If you want to check the style of your code before git commit, you
can do the following:
-$ pre-commit install -$ pre-commit run +$ pre-commit install +$ pre-commit runOr without installing the pre-commit hooks:
-$ cd icefall -$ pip install black==22.3.0 flake8==5.0.4 isort==5.10.1 -$ black --check your_changed_file.py -$ black your_changed_file.py # modify it in-place +diff --git a/contributing/doc.html b/contributing/doc.html index 952852d91..dcba822e3 100644 --- a/contributing/doc.html +++ b/contributing/doc.html @@ -88,8 +88,8 @@ for documentation.$ cd icefall +$ pip install black==22.3.0 flake8==5.0.4 isort==5.10.1 +$ black --check your_changed_file.py +$ black your_changed_file.py # modify it in-place $ -$ flake8 your_changed_file.py +$ flake8 your_changed_file.py $ -$ isort --check your_changed_file.py # modify it in-place -$ isort your_changed_file.py +$ isort --check your_changed_file.py # modify it in-place +$ isort your_changed_file.pyBefore writing documentation, you have to prepare the environment:
-$ cd docs -$ pip install -r requirements.txt +@@ -99,16 +99,16 @@ if you are not familiar with$ cd docs +$ pip install -r requirements.txtAfter writing some documentation, you can build the documentation locally to preview what it looks like if it is published:-$ cd docs -$ make html +$ cd docs +$ make htmlThe generated documentation is in
docs/build/htmland can be viewed with the following commands:-$ cd docs/build/html -$ python3 -m http.server +diff --git a/contributing/how-to-create-a-recipe.html b/contributing/how-to-create-a-recipe.html index cfb4ec2ce..08dccf3d8 100644 --- a/contributing/how-to-create-a-recipe.html +++ b/contributing/how-to-create-a-recipe.html @@ -140,12 +140,12 @@ $ touch README.md model.py train.py decode.py asr_datamodule.py pretrained.py$ cd docs/build/html +$ python3 -m http.serverFor instance , the
yesnorecipe has atdnnmodel and its directory structure looks like the following:egs/yesno/ASR/tdnn/ -|-- README.md -|-- asr_datamodule.py -|-- decode.py -|-- model.py -|-- pretrained.py -`-- train.py +|-- README.md +|-- asr_datamodule.py +|-- decode.py +|-- model.py +|-- pretrained.py +`-- train.pyFile description:
diff --git a/installation/index.html b/installation/index.html index edf748e0c..fab7b095c 100644 --- a/installation/index.html +++ b/installation/index.html @@ -166,11 +166,11 @@ to installlhotsePYTHONPATHto point to it.Assume you want to place
-icefallin the folder/tmp. The following commands show you how to setupicefall:cd /tmp -git clone https://github.com/k2-fsa/icefall -cd icefall -pip install -r requirements.txt -export PYTHONPATH=/tmp/icefall:$PYTHONPATH +cd /tmp +git clone https://github.com/k2-fsa/icefall +cd icefall +pip install -r requirements.txt +export PYTHONPATH=/tmp/icefall:$PYTHONPATH@@ -185,39 +185,39 @@ to point to the version you want.The following shows an example about setting up the environment.
(1) Create a virtual environment
-$ virtualenv -p python3.8 test-icefall +$ virtualenv -p python3.8 test-icefall -created virtual environment CPython3.8.6.final.0-64 in 1540ms - creator CPython3Posix(dest=/ceph-fj/fangjun/test-icefall, clear=False, no_vcs_ignore=False, global=False) - seeder FromAppData(download=False, pip=bundle, setuptools=bundle, wheel=bundle, via=copy, app_data_dir=/root/fangjun/.local/share/v +created virtual environment CPython3.8.6.final.0-64 in 1540ms + creator CPython3Posix(dest=/ceph-fj/fangjun/test-icefall, clear=False, no_vcs_ignore=False, global=False) + seeder FromAppData(download=False, pip=bundle, setuptools=bundle, wheel=bundle, via=copy, app_data_dir=/root/fangjun/.local/share/v irtualenv) - added seed packages: pip==21.1.3, setuptools==57.4.0, wheel==0.36.2 - activators BashActivator,CShellActivator,FishActivator,PowerShellActivator,PythonActivator,XonshActivator + added seed packages: pip==21.1.3, setuptools==57.4.0, wheel==0.36.2 + activators BashActivator,CShellActivator,FishActivator,PowerShellActivator,PythonActivator,XonshActivator(2) Activate your virtual environment
-$ source test-icefall/bin/activate +$ source test-icefall/bin/activate(3) Install k2
