||||
)
|
||||
errs_info = params.res_dir / f"wer-summary-{test_set_name}-{params.suffix}.txt"
|
||||
with open(errs_info, "w") as f:
|
||||
print("settings\tCER", file=f)
|
||||
for key, val in test_set_wers:
|
||||
|
||||
@ -325,17 +325,13 @@ def save_results(
|
||||
):
|
||||
test_set_wers = dict()
|
||||
for key, results in results_dict.items():
|
||||
recog_path = (
|
||||
params.res_dir / f"recogs-{test_set_name}-{key}-{params.suffix}.txt"
|
||||
)
|
||||
recog_path = params.res_dir / f"recogs-{test_set_name}-{params.suffix}.txt"
|
||||
results = sorted(results)
|
||||
store_transcripts(filename=recog_path, texts=results)
|
||||
|
||||
# The following prints out WERs, per-word error statistics and aligned
|
||||
# ref/hyp pairs.
|
||||
errs_filename = (
|
||||
params.res_dir / f"errs-{test_set_name}-{key}-{params.suffix}.txt"
|
||||
)
|
||||
errs_filename = params.res_dir / f"errs-{test_set_name}-{params.suffix}.txt"
|
||||
# we compute CER for aishell dataset.
|
||||
results_char = []
|
||||
for res in results:
|
||||
@ -349,9 +345,7 @@ def save_results(
|
||||
logging.info("Wrote detailed error stats to {}".format(errs_filename))
|
||||
|
||||
test_set_wers = sorted(test_set_wers.items(), key=lambda x: x[1])
|
||||
errs_info = (
|
||||
params.res_dir / f"wer-summary-{test_set_name}-{key}-{params.suffix}.txt"
|
||||
)
|
||||
errs_info = params.res_dir / f"wer-summary-{test_set_name}-{params.suffix}.txt"
|
||||
with open(errs_info, "w") as f:
|
||||
print("settings\tCER", file=f)
|
||||
for key, val in test_set_wers:
|
||||
|
||||
@ -370,18 +370,14 @@ def save_results(
|
||||
):
|
||||
test_set_wers = dict()
|
||||
for key, results in results_dict.items():
|
||||
recog_path = (
|
||||
params.res_dir / f"recogs-{test_set_name}-{key}-{params.suffix}.txt"
|
||||
)
|
||||
recog_path = params.res_dir / f"recogs-{test_set_name}-{params.suffix}.txt"
|
||||
results = sorted(results)
|
||||
store_transcripts(filename=recog_path, texts=results)
|
||||
logging.info(f"The transcripts are stored in {recog_path}")
|
||||
|
||||
# The following prints out WERs, per-word error statistics and aligned
|
||||
# ref/hyp pairs.
|
||||
errs_filename = (
|
||||
params.res_dir / f"errs-{test_set_name}-{key}-{params.suffix}.txt"
|
||||
)
|
||||
errs_filename = params.res_dir / f"errs-{test_set_name}-{params.suffix}.txt"
|
||||
# we compute CER for aishell dataset.
|
||||
results_char = []
|
||||
for res in results:
|
||||
@ -395,9 +391,7 @@ def save_results(
|
||||
logging.info("Wrote detailed error stats to {}".format(errs_filename))
|
||||
|
||||
test_set_wers = sorted(test_set_wers.items(), key=lambda x: x[1])
|
||||
errs_info = (
|
||||
params.res_dir / f"wer-summary-{test_set_name}-{key}-{params.suffix}.txt"
|
||||
)
|
||||
errs_info = params.res_dir / f"wer-summary-{test_set_name}-{params.suffix}.txt"
|
||||
with open(errs_info, "w") as f:
|
||||
print("settings\tCER", file=f)
|
||||
for key, val in test_set_wers:
|
||||
|
||||
@ -374,18 +374,14 @@ def save_results(
|
||||
):
|
||||
test_set_wers = dict()
|
||||
for key, results in results_dict.items():
|
||||
recog_path = (
|
||||
params.res_dir / f"recogs-{test_set_name}-{key}-{params.suffix}.txt"
|
||||
)
|
||||
recog_path = params.res_dir / f"recogs-{test_set_name}-{params.suffix}.txt"
|
||||
results = sorted(results)
|
||||
store_transcripts(filename=recog_path, texts=results)
|
||||
logging.info(f"The transcripts are stored in {recog_path}")
|
||||
|
||||
# The following prints out WERs, per-word error statistics and aligned
|
||||
# ref/hyp pairs.
|
||||
errs_filename = (
|
||||
params.res_dir / f"errs-{test_set_name}-{key}-{params.suffix}.txt"
|
||||
)
|
||||
errs_filename = params.res_dir / f"errs-{test_set_name}-{params.suffix}.txt"
|
||||
# we compute CER for aishell dataset.
|
||||
results_char = []
|
||||
for res in results:
|
||||
@ -399,9 +395,7 @@ def save_results(
|
||||
logging.info("Wrote detailed error stats to {}".format(errs_filename))
|
||||
|
||||
test_set_wers = sorted(test_set_wers.items(), key=lambda x: x[1])
|
||||
errs_info = (
|
||||
params.res_dir / f"wer-summary-{test_set_name}-{key}-{params.suffix}.txt"
|
||||
)
|
||||
errs_info = params.res_dir / f"wer-summary-{test_set_name}-{params.suffix}.txt"
|
||||
with open(errs_info, "w") as f:
|
||||
print("settings\tCER", file=f)
|
||||
for key, val in test_set_wers:
|
||||
|
||||
@ -543,18 +543,14 @@ def save_results(
|
||||
):
|
||||
test_set_wers = dict()
|
||||
for key, results in results_dict.items():
|
||||
recog_path = (
|
||||
params.res_dir / f"recogs-{test_set_name}-{key}-{params.suffix}.txt"
|
||||
)
|
||||
recog_path = params.res_dir / f"recogs-{test_set_name}-{params.suffix}.txt"
|
||||
results = sorted(results)
|
||||
store_transcripts(filename=recog_path, texts=results)
|
||||
logging.info(f"The transcripts are stored in {recog_path}")
|
||||
|
||||
# The following prints out WERs, per-word error statistics and aligned
|
||||
# ref/hyp pairs.
|
||||
errs_filename = (
|
||||
params.res_dir / f"errs-{test_set_name}-{key}-{params.suffix}.txt"
|
||||
)
|
||||
errs_filename = params.res_dir / f"errs-{test_set_name}-{params.suffix}.txt"
|
||||
with open(errs_filename, "w") as f:
|
||||
wer = write_error_stats(
|
||||
f, f"{test_set_name}-{key}", results, enable_log=True
|
||||
@ -564,9 +560,7 @@ def save_results(
|
||||
logging.info("Wrote detailed error stats to {}".format(errs_filename))
|
||||
|
||||
test_set_wers = sorted(test_set_wers.items(), key=lambda x: x[1])
|
||||
errs_info = (
|
||||
params.res_dir / f"wer-summary-{test_set_name}-{key}-{params.suffix}.txt"
|
||||
)
|
||||
errs_info = params.res_dir / f"wer-summary-{test_set_name}-{params.suffix}.txt"
|
||||
with open(errs_info, "w") as f:
|
||||
print("settings\tWER", file=f)
|
||||
for key, val in test_set_wers:
|
||||
|
||||
@ -406,18 +406,14 @@ def save_results(
|
||||
):
|
||||
test_set_wers = dict()
|
||||
for key, results in results_dict.items():
|
||||
recog_path = (
|
||||
params.res_dir / f"recogs-{test_set_name}-{key}-{params.suffix}.txt"
|
||||
)
|
||||
recog_path = params.res_dir / f"recogs-{test_set_name}-{params.suffix}.txt"
|
||||
results = sorted(results)
|
||||
store_transcripts(filename=recog_path, texts=results)
|
||||
logging.info(f"The transcripts are stored in {recog_path}")
|
||||
|
||||
# The following prints out WERs, per-word error statistics and aligned
|
||||
# ref/hyp pairs.
|
||||
errs_filename = (
|
||||
params.res_dir / f"errs-{test_set_name}-{key}-{params.suffix}.txt"
|
||||
)
|
||||
errs_filename = params.res_dir / f"errs-{test_set_name}-{params.suffix}.txt"
|
||||
with open(errs_filename, "w") as f:
|
||||
wer = write_error_stats(
|
||||
f, f"{test_set_name}-{key}", results, enable_log=True
|
||||
@ -427,9 +423,7 @@ def save_results(
|
||||
logging.info("Wrote detailed error stats to {}".format(errs_filename))
|
||||
|
||||
test_set_wers = sorted(test_set_wers.items(), key=lambda x: x[1])
|
||||
errs_info = (
|
||||
params.res_dir / f"wer-summary-{test_set_name}-{key}-{params.suffix}.txt"
|
||||
)
|
||||
errs_info = params.res_dir / f"wer-summary-{test_set_name}-{params.suffix}.txt"
|
||||
with open(errs_info, "w") as f:
|
||||
print("settings\tWER", file=f)
|
||||
for key, val in test_set_wers:
|
||||
|
||||
@ -391,18 +391,14 @@ def save_results(
|
||||
):
|
||||
test_set_wers = dict()
|
||||
for key, results in results_dict.items():
|
||||
recog_path = (
|
||||
params.res_dir / f"recogs-{test_set_name}-{key}-{params.suffix}.txt"
|
||||
)
|
||||
recog_path = params.res_dir / f"recogs-{test_set_name}-{params.suffix}.txt"
|
||||
results = sorted(results)
|
||||
store_transcripts(filename=recog_path, texts=results)
|
||||
logging.info(f"The transcripts are stored in {recog_path}")
