mirrors/icefall
1
0
Fork 0
mirror of https://github.com/k2-fsa/icefall.git synced 2025-12-11 06:55:27 +00:00
Code Issues Packages Projects Releases Wiki Activity

Minor fixes

Browse Source
This commit is contained in:
pkufool 2023-06-16 17:13:16 +08:00
parent 78ef1d1874
commit 3ef74b0630
2 changed files with 11 additions and 11 deletions
Show Stats Download Patch File Download Diff File Expand all files Collapse all files

2
egs/librispeech/ASR/zipformer/export-onnx-streaming.py
View File

@ -145,7 +145,7 @@ def get_parser():
"--tokens", "--tokens",
type=str, type=str,
default="data/lang_bpe_500/tokens.txt", default="data/lang_bpe_500/tokens.txt",
help="Path to the BPE model", help="Path to the tokens.txt",
) )
parser.add_argument( parser.add_argument(

20
egs/wenetspeech/ASR/zipformer/onnx_decode.py
View File

@ -22,29 +22,29 @@
This script loads ONNX exported models and uses them to decode the test sets. This script loads ONNX exported models and uses them to decode the test sets.
We use the pre-trained model from We use the pre-trained model from
https://huggingface.co/csukuangfj/icefall-asr-librispeech-pruned-transducer-stateless7-2022-11-11 https://huggingface.co/pkufool/icefall-asr-zipformer-wenetspeech-20230615
as an example to show how to use this file. as an example to show how to use this file.
1. Download the pre-trained model 1. Download the pre-trained model
cd egs/librispeech/ASR cd egs/wenetspeech/ASR
repo_url=https://huggingface.co/csukuangfj/icefall-asr-librispeech-pruned-transducer-stateless7-2022-11-11 repo_url=https://huggingface.co/pkufool/icefall-asr-zipformer-wenetspeech-20230615
GIT_LFS_SKIP_SMUDGE=1 git clone $repo_url GIT_LFS_SKIP_SMUDGE=1 git clone $repo_url
repo=$(basename $repo_url) repo=$(basename $repo_url)
pushd $repo pushd $repo
git lfs pull --include "data/lang_bpe_500/bpe.model" git lfs pull --include "data/lang_char/tokens.txt"
git lfs pull --include "exp/pretrained-epoch-30-avg-9.pt" git lfs pull --include "exp/pretrained.pt"
cd exp cd exp
ln -s pretrained-epoch-30-avg-9.pt epoch-9999.pt ln -s pretrained.pt epoch-9999.pt
popd popd
2. Export the model to ONNX 2. Export the model to ONNX
./pruned_transducer_stateless7/export-onnx.py \ ./zipformer/export-onnx.py \
--bpe-model $repo/data/lang_bpe_500/bpe.model \ --tokens $repo/data/lang_char/tokens.txt \
--epoch 9999 \ --epoch 9999 \
--avg 1 \ --avg 1 \
--exp-dir $repo/exp/ --exp-dir $repo/exp/
@ -57,8 +57,8 @@ It will generate the following 3 files inside $repo/exp:
2. Run this file 2. Run this file
./pruned_transducer_stateless7/onnx_decode.py \ ./zipformer/onnx_decode.py \
--exp-dir ./pruned_transducer_stateless7/exp \ --exp-dir ./zipformer/exp \
--max-duration 600 \ --max-duration 600 \
--encoder-model-filename $repo/exp/encoder-epoch-9999-avg-1.onnx \ --encoder-model-filename $repo/exp/encoder-epoch-9999-avg-1.onnx \
--decoder-model-filename $repo/exp/decoder-epoch-9999-avg-1.onnx \ --decoder-model-filename $repo/exp/decoder-epoch-9999-avg-1.onnx \