mirrors/icefall
1
0
Fork 0
mirror of https://github.com/k2-fsa/icefall.git synced 2025-09-06 23:54:17 +00:00
Code Issues Packages Projects Releases Wiki Activity

add tests for aishell conformer_ctc

Browse Source
This commit is contained in:
Fangjun Kuang 2023-09-30 22:04:53 +08:00
parent cb15b9029c
commit f54a430a44
1 changed files with 79 additions and 1 deletions
Show Stats Download Patch File Download Diff File Expand all files Collapse all files

80
.github/scripts/run-pre-trained-conformer-ctc.sh vendored
View File

@ -8,7 +8,7 @@ log() {
echo -e "$(date '+%Y-%m-%d %H:%M:%S') (${fname}:${BASH_LINENO[0]}:${FUNCNAME[1]}) $*"
}
cd egs/librispeech/ASR
pushd egs/librispeech/ASR
# repo_url=https://github.com/csukuangfj/icefall-asr-conformer-ctc-bpe-500
repo_url=https://huggingface.co/csukuangfj/icefall-asr-librispeech-conformer-ctc-jit-bpe-500-2021-11-09
@ -112,3 +112,81 @@ log "Decoding with HLG on CPU with OpenFst"
$repo/test_wavs/1089-134686-0001.wav \
$repo/test_wavs/1221-135766-0001.wav \
$repo/test_wavs/1221-135766-0002.wav
rm -rf $repo
popd
log "Test aishell"
pushd egs/aishell/ASR
repo_url=https://huggingface.co/csukuangfj/icefall_asr_aishell_conformer_ctc
log "Downloading pre-trained model from $repo_url"
GIT_LFS_SKIP_SMUDGE=1 git clone $repo_url
repo=$(basename $repo_url)
pushd $repo
git lfs pull --include "exp/pretrained.pt"
git lfs pull --include "data/lm/G_3_gram_char.fst.txt"
popd
log "Display test files"
tree $repo/
ls -lh $repo/test_wavs/*.wav
log "CTC decoding"
log "Exporting model with torchscript"
pushd $repo/exp
ln -s pretrained.pt epoch-99.pt
popd
./conformer_ctc/export.py \
--epoch 99 \
--avg 1 \
--exp-dir $repo/exp \
--tokens $repo/data/lang_char/tokens.txt \
--jit 1
ls -lh $repo/exp
log "Generating H.fst, HL.fst"
./local/prepare_lang_fst.py --lang-dir $repo/data/lang_char --ngram-G $repo/data/lm/G_3_gram_char.fst.txt
ls -lh $repo/data/char
log "Decoding with H on CPU with OpenFst"
./conformer_ctc/jit_pretrained_decode_with_H.py \
--nn-model $repo/exp/cpu_jit.pt \
--H $repo/data/lang_char/H.fst \
--tokens $repo/data/lang_char/tokens.txt \
$repo/test_wavs/0.wav \
$repo/test_wavs/1.wav \
$repo/test_wavs/2.wav
log "Decoding with HL on CPU with OpenFst"
./conformer_ctc/jit_pretrained_decode_with_HL.py \
--nn-model $repo/exp/cpu_jit.pt \
--HL $repo/data/lang_char/HL.fst \
--words $repo/data/lang_char/words.txt \
$repo/test_wavs/0.wav \
$repo/test_wavs/1.wav \
$repo/test_wavs/2.wav
log "Decoding with HLG on CPU with OpenFst"
./conformer_ctc/jit_pretrained_decode_with_HLG.py \
--nn-model $repo/exp/cpu_jit.pt \
--HLG $repo/data/lang_char/HLG.fst \
--words $repo/data/lang_char/words.txt \
$repo/test_wavs/0.wav \
$repo/test_wavs/1.wav \
$repo/test_wavs/2.wav
rm -rf $repo