|
||||
|
||||
# The following prints out WERs, per-word error statistics and aligned
|
||||
# ref/hyp pairs.
|
||||
errs_filename = (
|
||||
params.res_dir / f"errs-{test_set_name}-{key}-{params.suffix}.txt"
|
||||
)
|
||||
errs_filename = params.res_dir / f"errs-{test_set_name}-{params.suffix}.txt"
|
||||
with open(errs_filename, "w") as f:
|
||||
wer = write_error_stats(
|
||||
f, f"{test_set_name}-{key}", results, enable_log=True
|
||||
@ -412,9 +408,7 @@ def save_results(
|
||||
logging.info("Wrote detailed error stats to {}".format(errs_filename))
|
||||
|
||||
test_set_wers = sorted(test_set_wers.items(), key=lambda x: x[1])
|
||||
errs_info = (
|
||||
params.res_dir / f"wer-summary-{test_set_name}-{key}-{params.suffix}.txt"
|
||||
)
|
||||
errs_info = params.res_dir / f"wer-summary-{test_set_name}-{params.suffix}.txt"
|
||||
with open(errs_info, "w") as f:
|
||||
print("settings\tWER", file=f)
|
||||
for key, val in test_set_wers:
|
||||
|
||||
@ -462,18 +462,14 @@ def save_results(
|
||||
):
|
||||
test_set_wers = dict()
|
||||
for key, results in results_dict.items():
|
||||
recog_path = (
|
||||
params.res_dir / f"recogs-{test_set_name}-{key}-{params.suffix}.txt"
|
||||
)
|
||||
recog_path = params.res_dir / f"recogs-{test_set_name}-{params.suffix}.txt"
|
||||
results = sorted(results)
|
||||
store_transcripts(filename=recog_path, texts=results)
|
||||
logging.info(f"The transcripts are stored in {recog_path}")
|
||||
|
||||
# The following prints out WERs, per-word error statistics and aligned
|
||||
# ref/hyp pairs.
|
||||
errs_filename = (
|
||||
params.res_dir / f"errs-{test_set_name}-{key}-{params.suffix}.txt"
|
||||
)
|
||||
errs_filename = params.res_dir / f"errs-{test_set_name}-{params.suffix}.txt"
|
||||
with open(errs_filename, "w") as f:
|
||||
wer = write_error_stats(
|
||||
f, f"{test_set_name}-{key}", results, enable_log=True
|
||||
@ -483,9 +479,7 @@ def save_results(
|
||||
logging.info("Wrote detailed error stats to {}".format(errs_filename))
|
||||
|
||||
test_set_wers = sorted(test_set_wers.items(), key=lambda x: x[1])
|
||||
errs_info = (
|
||||
params.res_dir / f"wer-summary-{test_set_name}-{key}-{params.suffix}.txt"
|
||||
)
|
||||
errs_info = params.res_dir / f"wer-summary-{test_set_name}-{params.suffix}.txt"
|
||||
with open(errs_info, "w") as f:
|
||||
print("settings\tWER", file=f)
|
||||
for key, val in test_set_wers:
|
||||
|
||||
@ -478,17 +478,13 @@ def save_results(
|
||||
test_set_wers = dict()
|
||||
test_set_cers = dict()
|
||||
for key, results in results_dict.items():
|
||||
recog_path = (
|
||||
params.res_dir / f"recogs-{test_set_name}-{key}-{params.suffix}.txt"
|
||||
)
|
||||
recog_path = params.res_dir / f"recogs-{test_set_name}-{params.suffix}.txt"
|
||||
store_transcripts(filename=recog_path, texts=results)
|
||||
logging.info(f"The transcripts are stored in {recog_path}")
|
||||
|
||||
# The following prints out WERs, per-word error statistics and aligned
|
||||
# ref/hyp pairs.
|
||||
wers_filename = (
|
||||
params.res_dir / f"wers-{test_set_name}-{key}-{params.suffix}.txt"
|
||||
)
|
||||
wers_filename = params.res_dir / f"wers-{test_set_name}-{params.suffix}.txt"
|
||||
with open(wers_filename, "w") as f:
|
||||
wer = write_error_stats(
|
||||
f, f"{test_set_name}-{key}", results, enable_log=True
|
||||
@ -499,9 +495,7 @@ def save_results(
|
||||
results_char = []
|
||||
for res in results:
|
||||
results_char.append((res[0], list("".join(res[1])), list("".join(res[2]))))
|
||||
cers_filename = (
|
||||
params.res_dir / f"cers-{test_set_name}-{key}-{params.suffix}.txt"
|
||||
)
|
||||
cers_filename = params.res_dir / f"cers-{test_set_name}-{params.suffix}.txt"
|
||||
with open(cers_filename, "w") as f:
|
||||
cer = write_error_stats(
|
||||
f, f"{test_set_name}-{key}", results_char, enable_log=True
|
||||
@ -512,9 +506,7 @@ def save_results(
|
||||
|
||||
test_set_wers = {k: v for k, v in sorted(test_set_wers.items(), key=lambda x: x[1])}
|
||||
test_set_cers = {k: v for k, v in sorted(test_set_cers.items(), key=lambda x: x[1])}
|
||||
errs_info = (
|
||||
params.res_dir / f"wer-summary-{test_set_name}-{key}-{params.suffix}.txt"
|
||||
)
|
||||
errs_info = params.res_dir / f"wer-summary-{test_set_name}-{params.suffix}.txt"
|
||||
with open(errs_info, "w") as f:
|
||||
print("settings\tWER\tCER", file=f)
|
||||
for key in test_set_wers:
|
||||
|
||||
@ -599,9 +599,7 @@ def save_results(
|
||||
):
|
||||
test_set_wers = dict()
|
||||
for key, results in results_dict.items():
|
||||
recog_path = (
|
||||
params.res_dir / f"recogs-{test_set_name}-{key}-{params.suffix}.txt"
|
||||
)
|
||||
recog_path = params.res_dir / f"recogs-{test_set_name}-{params.suffix}.txt"
|
||||
results = sorted(results)
|
||||
store_transcripts(filename=recog_path, texts=results)
|
||||
|
||||
@ -609,9 +607,7 @@ def save_results(
|
||||
|
||||
# The following prints out WERs, per-word error statistics and aligned
|
||||
# ref/hyp pairs.
|
||||
errs_filename = (
|
||||
params.res_dir / f"errs-{test_set_name}-{key}-{params.suffix}.txt"
|
||||
)
|
||||
errs_filename = params.res_dir / f"errs-{test_set_name}-{params.suffix}.txt"
|
||||
with open(errs_filename, "w") as f:
|
||||
wer = write_error_stats(
|
||||
f, f"{test_set_name}-{key}", results, enable_log=True
|
||||
@ -621,9 +617,7 @@ def save_results(
|
||||
logging.info("Wrote detailed error stats to {}".format(errs_filename))
|
||||
|
||||
test_set_wers = sorted(test_set_wers.items(), key=lambda x: x[1])
|
||||
errs_info = (
|
||||
params.res_dir / f"wer-summary-{test_set_name}-{key}-{params.suffix}.txt"
|
||||
)
|
||||
errs_info = params.res_dir / f"wer-summary-{test_set_name}-{params.suffix}.txt"
|
||||
with open(errs_info, "w") as f:
|
||||
print("settings\tWER", file=f)
|
||||
for key, val in test_set_wers:
|
||||
|
||||
@ -399,9 +399,7 @@ def save_results(
|
||||
):
|
||||
test_set_wers = dict()
|
||||
for key, results in results_dict.items():
|
||||
recog_path = (
|
||||
params.res_dir / f"recogs-{test_set_name}-{key}-{params.suffix}.txt"
|
||||
)
|
||||
recog_path = params.res_dir / f"recogs-{test_set_name}-{params.suffix}.txt"
|
||||
results = post_processing(results)
|
||||
results = sorted(results)
|
||||
store_transcripts(filename=recog_path, texts=results)
|
||||
@ -409,9 +407,7 @@ def save_results(
|
||||
|
||||
# The following prints out WERs, per-word error statistics and aligned
|
||||
# ref/hyp pairs.
|
||||
errs_filename = (
|
||||
params.res_dir / f"errs-{test_set_name}-{key}-{params.suffix}.txt"
|
||||
)
|
||||
errs_filename = params.res_dir / f"errs-{test_set_name}-{params.suffix}.txt"
|
||||
with open(errs_filename, "w") as f:
|
||||
wer = write_error_stats(
|
||||
f, f"{test_set_name}-{key}", results, enable_log=True
|
||||
@ -421,9 +417,7 @@ def save_results(
|
||||
logging.info("Wrote detailed error stats to {}".format(errs_filename))
|
||||
|
||||
test_set_wers = sorted(test_set_wers.items(), key=lambda x: x[1])
|
||||
errs_info = (
|
||||
params.res_dir / f"wer-summary-{test_set_name}-{key}-{params.suffix}.txt"
|
||||
)
|
||||
errs_info = params.res_dir / f"wer-summary-{test_set_name}-{params.suffix}.txt"
|
||||
with open(errs_info, "w") as f:
|
||||
print("settings\tWER", file=f)
|
||||
for key, val in test_set_wers:
|
||||
|
||||
@ -728,18 +728,14 @@ def save_results(
|
||||
test_set_wers = dict()
|
||||
test_set_delays = dict()
|
||||
for key, results in results_dict.items():
|
||||
recog_path = (
|
||||
params.res_dir / f"recogs-{test_set_name}-{key}-{params.suffix}.txt"
|
||||
)
|
||||
recog_path = params.res_dir / f"recogs-{test_set_name}-{params.suffix}.txt"
|
||||
results = sorted(results)
|
||||
store_transcripts_and_timestamps(filename=recog_path, texts=results)
|
||||
logging.info(f"The transcripts are stored in {recog_path}")
|
||||
|
||||
# The following prints out WERs, per-word error statistics and aligned
|
||||
# ref/hyp pairs.
|
||||
errs_filename = (
|
||||
params.res_dir / f"errs-{test_set_name}-{key}-{params.suffix}.txt"
|
||||
)
|
||||
errs_filename = params.res_dir / f"errs-{test_set_name}-{params.suffix}.txt"
|
||||
with open(errs_filename, "w") as f:
|
||||
wer, mean_delay, var_delay = write_error_stats_with_timestamps(
|
||||
f,
|
||||
@ -754,9 +750,7 @@ def save_results(
|
||||
logging.info("Wrote detailed error stats to {}".format(errs_filename))
|
||||
|
||||
test_set_wers = sorted(test_set_wers.items(), key=lambda x: x[1])
|
||||
errs_info = (
|
||||
params.res_dir / f"wer-summary-{test_set_name}-{key}-{params.suffix}.txt"
|
||||
)
|
||||
errs_info = params.res_dir / f"wer-summary-{test_set_name}-{params.suffix}.txt"
|
||||
with open(errs_info, "w") as f:
|
||||
print("settings\tWER", file=f)
|
||||
for key, val in test_set_wers:
|
||||
@ -765,8 +759,7 @@ def save_results(
|
||||
# sort according to the mean start symbol delay
|
||||
test_set_delays = sorted(test_set_delays.items(), key=lambda x: x[1][0][0])
|
||||
delays_info = (
|
||||
params.res_dir
|
||||
/ f"symbol-delay-summary-{test_set_name}-{key}-{params.suffix}.txt"
|
||||
params.res_dir / f"symbol-delay-summary-{test_set_name}-{params.suffix}.txt"
|
||||
)
|
||||
with open(delays_info, "w") as f:
|
||||
print("settings\t(start, end) symbol-delay (s) (start, end)", file=f)
|
||||
|
||||
@ -432,18 +432,14 @@ def save_results(
|
||||
):
|
||||
test_set_wers = dict()
|
||||
for key, results in results_dict.items():
|
||||
recog_path = (
|
||||
params.res_dir / f"recogs-{test_set_name}-{key}-{params.suffix}.txt"
|
||||
)
|
||||
recog_path = params.res_dir / f"recogs-{test_set_name}-{params.suffix}.txt"
|
||||
results = sorted(results)
|
||||
store_transcripts(filename=recog_path, texts=results)
|
||||
logging.info(f"The transcripts are stored in {recog_path}")
|
||||
|
||||
# The following prints out WERs, per-word error statistics and aligned
|
||||
# ref/hyp pairs.
|
||||
errs_filename = (
|
||||
params.res_dir / f"errs-{test_set_name}-{key}-{params.suffix}.txt"
|
||||
)
|
||||
errs_filename = params.res_dir / f"errs-{test_set_name}-{params.suffix}.txt"
|
||||
with open(errs_filename, "w") as f:
|
||||
wer = write_error_stats(
|
||||
f, f"{test_set_name}-{key}", results, enable_log=True
|
||||
@ -453,9 +449,7 @@ def save_results(
|
||||
logging.info("Wrote detailed error stats to {}".format(errs_filename))
|
||||
|
||||
test_set_wers = sorted(test_set_wers.items(), key=lambda x: x[1])
|
||||
errs_info = (
|
||||
params.res_dir / f"wer-summary-{test_set_name}-{key}-{params.suffix}.txt"
|
||||
)
|
||||
errs_info = params.res_dir / f"wer-summary-{test_set_name}-{params.suffix}.txt"
|
||||
with open(errs_info, "w") as f:
|
||||
print("settings\tWER", file=f)
|
||||
for key, val in test_set_wers:
|
||||
|
||||
@ -750,17 +750,13 @@ def save_results(
|
||||
):
|
||||
test_set_wers = dict()
|
||||
for key, results in results_dict.items():
|
||||
recog_path = (
|
||||
params.res_dir / f"recogs-{test_set_name}-{key}-{params.suffix}.txt"
|
||||
)
|
||||
recog_path = params.res_dir / f"recogs-{test_set_name}-{params.suffix}.txt"
|
||||
store_transcripts(filename=recog_path, texts=sorted(results))
|
||||
logging.info(f"The transcripts are stored in {recog_path}")
|
||||
|
||||
# The following prints out WERs, per-word error statistics and aligned
|
||||
# ref/hyp pairs.
|
||||
errs_filename = (
|
||||
params.res_dir / f"errs-{test_set_name}-{key}-{params.suffix}.txt"
|
||||
)
|
||||
errs_filename = params.res_dir / f"errs-{test_set_name}-{params.suffix}.txt"
|
||||
with open(errs_filename, "w") as f:
|
||||
wer = write_error_stats(
|
||||
f, f"{test_set_name}-{key}", results, enable_log=True
|
||||
@ -770,9 +766,7 @@ def save_results(
|
||||
logging.info("Wrote detailed error stats to {}".format(errs_filename))
|
||||
|
||||
test_set_wers = sorted(test_set_wers.items(), key=lambda x: x[1])
|
||||
errs_info = (
|
||||
params.res_dir / f"wer-summary-{test_set_name}-{key}-{params.suffix}.txt"
|
||||
)
|
||||
errs_info = params.res_dir / f"wer-summary-{test_set_name}-{params.suffix}.txt"
|
||||
with open(errs_info, "w") as f:
|
||||
print("settings\tWER", file=f)
|
||||
for key, val in test_set_wers:
|
||||
|
||||
@ -432,18 +432,14 @@ def save_results(
|
||||
):
|
||||
test_set_wers = dict()
|
||||
for key, results in results_dict.items():
|
||||
recog_path = (
|
||||
params.res_dir / f"recogs-{test_set_name}-{key}-{params.suffix}.txt"
|
||||
)
|
||||
recog_path = params.res_dir / f"recogs-{test_set_name}-{params.suffix}.txt"
|
||||
results = sorted(results)
|
||||
store_transcripts(filename=recog_path, texts=results)
|
||||
logging.info(f"The transcripts are stored in {recog_path}")
|
||||
|
||||
# The following prints out WERs, per-word error statistics and aligned
|
||||
# ref/hyp pairs.
|
||||
errs_filename = (
|
||||
params.res_dir / f"errs-{test_set_name}-{key}-{params.suffix}.txt"
|
||||
)
|
||||
errs_filename = params.res_dir / f"errs-{test_set_name}-{params.suffix}.txt"
|
||||
with open(errs_filename, "w") as f:
|
||||
wer = write_error_stats(
|
||||
f, f"{test_set_name}-{key}", results, enable_log=True
|
||||
@ -453,9 +449,7 @@ def save_results(
|
||||
logging.info("Wrote detailed error stats to {}".format(errs_filename))
|
||||
|
||||
test_set_wers = sorted(test_set_wers.items(), key=lambda x: x[1])
|
||||
errs_info = (
|
||||
params.res_dir / f"wer-summary-{test_set_name}-{key}-{params.suffix}.txt"
|
||||
)
|
||||
errs_info = params.res_dir / f"wer-summary-{test_set_name}-{params.suffix}.txt"
|
||||
with open(errs_info, "w") as f:
|
||||
print("settings\tWER", file=f)
|
||||
for key, val in test_set_wers:
|
||||
|
||||
@ -750,17 +750,13 @@ def save_results(
|
||||
):
|
||||
test_set_wers = dict()
|
||||
for key, results in results_dict.items():
|
||||
recog_path = (
|
||||
params.res_dir / f"recogs-{test_set_name}-{key}-{params.suffix}.txt"
|
||||
)
|
||||
recog_path = params.res_dir / f"recogs-{test_set_name}-{params.suffix}.txt"
|
||||
store_transcripts(filename=recog_path, texts=sorted(results))
|
||||
logging.info(f"The transcripts are stored in {recog_path}")
|
||||
|
||||
# The following prints out WERs, per-word error statistics and aligned
|
||||
# ref/hyp pairs.
|
||||
errs_filename = (
|
||||
params.res_dir / f"errs-{test_set_name}-{key}-{params.suffix}.txt"
|
||||
)
|
||||
errs_filename = params.res_dir / f"errs-{test_set_name}-{params.suffix}.txt"
|
||||
with open(errs_filename, "w") as f:
|
||||
wer = write_error_stats(
|
||||
f, f"{test_set_name}-{key}", results, enable_log=True
|
||||
@ -770,9 +766,7 @@ def save_results(
|
||||
logging.info("Wrote detailed error stats to {}".format(errs_filename))
|
||||
|
||||
test_set_wers = sorted(test_set_wers.items(), key=lambda x: x[1])
|
||||
errs_info = (
|
||||
params.res_dir / f"wer-summary-{test_set_name}-{key}-{params.suffix}.txt"
|
||||
)
|
||||
errs_info = params.res_dir / f"wer-summary-{test_set_name}-{params.suffix}.txt"
|
||||
with open(errs_info, "w") as f:
|
||||
print("settings\tWER", file=f)
|
||||
for key, val in test_set_wers:
|
||||
|
||||
@ -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)
|
||||
|
||||
@ -566,18 +566,14 @@ def save_results(
|
||||
):
|
||||
test_set_wers = dict()
|
||||
for key, results in results_dict.items():
|
||||
recog_path = (
|
||||
params.res_dir / f"recogs-{test_set_name}-{key}-{params.suffix}.txt"
|
||||
)
|
||||
recog_path = params.res_dir / f"recogs-{test_set_name}-{params.suffix}.txt"
|
||||
results = sorted(results)
|
||||
store_transcripts(filename=recog_path, texts=results)
|
||||
logging.info(f"The transcripts are stored in {recog_path}")
|
||||
|
||||
# The following prints out WERs, per-word error statistics and aligned
|
||||
# ref/hyp pairs.
|
||||
errs_filename = (
|
||||
params.res_dir / f"errs-{test_set_name}-{key}-{params.suffix}.txt"
|
||||
)
|
||||
errs_filename = params.res_dir / f"errs-{test_set_name}-{params.suffix}.txt"
|
||||
with open(errs_filename, "w") as f:
|
||||
wer = write_error_stats(
|
||||
f, f"{test_set_name}-{key}", results, enable_log=True
|
||||
@ -587,9 +583,7 @@ def save_results(
|
||||
logging.info("Wrote detailed error stats to {}".format(errs_filename))
|
||||
|
||||
test_set_wers = sorted(test_set_wers.items(), key=lambda x: x[1])
|
||||
errs_info = (
|
||||
params.res_dir / f"wer-summary-{test_set_name}-{key}-{params.suffix}.txt"
|
||||
)
|
||||
errs_info = params.res_dir / f"wer-summary-{test_set_name}-{params.suffix}.txt"
|
||||
with open(errs_info, "w") as f:
|
||||
print("settings\tWER", file=f)
|
||||
for key, val in test_set_wers:
|
||||
|
||||
@ -742,17 +742,13 @@ def save_results(
|
||||
):
|
||||
test_set_wers = dict()
|
||||
for key, results in results_dict.items():
|
||||
recog_path = (
|
||||
params.res_dir / f"recogs-{test_set_name}-{key}-{params.suffix}.txt"
|
||||
)
|
||||
recog_path = params.res_dir / f"recogs-{test_set_name}-{params.suffix}.txt"
|
||||
store_transcripts(filename=recog_path, texts=sorted(results))
|
||||
logging.info(f"The transcripts are stored in {recog_path}")
|
||||
|
||||
# The following prints out WERs, per-word error statistics and aligned
|
||||
# ref/hyp pairs.
|
||||
errs_filename = (
|
||||
params.res_dir / f"errs-{test_set_name}-{key}-{params.suffix}.txt"
|
||||
)
|
||||
errs_filename = params.res_dir / f"errs-{test_set_name}-{params.suffix}.txt"
|
||||
with open(errs_filename, "w") as f:
|
||||
wer = write_error_stats(
|
||||
f, f"{test_set_name}-{key}", results, enable_log=True
|
||||
@ -762,9 +758,7 @@ def save_results(
|
||||
logging.info("Wrote detailed error stats to {}".format(errs_filename))
|
||||
|
||||
test_set_wers = sorted(test_set_wers.items(), key=lambda x: x[1])
|
||||
errs_info = (
|
||||
params.res_dir / f"wer-summary-{test_set_name}-{key}-{params.suffix}.txt"
|
||||
)
|
||||
errs_info = params.res_dir / f"wer-summary-{test_set_name}-{params.suffix}.txt"
|
||||
with open(errs_info, "w") as f:
|
||||
print("settings\tWER", file=f)
|
||||
for key, val in test_set_wers:
|
||||
|
||||
@ -702,18 +702,14 @@ def save_results(
|
||||
):
|
||||
test_set_wers = dict()
|
||||
for key, results in results_dict.items():
|
||||
recog_path = (
|
||||
params.res_dir / f"recogs-{test_set_name}-{key}-{params.suffix}.txt"
|
||||
)
|
||||
recog_path = params.res_dir / f"recogs-{test_set_name}-{params.suffix}.txt"
|
||||
results = sorted(results)
|
||||
store_transcripts(filename=recog_path, texts=results)
|
||||
logging.info(f"The transcripts are stored in {recog_path}")
|
||||
|
||||
# The following prints out WERs, per-word error statistics and aligned
|
||||
# ref/hyp pairs.
|
||||
errs_filename = (
|
||||
params.res_dir / f"errs-{test_set_name}-{key}-{params.suffix}.txt"
|
||||
)
|
||||
errs_filename = params.res_dir / f"errs-{test_set_name}-{params.suffix}.txt"
|
||||
with open(errs_filename, "w") as f:
|
||||
wer = write_error_stats(
|
||||
f, f"{test_set_name}-{key}", results, enable_log=True
|
||||
@ -723,9 +719,7 @@ def save_results(
|
||||
logging.info("Wrote detailed error stats to {}".format(errs_filename))
|
||||
|
||||
test_set_wers = sorted(test_set_wers.items(), key=lambda x: x[1])
|
||||
errs_info = (
|
||||
params.res_dir / f"wer-summary-{test_set_name}-{key}-{params.suffix}.txt"
|
||||
)
|
||||
errs_info = params.res_dir / f"wer-summary-{test_set_name}-{params.suffix}.txt"
|
||||
with open(errs_info, "w") as f:
|
||||
print("settings\tWER", file=f)
|
||||
for key, val in test_set_wers:
|
||||
|
||||
@ -611,18 +611,14 @@ def save_results(
|
||||
test_set_wers = dict()
|
||||
test_set_delays = dict()
|
||||
for key, results in results_dict.items():
|
||||
recog_path = (
|
||||
params.res_dir / f"recogs-{test_set_name}-{key}-{params.suffix}.txt"
|
||||
)
|
||||
recog_path = params.res_dir / f"recogs-{test_set_name}-{params.suffix}.txt"
|
||||
results = sorted(results)
|
||||
store_transcripts_and_timestamps(filename=recog_path, texts=results)
|
||||
logging.info(f"The transcripts are stored in {recog_path}")
|
||||
|
||||
# The following prints out WERs, per-word error statistics and aligned
|
||||
# ref/hyp pairs.
|
||||
errs_filename = (
|
||||
params.res_dir / f"errs-{test_set_name}-{key}-{params.suffix}.txt"
|
||||
)
|
||||
errs_filename = params.res_dir / f"errs-{test_set_name}-{params.suffix}.txt"
|
||||
with open(errs_filename, "w") as f:
|
||||
wer, mean_delay, var_delay = write_error_stats_with_timestamps(
|
||||
f, f"{test_set_name}-{key}", results, enable_log=True
|
||||
@ -633,9 +629,7 @@ def save_results(
|
||||
logging.info("Wrote detailed error stats to {}".format(errs_filename))
|
||||
|
||||
test_set_wers = sorted(test_set_wers.items(), key=lambda x: x[1])
|
||||
errs_info = (
|
||||
params.res_dir / f"wer-summary-{test_set_name}-{key}-{params.suffix}.txt"
|
||||
)
|
||||
errs_info = params.res_dir / f"wer-summary-{test_set_name}-{params.suffix}.txt"
|
||||
with open(errs_info, "w") as f:
|
||||
print("settings\tWER", file=f)
|
||||
for key, val in test_set_wers:
|
||||
@ -643,8 +637,7 @@ def save_results(
|
||||
|
||||
test_set_delays = sorted(test_set_delays.items(), key=lambda x: x[1][0])
|
||||
delays_info = (
|
||||
params.res_dir
|
||||
/ f"symbol-delay-summary-{test_set_name}-{key}-{params.suffix}.txt"
|
||||
params.res_dir / f"symbol-delay-summary-{test_set_name}-{params.suffix}.txt"
|
||||
)
|
||||
with open(delays_info, "w") as f:
|
||||
print("settings\tsymbol-delay", file=f)
|
||||
|
||||
@ -742,17 +742,13 @@ def save_results(
|
||||
):
|
||||
test_set_wers = dict()
|
||||
for key, results in results_dict.items():
|
||||
recog_path = (
|
||||
params.res_dir / f"recogs-{test_set_name}-{key}-{params.suffix}.txt"
|
||||
)
|
||||
recog_path = params.res_dir / f"recogs-{test_set_name}-{params.suffix}.txt"
|
||||
store_transcripts(filename=recog_path, texts=sorted(results))
|
||||
logging.info(f"The transcripts are stored in {recog_path}")
|
||||
|
||||
# The following prints out WERs, per-word error statistics and aligned
|
||||
# ref/hyp pairs.
|
||||
errs_filename = (
|
||||
params.res_dir / f"errs-{test_set_name}-{key}-{params.suffix}.txt"
|
||||
)
|
||||
errs_filename = params.res_dir / f"errs-{test_set_name}-{params.suffix}.txt"
|
||||
with open(errs_filename, "w") as f:
|
||||
wer = write_error_stats(
|
||||
f, f"{test_set_name}-{key}", results, enable_log=True
|
||||
@ -762,9 +758,7 @@ def save_results(
|
||||
logging.info("Wrote detailed error stats to {}".format(errs_filename))
|
||||
|
||||
test_set_wers = sorted(test_set_wers.items(), key=lambda x: x[1])
|
||||
errs_info = (
|
||||
params.res_dir / f"wer-summary-{test_set_name}-{key}-{params.suffix}.txt"
|
||||
)
|
||||
errs_info = params.res_dir / f"wer-summary-{test_set_name}-{params.suffix}.txt"
|
||||
with open(errs_info, "w") as f:
|
||||
print("settings\tWER", file=f)
|
||||
for key, val in test_set_wers:
|
||||
|
||||
@ -386,17 +386,13 @@ def save_results(
|
||||
):
|
||||
test_set_wers = dict()
|
||||
for key, results in results_dict.items():
|
||||
recog_path = (
|
||||
params.res_dir / f"recogs-{test_set_name}-{key}-{params.suffix}.txt"
|
||||
)
|
||||
recog_path = params.res_dir / f"recogs-{test_set_name}-{params.suffix}.txt"
|
||||
store_transcripts(filename=recog_path, texts=results)
|
||||
logging.info(f"The transcripts are stored in {recog_path}")
|
||||
|
||||
# The following prints out WERs, per-word error statistics and aligned
|
||||
# ref/hyp pairs.
|
||||
errs_filename = (
|
||||
params.res_dir / f"errs-{test_set_name}-{key}-{params.suffix}.txt"
|
||||
)
|
||||
errs_filename = params.res_dir / f"errs-{test_set_name}-{params.suffix}.txt"
|
||||
with open(errs_filename, "w") as f:
|
||||
wer = write_error_stats(
|
||||
f, f"{test_set_name}-{key}", results, enable_log=True
|
||||
@ -406,9 +402,7 @@ def save_results(
|
||||
logging.info("Wrote detailed error stats to {}".format(errs_filename))
|
||||
|
||||
test_set_wers = sorted(test_set_wers.items(), key=lambda x: x[1])
|
||||
errs_info = (
|
||||
params.res_dir / f"wer-summary-{test_set_name}-{key}-{params.suffix}.txt"
|
||||
)
|
||||
errs_info = params.res_dir / f"wer-summary-{test_set_name}-{params.suffix}.txt"
|
||||
with open(errs_info, "w") as f:
|
||||
print("settings\tWER", file=f)
|
||||
for key, val in test_set_wers:
|
||||
|
||||
@ -420,18 +420,14 @@ def save_results(
|
||||
):
|
||||
test_set_wers = dict()
|
||||
for key, results in results_dict.items():
|
||||
recog_path = (
|
||||
params.res_dir / f"recogs-{test_set_name}-{key}-{params.suffix}.txt"
|
||||
)
|
||||
recog_path = params.res_dir / f"recogs-{test_set_name}-{params.suffix}.txt"
|
||||
results = sorted(results)
|
||||
store_transcripts(filename=recog_path, texts=results)
|
||||
logging.info(f"The transcripts are stored in {recog_path}")
|
||||
|
||||
# The following prints out WERs, per-word error statistics and aligned
|
||||
# ref/hyp pairs.
|
||||
errs_filename = (
|
||||
params.res_dir / f"errs-{test_set_name}-{key}-{params.suffix}.txt"
|
||||
)
|
||||
errs_filename = params.res_dir / f"errs-{test_set_name}-{params.suffix}.txt"
|
||||
with open(errs_filename, "w") as f:
|
||||
wer = write_error_stats(
|
||||
f, f"{test_set_name}-{key}", results, enable_log=True
|
||||
@ -441,9 +437,7 @@ def save_results(
|
||||
logging.info("Wrote detailed error stats to {}".format(errs_filename))
|
||||
|
||||
test_set_wers = sorted(test_set_wers.items(), key=lambda x: x[1])
|
||||
errs_info = (
|
||||
params.res_dir / f"wer-summary-{test_set_name}-{key}-{params.suffix}.txt"
|
||||
)
|
||||
errs_info = params.res_dir / f"wer-summary-{test_set_name}-{params.suffix}.txt"
|
||||
with open(errs_info, "w") as f:
|
||||
print("settings\tWER", file=f)
|
||||
for key, val in test_set_wers:
|
||||
|
||||
@ -585,18 +585,14 @@ def save_results(
|
||||
):
|
||||
test_set_wers = dict()
|
||||
for key, results in results_dict.items():
|
||||
recog_path = (
|
||||
params.res_dir / f"recogs-{test_set_name}-{key}-{params.suffix}.txt"
|
||||
)
|
||||
recog_path = params.res_dir / f"recogs-{test_set_name}-{params.suffix}.txt"
|
||||
results = sorted(results)
|
||||
store_transcripts(filename=recog_path, texts=results)
|
||||
logging.info(f"The transcripts are stored in {recog_path}")
|
||||
|
||||
# The following prints out WERs, per-word error statistics and aligned
|
||||
# ref/hyp pairs.
|
||||
errs_filename = (
|
||||
params.res_dir / f"errs-{test_set_name}-{key}-{params.suffix}.txt"
|
||||
)
|
||||
errs_filename = params.res_dir / f"errs-{test_set_name}-{params.suffix}.txt"
|
||||
with open(errs_filename, "w") as f:
|
||||
wer = write_error_stats(
|
||||
f, f"{test_set_name}-{key}", results, enable_log=True
|
||||
@ -606,9 +602,7 @@ def save_results(
|
||||
logging.info("Wrote detailed error stats to {}".format(errs_filename))
|
||||
|
||||
test_set_wers = sorted(test_set_wers.items(), key=lambda x: x[1])
|
||||
errs_info = (
|
||||
params.res_dir / f"wer-summary-{test_set_name}-{key}-{params.suffix}.txt"
|
||||
)
|
||||
errs_info = params.res_dir / f"wer-summary-{test_set_name}-{params.suffix}.txt"
|
||||
with open(errs_info, "w") as f:
|
||||
print("settings\tWER", file=f)
|
||||
for key, val in test_set_wers:
|
||||
|
||||
@ -423,9 +423,7 @@ def save_results(
|
||||
):
|
||||
test_set_wers = dict()
|
||||
for key, results in results_dict.items():
|
||||
recog_path = (
|
||||
params.res_dir / f"recogs-{test_set_name}-{key}-{params.suffix}.txt"
|
||||
)
|
||||
recog_path = params.res_dir / f"recogs-{test_set_name}-{params.suffix}.txt"
|
||||
# sort results so we can easily compare the difference between two
|
||||
# recognition results
|
||||
results = sorted(results)
|
||||
@ -434,9 +432,7 @@ def save_results(
|
||||
|
||||
# The following prints out WERs, per-word error statistics and aligned
|
||||
# ref/hyp pairs.
|
||||
errs_filename = (
|
||||
params.res_dir / f"errs-{test_set_name}-{key}-{params.suffix}.txt"
|
||||
)
|
||||
errs_filename = params.res_dir / f"errs-{test_set_name}-{params.suffix}.txt"
|
||||
with open(errs_filename, "w") as f:
|
||||
wer = write_error_stats(
|
||||
f, f"{test_set_name}-{key}", results, enable_log=True
|
||||
@ -446,9 +442,7 @@ def save_results(
|
||||
logging.info("Wrote detailed error stats to {}".format(errs_filename))
|
||||
|
||||
test_set_wers = sorted(test_set_wers.items(), key=lambda x: x[1])
|
||||
errs_info = (
|
||||
params.res_dir / f"wer-summary-{test_set_name}-{key}-{params.suffix}.txt"
|
||||
)
|
||||
errs_info = params.res_dir / f"wer-summary-{test_set_name}-{params.suffix}.txt"
|
||||
with open(errs_info, "w") as f:
|
||||
print("settings\tWER", file=f)
|
||||
for key, val in test_set_wers:
|
||||
|
||||
@ -609,18 +609,14 @@ def save_results(
|
||||
):
|
||||
test_set_wers = dict()
|
||||
for key, results in results_dict.items():
|
||||
recog_path = (
|
||||
params.res_dir / f"recogs-{test_set_name}-{key}-{params.suffix}.txt"
|
||||
)
|
||||
recog_path = params.res_dir / f"recogs-{test_set_name}-{params.suffix}.txt"
|
||||
results = sorted(results)
|
||||
store_transcripts(filename=recog_path, texts=results)
|
||||
logging.info(f"The transcripts are stored in {recog_path}")
|
||||
|
||||
# The following prints out WERs, per-word error statistics and aligned
|
||||
# ref/hyp pairs.
|
||||
errs_filename = (
|
||||
params.res_dir / f"errs-{test_set_name}-{key}-{params.suffix}.txt"
|
||||
)
|
||||
errs_filename = params.res_dir / f"errs-{test_set_name}-{params.suffix}.txt"
|
||||
with open(errs_filename, "w") as f:
|
||||
wer = write_error_stats(
|
||||
f, f"{test_set_name}-{key}", results, enable_log=True
|
||||
@ -630,9 +626,7 @@ def save_results(
|
||||
logging.info("Wrote detailed error stats to {}".format(errs_filename))
|
||||
|
||||
test_set_wers = sorted(test_set_wers.items(), key=lambda x: x[1])
|
||||
errs_info = (
|
||||
params.res_dir / f"wer-summary-{test_set_name}-{key}-{params.suffix}.txt"
|
||||
)
|
||||
errs_info = params.res_dir / f"wer-summary-{test_set_name}-{params.suffix}.txt"
|
||||
with open(errs_info, "w") as f:
|
||||
print("settings\tWER", file=f)
|
||||
for key, val in test_set_wers:
|
||||
|
||||
@ -425,9 +425,7 @@ def save_results(
|
||||
):
|
||||
test_set_wers = dict()
|
||||
for key, results in results_dict.items():
|
||||
recog_path = (
|
||||
params.res_dir / f"recogs-{test_set_name}-{key}-{params.suffix}.txt"
|
||||
)
|
||||
recog_path = params.res_dir / f"recogs-{test_set_name}-{params.suffix}.txt"
|
||||
# sort results so we can easily compare the difference between two
|
||||
# recognition results
|
||||
results = sorted(results)
|
||||
@ -436,9 +434,7 @@ def save_results(
|
||||
|
||||
# The following prints out WERs, per-word error statistics and aligned
|
||||
# ref/hyp pairs.
|
||||
errs_filename = (
|
||||
params.res_dir / f"errs-{test_set_name}-{key}-{params.suffix}.txt"
|
||||
)
|
||||
errs_filename = params.res_dir / f"errs-{test_set_name}-{params.suffix}.txt"
|
||||
with open(errs_filename, "w") as f:
|
||||
wer = write_error_stats(
|
||||
f, f"{test_set_name}-{key}", results, enable_log=True
|
||||
@ -448,9 +444,7 @@ def save_results(
|
||||
logging.info("Wrote detailed error stats to {}".format(errs_filename))
|
||||
|
||||
test_set_wers = sorted(test_set_wers.items(), key=lambda x: x[1])
|
||||
errs_info = (
|
||||
params.res_dir / f"wer-summary-{test_set_name}-{key}-{params.suffix}.txt"
|
||||
)
|
||||
errs_info = params.res_dir / f"wer-summary-{test_set_name}-{params.suffix}.txt"
|
||||
with open(errs_info, "w") as f:
|
||||
print("settings\tWER", file=f)
|
||||
for key, val in test_set_wers:
|
||||
|
||||
@ -869,18 +869,14 @@ def save_results(
|
||||
):
|
||||
test_set_wers = dict()
|
||||
for key, results in results_dict.items():
|
||||
recog_path = (
|
||||
params.res_dir / f"recogs-{test_set_name}-{key}-{params.suffix}.txt"
|
||||
)
|
||||
recog_path = params.res_dir / f"recogs-{test_set_name}-{params.suffix}.txt"
|
||||
results = sorted(results)
|
||||
store_transcripts(filename=recog_path, texts=results)
|
||||
logging.info(f"The transcripts are stored in {recog_path}")
|
||||
|
||||
# The following prints out WERs, per-word error statistics and aligned
|
||||
# ref/hyp pairs.
|
||||
errs_filename = (
|
||||
params.res_dir / f"errs-{test_set_name}-{key}-{params.suffix}.txt"
|
||||
)
|
||||
errs_filename = params.res_dir / f"errs-{test_set_name}-{params.suffix}.txt"
|
||||
with open(errs_filename, "w") as f:
|
||||
wer = write_error_stats(
|
||||
f, f"{test_set_name}-{key}", results, enable_log=True
|
||||
@ -890,9 +886,7 @@ def save_results(
|
||||
logging.info("Wrote detailed error stats to {}".format(errs_filename))
|
||||
|
||||
test_set_wers = sorted(test_set_wers.items(), key=lambda x: x[1])
|
||||
errs_info = (
|
||||
params.res_dir / f"wer-summary-{test_set_name}-{key}-{params.suffix}.txt"
|
||||
)
|
||||
errs_info = params.res_dir / f"wer-summary-{test_set_name}-{params.suffix}.txt"
|
||||
with open(errs_info, "w") as f:
|
||||
print("settings\tWER", file=f)
|
||||
for key, val in test_set_wers:
|
||||
|
||||
@ -426,18 +426,14 @@ def save_results(
|
||||
):
|
||||
test_set_wers = dict()
|
||||
for key, results in results_dict.items():
|
||||
recog_path = (
|
||||
params.res_dir / f"recogs-{test_set_name}-{key}-{params.suffix}.txt"
|
||||
)
|
||||
recog_path = params.res_dir / f"recogs-{test_set_name}-{params.suffix}.txt"
|
||||
results = sorted(results)
|
||||
store_transcripts(filename=recog_path, texts=results)
|
||||
logging.info(f"The transcripts are stored in {recog_path}")
|
||||
|
||||
# The following prints out WERs, per-word error statistics and aligned
|
||||
# ref/hyp pairs.
|
||||
errs_filename = (
|
||||
params.res_dir / f"errs-{test_set_name}-{key}-{params.suffix}.txt"
|
||||
)
|
||||
errs_filename = params.res_dir / f"errs-{test_set_name}-{params.suffix}.txt"
|
||||
with open(errs_filename, "w") as f:
|
||||
wer = write_error_stats(
|
||||
f, f"{test_set_name}-{key}", results, enable_log=True
|
||||
@ -447,9 +443,7 @@ def save_results(
|
||||
logging.info("Wrote detailed error stats to {}".format(errs_filename))
|
||||
|
||||
test_set_wers = sorted(test_set_wers.items(), key=lambda x: x[1])
|
||||
errs_info = (
|
||||
params.res_dir / f"wer-summary-{test_set_name}-{key}-{params.suffix}.txt"
|
||||
)
|
||||
errs_info = params.res_dir / f"wer-summary-{test_set_name}-{params.suffix}.txt"
|
||||
with open(errs_info, "w") as f:
|
||||
print("settings\tWER", file=f)
|
||||
for key, val in test_set_wers:
|
||||
|
||||
@ -655,18 +655,14 @@ def save_results(
|
||||
test_set_wers = dict()
|
||||
test_set_delays = dict()
|
||||
for key, results in results_dict.items():
|
||||
recog_path = (
|
||||
params.res_dir / f"recogs-{test_set_name}-{key}-{params.suffix}.txt"
|
||||
)
|
||||
recog_path = params.res_dir / f"recogs-{test_set_name}-{params.suffix}.txt"
|
||||
results = sorted(results)
|
||||
store_transcripts_and_timestamps(filename=recog_path, texts=results)
|
||||
logging.info(f"The transcripts are stored in {recog_path}")
|
||||
|
||||
# The following prints out WERs, per-word error statistics and aligned
|
||||
# ref/hyp pairs.
|
||||
errs_filename = (
|
||||
params.res_dir / f"errs-{test_set_name}-{key}-{params.suffix}.txt"
|
||||
)
|
||||
errs_filename = params.res_dir / f"errs-{test_set_name}-{params.suffix}.txt"
|
||||
with open(errs_filename, "w") as f:
|
||||
wer, mean_delay, var_delay = write_error_stats_with_timestamps(
|
||||
f, f"{test_set_name}-{key}", results, enable_log=True
|
||||
@ -677,9 +673,7 @@ def save_results(
|
||||
logging.info("Wrote detailed error stats to {}".format(errs_filename))
|
||||
|
||||
test_set_wers = sorted(test_set_wers.items(), key=lambda x: x[1])
|
||||
errs_info = (
|
||||
params.res_dir / f"wer-summary-{test_set_name}-{key}-{params.suffix}.txt"
|
||||
)
|
||||
errs_info = params.res_dir / f"wer-summary-{test_set_name}-{params.suffix}.txt"
|
||||
with open(errs_info, "w") as f:
|
||||
print("settings\tWER", file=f)
|
||||
for key, val in test_set_wers:
|
||||
@ -687,8 +681,7 @@ def save_results(
|
||||
|
||||
test_set_delays = sorted(test_set_delays.items(), key=lambda x: x[1][0])
|
||||
delays_info = (
|
||||
params.res_dir
|
||||
/ f"symbol-delay-summary-{test_set_name}-{key}-{params.suffix}.txt"
|
||||
params.res_dir / f"symbol-delay-summary-{test_set_name}-{params.suffix}.txt"
|
||||
)
|
||||
with open(delays_info, "w") as f:
|
||||
print("settings\tsymbol-delay", file=f)
|
||||
|
||||
@ -442,18 +442,14 @@ def save_results(
|
||||
):
|
||||
test_set_wers = dict()
|
||||
for key, results in results_dict.items():
|
||||
recog_path = (
|
||||
params.res_dir / f"recogs-{test_set_name}-{key}-{params.suffix}.txt"
|
||||
)
|
||||
recog_path = params.res_dir / f"recogs-{test_set_name}-{params.suffix}.txt"
|
||||
results = sorted(results)
|
||||
store_transcripts(filename=recog_path, texts=results)
|
||||
logging.info(f"The transcripts are stored in {recog_path}")
|
||||
|
||||
# The following prints out WERs, per-word error statistics and aligned
|
||||
# ref/hyp pairs.
|
||||
errs_filename = (
|
||||
params.res_dir / f"errs-{test_set_name}-{key}-{params.suffix}.txt"
|
||||
)
|
||||
errs_filename = params.res_dir / f"errs-{test_set_name}-{params.suffix}.txt"
|
||||
with open(errs_filename, "w") as f:
|
||||
wer = write_error_stats(
|
||||
f, f"{test_set_name}-{key}", results, enable_log=True
|
||||
@ -463,9 +459,7 @@ def save_results(
|
||||
logging.info("Wrote detailed error stats to {}".format(errs_filename))
|
||||
|
||||
test_set_wers = sorted(test_set_wers.items(), key=lambda x: x[1])
|
||||
errs_info = (
|
||||
params.res_dir / f"wer-summary-{test_set_name}-{key}-{params.suffix}.txt"
|
||||
)
|
||||
errs_info = params.res_dir / f"wer-summary-{test_set_name}-{params.suffix}.txt"
|
||||
with open(errs_info, "w") as f:
|
||||
print("settings\tWER", file=f)
|
||||
for key, val in test_set_wers:
|
||||
|
||||
@ -735,18 +735,14 @@ def save_results(
|
||||
):
|
||||
test_set_wers = dict()
|
||||
for key, results in results_dict.items():
|
||||
recog_path = (
|
||||
params.res_dir / f"recogs-{test_set_name}-{key}-{params.suffix}.txt"
|
||||
)
|
||||
recog_path = params.res_dir / f"recogs-{test_set_name}-{params.suffix}.txt"
|
||||
results = sorted(results)
|
||||
store_transcripts(filename=recog_path, texts=results)
|
||||
logging.info(f"The transcripts are stored in {recog_path}")
|
||||
|
||||
# The following prints out WERs, per-word error statistics and aligned
|
||||
# ref/hyp pairs.
|
||||
errs_filename = (
|
||||
params.res_dir / f"errs-{test_set_name}-{key}-{params.suffix}.txt"
|
||||
)
|
||||
errs_filename = params.res_dir / f"errs-{test_set_name}-{params.suffix}.txt"
|
||||
with open(errs_filename, "w") as f:
|
||||
wer = write_error_stats(
|
||||
f, f"{test_set_name}-{key}", results, enable_log=True
|
||||
@ -756,9 +752,7 @@ def save_results(
|
||||
logging.info("Wrote detailed error stats to {}".format(errs_filename))
|
||||
|
||||
test_set_wers = sorted(test_set_wers.items(), key=lambda x: x[1])
|
||||
errs_info = (
|
||||
params.res_dir / f"wer-summary-{test_set_name}-{key}-{params.suffix}.txt"
|
||||
)
|
||||
errs_info = params.res_dir / f"wer-summary-{test_set_name}-{params.suffix}.txt"
|
||||
with open(errs_info, "w") as f:
|
||||
print("settings\tWER", file=f)
|
||||
for key, val in test_set_wers:
|
||||
|
||||
@ -442,18 +442,14 @@ def save_results(
|
||||
):
|
||||
test_set_wers = dict()
|
||||
for key, results in results_dict.items():
|
||||
recog_path = (
|
||||
params.res_dir / f"recogs-{test_set_name}-{key}-{params.suffix}.txt"
|
||||
)
|
||||
recog_path = params.res_dir / f"recogs-{test_set_name}-{params.suffix}.txt"
|
||||
results = sorted(results)
|
||||
store_transcripts(filename=recog_path, texts=results)
|
||||
logging.info(f"The transcripts are stored in {recog_path}")
|
||||
|
||||
# The following prints out WERs, per-word error statistics and aligned
|
||||
# ref/hyp pairs.
|
||||
errs_filename = (
|
||||
params.res_dir / f"errs-{test_set_name}-{key}-{params.suffix}.txt"
|
||||
)
|
||||
errs_filename = params.res_dir / f"errs-{test_set_name}-{params.suffix}.txt"
|
||||
with open(errs_filename, "w") as f:
|
||||
wer = write_error_stats(
|
||||
f, f"{test_set_name}-{key}", results, enable_log=True
|
||||
@ -463,9 +459,7 @@ def save_results(
|
||||
logging.info("Wrote detailed error stats to {}".format(errs_filename))
|
||||
|
||||
test_set_wers = sorted(test_set_wers.items(), key=lambda x: x[1])
|
||||
errs_info = (
|
||||
params.res_dir / f"wer-summary-{test_set_name}-{key}-{params.suffix}.txt"
|
||||
)
|
||||
errs_info = params.res_dir / f"wer-summary-{test_set_name}-{params.suffix}.txt"
|
||||
with open(errs_info, "w") as f:
|
||||
print("settings\tWER", file=f)
|
||||
for key, val in test_set_wers:
|
||||
|
||||
@ -416,18 +416,14 @@ def save_results(
|
||||
):
|
||||
test_set_wers = dict()
|
||||
for key, results in results_dict.items():
|
||||
recog_path = (
|
||||
params.res_dir / f"recogs-{test_set_name}-{key}-{params.suffix}.txt"
|
||||
)
|
||||
recog_path = params.res_dir / f"recogs-{test_set_name}-{params.suffix}.txt"
|
||||
results = sorted(results)
|
||||
store_transcripts(filename=recog_path, texts=results)
|
||||
logging.info(f"The transcripts are stored in {recog_path}")
|
||||
|
||||
# The following prints out WERs, per-word error statistics and aligned
|
||||
# ref/hyp pairs.
|
||||
errs_filename = (
|
||||
params.res_dir / f"errs-{test_set_name}-{key}-{params.suffix}.txt"
|
||||
)
|
||||
errs_filename = params.res_dir / f"errs-{test_set_name}-{params.suffix}.txt"
|
||||
with open(errs_filename, "w") as f:
|
||||
wer = write_error_stats(
|
||||
f, f"{test_set_name}-{key}", results, enable_log=True
|
||||
@ -437,9 +433,7 @@ def save_results(
|
||||
logging.info("Wrote detailed error stats to {}".format(errs_filename))
|
||||
|
||||
test_set_wers = sorted(test_set_wers.items(), key=lambda x: x[1])
|
||||
errs_info = (
|
||||
params.res_dir / f"wer-summary-{test_set_name}-{key}-{params.suffix}.txt"
|
||||
)
|
||||
errs_info = params.res_dir / f"wer-summary-{test_set_name}-{params.suffix}.txt"
|
||||
with open(errs_info, "w") as f:
|
||||
print("settings\tWER", file=f)
|
||||
for key, val in test_set_wers:
|
||||
|
||||
@ -722,18 +722,14 @@ def save_results(
|
||||
):
|
||||
test_set_wers = dict()
|
||||
for key, results in results_dict.items():
|
||||
recog_path = (
|
||||
params.res_dir / f"recogs-{test_set_name}-{key}-{params.suffix}.txt"
|
||||
)
|
||||
recog_path = params.res_dir / f"recogs-{test_set_name}-{params.suffix}.txt"
|
||||
results = sorted(results)
|
||||
store_transcripts(filename=recog_path, texts=results)
|
||||
logging.info(f"The transcripts are stored in {recog_path}")
|
||||
|
||||
# The following prints out WERs, per-word error statistics and aligned
|
||||
# ref/hyp pairs.
|
||||
errs_filename = (
|
||||
params.res_dir / f"errs-{test_set_name}-{key}-{params.suffix}.txt"
|
||||
)
|
||||
errs_filename = params.res_dir / f"errs-{test_set_name}-{params.suffix}.txt"
|
||||
with open(errs_filename, "w") as f:
|
||||
wer = write_error_stats(
|
||||
f, f"{test_set_name}-{key}", results, enable_log=True
|
||||
@ -743,9 +739,7 @@ def save_results(
|
||||
logging.info("Wrote detailed error stats to {}".format(errs_filename))
|
||||
|
||||
test_set_wers = sorted(test_set_wers.items(), key=lambda x: x[1])
|
||||
errs_info = (
|
||||
params.res_dir / f"wer-summary-{test_set_name}-{key}-{params.suffix}.txt"
|
||||
)
|
||||
errs_info = params.res_dir / f"wer-summary-{test_set_name}-{params.suffix}.txt"
|
||||
with open(errs_info, "w") as f:
|
||||
print("settings\tWER", file=f)
|
||||
for key, val in test_set_wers:
|
||||
|
||||
@ -541,18 +541,14 @@ def save_results(
|
||||
):
|
||||
test_set_wers = dict()
|
||||
for key, results in results_dict.items():
|
||||
recog_path = (
|
||||
params.res_dir / f"recogs-{test_set_name}-{key}-{params.suffix}.txt"
|
||||
)
|
||||
recog_path = params.res_dir / f"recogs-{test_set_name}-{params.suffix}.txt"
|
||||
results = sorted(results)
|
||||
store_transcripts(filename=recog_path, texts=results)
|
||||
logging.info(f"The transcripts are stored in {recog_path}")
|
||||
|
||||
# The following prints out WERs, per-word error statistics and aligned
|
||||
# ref/hyp pairs.
|
||||
errs_filename = (
|
||||
params.res_dir / f"errs-{test_set_name}-{key}-{params.suffix}.txt"
|
||||
)
|
||||
errs_filename = params.res_dir / f"errs-{test_set_name}-{params.suffix}.txt"
|
||||
with open(errs_filename, "w") as f:
|
||||
wer = write_error_stats(f, f"{test_set_name}-{key}", results)
|
||||
test_set_wers[key] = wer
|
||||
@ -560,9 +556,7 @@ def save_results(
|
||||
logging.info("Wrote detailed error stats to {}".format(errs_filename))
|
||||
|
||||
test_set_wers = sorted(test_set_wers.items(), key=lambda x: x[1])
|
||||
errs_info = (
|
||||
params.res_dir / f"wer-summary-{test_set_name}-{key}-{params.suffix}.txt"
|
||||
)
|
||||
errs_info = params.res_dir / f"wer-summary-{test_set_name}-{params.suffix}.txt"
|
||||
with open(errs_info, "w") as f:
|
||||
print("settings\tWER", file=f)
|
||||
for key, val in test_set_wers:
|
||||
|
||||
@ -593,18 +593,14 @@ def save_results(
|
||||
):
|
||||
test_set_wers = dict()
|
||||
for key, results in results_dict.items():
|
||||
recog_path = (
|
||||
params.res_dir / f"recogs-{test_set_name}-{key}-{params.suffix}.txt"
|
||||
)
|
||||
recog_path = params.res_dir / f"recogs-{test_set_name}-{params.suffix}.txt"
|
||||
results = sorted(results)
|
||||
store_transcripts(filename=recog_path, texts=results)
|
||||
logging.info(f"The transcripts are stored in {recog_path}")
|
||||
|
||||
# The following prints out WERs, per-word error statistics and aligned
|
||||
# ref/hyp pairs.
|
||||
errs_filename = (
|
||||
params.res_dir / f"errs-{test_set_name}-{key}-{params.suffix}.txt"
|
||||
)
|
||||
errs_filename = params.res_dir / f"errs-{test_set_name}-{params.suffix}.txt"
|
||||
with open(errs_filename, "w") as f:
|
||||
wer = write_error_stats(
|
||||
f, f"{test_set_name}-{key}", results, enable_log=True
|
||||
@ -614,9 +610,7 @@ def save_results(
|
||||
logging.info("Wrote detailed error stats to {}".format(errs_filename))
|
||||
|
||||
test_set_wers = sorted(test_set_wers.items(), key=lambda x: x[1])
|
||||
errs_info = (
|
||||
params.res_dir / f"wer-summary-{test_set_name}-{key}-{params.suffix}.txt"
|
||||
)
|
||||
errs_info = params.res_dir / f"wer-summary-{test_set_name}-{params.suffix}.txt"
|
||||
with open(errs_info, "w") as f:
|
||||
print("settings\tWER", file=f)
|
||||
for key, val in test_set_wers:
|
||||
|
||||
@ -532,18 +532,14 @@ def save_results(
|
||||
):
|
||||
test_set_wers = dict()
|
||||
for key, results in results_dict.items():
|
||||
recog_path = (
|
||||
params.res_dir / f"recogs-{test_set_name}-{key}-{params.suffix}.txt"
|
||||
)
|
||||
recog_path = params.res_dir / f"recogs-{test_set_name}-{params.suffix}.txt"
|
||||
results = sorted(results)
|
||||
store_transcripts(filename=recog_path, texts=results)
|
||||
logging.info(f"The transcripts are stored in {recog_path}")
|
||||
|
||||
# The following prints out WERs, per-word error statistics and aligned
|
||||
# ref/hyp pairs.
|
||||
errs_filename = (
|
||||
params.res_dir / f"errs-{test_set_name}-{key}-{params.suffix}.txt"
|
||||
)
|
||||
errs_filename = params.res_dir / f"errs-{test_set_name}-{params.suffix}.txt"
|
||||
with open(errs_filename, "w") as f:
|
||||
wer = write_error_stats(f, f"{test_set_name}-{key}", results)
|
||||
test_set_wers[key] = wer
|
||||
@ -551,9 +547,7 @@ def save_results(
|
||||
logging.info("Wrote detailed error stats to {}".format(errs_filename))
|
||||
|
||||
test_set_wers = sorted(test_set_wers.items(), key=lambda x: x[1])
|
||||
errs_info = (
|
||||
params.res_dir / f"wer-summary-{test_set_name}-{key}-{params.suffix}.txt"
|
||||
)
|
||||
errs_info = params.res_dir / f"wer-summary-{test_set_name}-{params.suffix}.txt"
|
||||
with open(errs_info, "w") as f:
|
||||
print("settings\tWER", file=f)
|
||||
for key, val in test_set_wers:
|
||||
|
||||
@ -593,18 +593,14 @@ def save_results(
|
||||
):
|
||||
test_set_wers = dict()
|
||||
for key, results in results_dict.items():
|
||||
recog_path = (
|
||||
params.res_dir / f"recogs-{test_set_name}-{key}-{params.suffix}.txt"
|
||||
)
|
||||
recog_path = params.res_dir / f"recogs-{test_set_name}-{params.suffix}.txt"
|
||||
results = sorted(results)
|
||||
store_transcripts(filename=recog_path, texts=results)
|
||||
logging.info(f"The transcripts are stored in {recog_path}")
|
||||
|
||||
# The following prints out WERs, per-word error statistics and aligned
|
||||
# ref/hyp pairs.
|
||||
errs_filename = (
|
||||
params.res_dir / f"errs-{test_set_name}-{key}-{params.suffix}.txt"
|
||||
)
|
||||
errs_filename = params.res_dir / f"errs-{test_set_name}-{params.suffix}.txt"
|
||||
with open(errs_filename, "w") as f:
|
||||
wer = write_error_stats(
|
||||
f, f"{test_set_name}-{key}", results, enable_log=True
|
||||
@ -614,9 +610,7 @@ def save_results(
|
||||
logging.info("Wrote detailed error stats to {}".format(errs_filename))
|
||||
|
||||
test_set_wers = sorted(test_set_wers.items(), key=lambda x: x[1])
|
||||
errs_info = (
|
||||
params.res_dir / f"wer-summary-{test_set_name}-{key}-{params.suffix}.txt"
|
||||
)
|
||||
errs_info = params.res_dir / f"wer-summary-{test_set_name}-{params.suffix}.txt"
|
||||
with open(errs_info, "w") as f:
|
||||
print("settings\tWER", file=f)
|
||||
for key, val in test_set_wers:
|
||||
|
||||
@ -568,18 +568,14 @@ def save_results(
|
||||
):
|
||||
test_set_wers = dict()
|
||||
for key, results in results_dict.items():
|
||||
recog_path = (
|
||||
params.res_dir / f"recogs-{test_set_name}-{key}-{params.suffix}.txt"
|
||||
)
|
||||
recog_path = params.res_dir / f"recogs-{test_set_name}-{params.suffix}.txt"
|
||||
results = sorted(results)
|
||||
store_transcripts(filename=recog_path, texts=results)
|
||||
logging.info(f"The transcripts are stored in {recog_path}")
|
||||
|
||||
# The following prints out WERs, per-word error statistics and aligned
|
||||
# ref/hyp pairs.
|
||||
errs_filename = (
|
||||
params.res_dir / f"errs-{test_set_name}-{key}-{params.suffix}.txt"
|
||||
)
|
||||
errs_filename = params.res_dir / f"errs-{test_set_name}-{params.suffix}.txt"
|
||||
with open(errs_filename, "w") as f:
|
||||
wer = write_error_stats(
|
||||
f, f"{test_set_name}-{key}", results, enable_log=True
|
||||
@ -589,9 +585,7 @@ def save_results(
|
||||
logging.info("Wrote detailed error stats to {}".format(errs_filename))
|
||||
|
||||
test_set_wers = sorted(test_set_wers.items(), key=lambda x: x[1])
|
||||
errs_info = (
|
||||
params.res_dir / f"wer-summary-{test_set_name}-{key}-{params.suffix}.txt"
|
||||
)
|
||||
errs_info = params.res_dir / f"wer-summary-{test_set_name}-{params.suffix}.txt"
|
||||
with open(errs_info, "w") as f:
|
||||
print("settings\tWER", file=f)
|
||||
for key, val in test_set_wers:
|
||||
|
||||
@ -409,18 +409,14 @@ def save_results(
|
||||
):
|
||||
test_set_wers = dict()
|
||||
for key, results in results_dict.items():
|
||||
recog_path = (
|
||||
params.res_dir / f"recogs-{test_set_name}-{key}-{params.suffix}.txt"
|
||||
)
|
||||
recog_path = params.res_dir / f"recogs-{test_set_name}-{params.suffix}.txt"
|
||||
results = sorted(results)
|
||||
store_transcripts(filename=recog_path, texts=results)
|
||||
logging.info(f"The transcripts are stored in {recog_path}")
|
||||
|
||||
# The following prints out WERs, per-word error statistics and aligned
|
||||
# ref/hyp pairs.
|
||||
errs_filename = (
|
||||
params.res_dir / f"errs-{test_set_name}-{key}-{params.suffix}.txt"
|
||||
)
|
||||
errs_filename = params.res_dir / f"errs-{test_set_name}-{params.suffix}.txt"
|
||||
with open(errs_filename, "w") as f:
|
||||
wer = write_error_stats(
|
||||
f, f"{test_set_name}-{key}", results, enable_log=True
|
||||
@ -430,9 +426,7 @@ def save_results(
|
||||
logging.info("Wrote detailed error stats to {}".format(errs_filename))
|
||||
|
||||
test_set_wers = sorted(test_set_wers.items(), key=lambda x: x[1])
|
||||
errs_info = (
|
||||
params.res_dir / f"wer-summary-{test_set_name}-{key}-{params.suffix}.txt"
|
||||
)
|
||||
errs_info = params.res_dir / f"wer-summary-{test_set_name}-{params.suffix}.txt"
|
||||
with open(errs_info, "w") as f:
|
||||
print("settings\tWER", file=f)
|
||||
for key, val in test_set_wers:
|
||||
|
||||
@ -594,18 +594,14 @@ def save_results(
|
||||
):
|
||||
test_set_wers = dict()
|
||||
for key, results in results_dict.items():
|
||||
recog_path = (
|
||||
params.res_dir / f"recogs-{test_set_name}-{key}-{params.suffix}.txt"
|
||||
)
|
||||
recog_path = params.res_dir / f"recogs-{test_set_name}-{params.suffix}.txt"
|
||||
results = sorted(results)
|
||||
store_transcripts(filename=recog_path, texts=results)
|
||||
logging.info(f"The transcripts are stored in {recog_path}")
|
||||
|
||||
# The following prints out WERs, per-word error statistics and aligned
|
||||
# ref/hyp pairs.
|
||||
errs_filename = (
|
||||
params.res_dir / f"errs-{test_set_name}-{key}-{params.suffix}.txt"
|
||||
)
|
||||
errs_filename = params.res_dir / f"errs-{test_set_name}-{params.suffix}.txt"
|
||||
with open(errs_filename, "w") as f:
|
||||
wer = write_error_stats(
|
||||
f, f"{test_set_name}-{key}", results, enable_log=True
|
||||
@ -615,9 +611,7 @@ def save_results(
|
||||
logging.info("Wrote detailed error stats to {}".format(errs_filename))
|
||||
|
||||
test_set_wers = sorted(test_set_wers.items(), key=lambda x: x[1])
|
||||
errs_info = (
|
||||
params.res_dir / f"wer-summary-{test_set_name}-{key}-{params.suffix}.txt"
|
||||
)
|
||||
errs_info = params.res_dir / f"wer-summary-{test_set_name}-{params.suffix}.txt"
|
||||
with open(errs_info, "w") as f:
|
||||
print("settings\tWER", file=f)
|
||||
for key, val in test_set_wers:
|
||||
|
||||
@ -325,18 +325,14 @@ def save_results(
|
||||
):
|
||||
test_set_wers = dict()
|
||||
for key, results in results_dict.items():
|
||||
recog_path = (
|
||||
params.res_dir / f"recogs-{test_set_name}-{key}-{params.suffix}.txt"
|
||||
)
|
||||
recog_path = params.res_dir / f"recogs-{test_set_name}-{params.suffix}.txt"
|
||||
results = sorted(results)
|
||||
store_transcripts(filename=recog_path, texts=results)
|
||||
logging.info(f"The transcripts are stored in {recog_path}")
|
||||
|
||||
# The following prints out WERs, per-word error statistics and aligned
|
||||
# ref/hyp pairs.
|
||||
errs_filename = (
|
||||
params.res_dir / f"errs-{test_set_name}-{key}-{params.suffix}.txt"
|
||||
)
|
||||
errs_filename = params.res_dir / f"errs-{test_set_name}-{params.suffix}.txt"
|
||||
with open(errs_filename, "w") as f:
|
||||
wer = write_error_stats(
|
||||
f, f"{test_set_name}-{key}", results, enable_log=True
|
||||
@ -346,9 +342,7 @@ def save_results(
|
||||
logging.info("Wrote detailed error stats to {}".format(errs_filename))
|
||||
|
||||
test_set_wers = sorted(test_set_wers.items(), key=lambda x: x[1])
|
||||
errs_info = (
|
||||
params.res_dir / f"wer-summary-{test_set_name}-{key}-{params.suffix}.txt"
|
||||
)
|
||||
errs_info = params.res_dir / f"wer-summary-{test_set_name}-{params.suffix}.txt"
|
||||
with open(errs_info, "w") as f:
|
||||
print("settings\tWER", file=f)
|
||||
for key, val in test_set_wers:
|
||||
|
||||
Some files were not shown because too many files have changed in this diff Show More
Loading…
x
Reference in New Issue
Block a user