-$ pip install k2==1.4.dev20210822+cpu.torch1.9.0 -f https://k2-fsa.org/nightly/index.html +$ pip install k2==1.4.dev20210822+cpu.torch1.9.0 -f https://k2-fsa.org/nightly/index.html -Looking in links: https://k2-fsa.org/nightly/index.html -Collecting k2==1.4.dev20210822+cpu.torch1.9.0 - Downloading https://k2-fsa.org/nightly/whl/k2-1.4.dev20210822%2Bcpu.torch1.9.0-cp38-cp38-linux_x86_64.whl (1.6 MB) - |________________________________| 1.6 MB 185 kB/s -Collecting graphviz - Downloading graphviz-0.17-py3-none-any.whl (18 kB) -Collecting torch==1.9.0 - Using cached torch-1.9.0-cp38-cp38-manylinux1_x86_64.whl (831.4 MB) -Collecting typing-extensions - Using cached typing_extensions-3.10.0.0-py3-none-any.whl (26 kB) -Installing collected packages: typing-extensions, torch, graphviz, k2 -Successfully installed graphviz-0.17 k2-1.4.dev20210822+cpu.torch1.9.0 torch-1.9.0 typing-extensions-3.10.0.0 +Looking in links: https://k2-fsa.org/nightly/index.html +Collecting k2==1.4.dev20210822+cpu.torch1.9.0 + Downloading https://k2-fsa.org/nightly/whl/k2-1.4.dev20210822%2Bcpu.torch1.9.0-cp38-cp38-linux_x86_64.whl (1.6 MB) + |________________________________| 1.6 MB 185 kB/s +Collecting graphviz + Downloading graphviz-0.17-py3-none-any.whl (18 kB) +Collecting torch==1.9.0 + Using cached torch-1.9.0-cp38-cp38-manylinux1_x86_64.whl (831.4 MB) +Collecting typing-extensions + Using cached typing_extensions-3.10.0.0-py3-none-any.whl (26 kB) +Installing collected packages: typing-extensions, torch, graphviz, k2 +Successfully installed graphviz-0.17 k2-1.4.dev20210822+cpu.torch1.9.0 torch-1.9.0 typing-extensions-3.10.0.0@@ -393,10 +393,10 @@ theData preparation
-$ export PYTHONPATH=/tmp/icefall:$PYTHONPATH -$ cd /tmp/icefall -$ cd egs/yesno/ASR -$ ./prepare.sh +$ export PYTHONPATH=/tmp/icefall:$PYTHONPATH +$ cd /tmp/icefall +$ cd egs/yesno/ASR +$ ./prepare.shThe log of running
@@ -457,7 +457,7 @@ even if there are GPUs available../prepare.shis:Hint
In case you get a
Segmentation fault (core dump)error, please use:--export PROTOCOL_BUFFERS_PYTHON_IMPLEMENTATION=python +diff --git a/model-export/export-model-state-dict.html b/model-export/export-model-state-dict.html index 7b7e0afc3..84dafe496 100644 --- a/model-export/export-model-state-dict.html +++ b/model-export/export-model-state-dict.html @@ -123,13 +123,13 @@ as an example.export PROTOCOL_BUFFERS_PYTHON_IMPLEMENTATION=pythonNote
The steps for other recipes are almost the same.
cd egs/librispeech/ASR +cd egs/librispeech/ASR -./pruned_transducer_stateless3/export.py \ - --exp-dir ./pruned_transducer_stateless3/exp \ - --bpe-model data/lang_bpe_500/bpe.model \ - --epoch 20 \ - --avg 10 +./pruned_transducer_stateless3/export.py \ + --exp-dir ./pruned_transducer_stateless3/exp \ + --bpe-model data/lang_bpe_500/bpe.model \ + --epoch 20 \ + --avg 10will generate a file
pruned_transducer_stateless3/exp/pretrained.pt, which @@ -141,10 +141,10 @@ is a dict containingRESULTS.mdof each dataset.In the following, we demonstrate how to use the pretrained model from https://huggingface.co/csukuangfj/icefall-asr-librispeech-pruned-transducer-stateless3-2022-05-13.
-cd egs/librispeech/ASR +cd egs/librispeech/ASR -git lfs install -git clone https://huggingface.co/csukuangfj/icefall-asr-librispeech-pruned-transducer-stateless3-2022-05-13 +git lfs install +git clone https://huggingface.co/csukuangfj/icefall-asr-librispeech-pruned-transducer-stateless3-2022-05-13After cloning the repo with
git lfs, you will find several files in the folder @@ -153,15 +153,15 @@ that have a prefixexported by the aboveexport.py.In each recipe, there is also a file
-pretrained.py, which can usepretrained-xxx.ptto decode waves. The following is an example:cd egs/librispeech/ASR +cd egs/librispeech/ASR -./pruned_transducer_stateless3/pretrained.py \ - --checkpoint ./icefall-asr-librispeech-pruned-transducer-stateless3-2022-05-13/exp/pretrained-iter-1224000-avg-14.pt \ - --bpe-model ./icefall-asr-librispeech-pruned-transducer-stateless3-2022-05-13/data/lang_bpe_500/bpe.model \ - --method greedy_search \ - ./icefall-asr-librispeech-pruned-transducer-stateless3-2022-05-13/test_wavs/1089-134686-0001.wav \ - ./icefall-asr-librispeech-pruned-transducer-stateless3-2022-05-13/test_wavs/1221-135766-0001.wav \ - ./icefall-asr-librispeech-pruned-transducer-stateless3-2022-05-13/test_wavs/1221-135766-0002.wav +./pruned_transducer_stateless3/pretrained.py \ + --checkpoint ./icefall-asr-librispeech-pruned-transducer-stateless3-2022-05-13/exp/pretrained-iter-1224000-avg-14.pt \ + --bpe-model ./icefall-asr-librispeech-pruned-transducer-stateless3-2022-05-13/data/lang_bpe_500/bpe.model \ + --method greedy_search \ + ./icefall-asr-librispeech-pruned-transducer-stateless3-2022-05-13/test_wavs/1089-134686-0001.wav \ + ./icefall-asr-librispeech-pruned-transducer-stateless3-2022-05-13/test_wavs/1221-135766-0001.wav \ + ./icefall-asr-librispeech-pruned-transducer-stateless3-2022-05-13/test_wavs/1221-135766-0002.wavThe above commands show how to use the exported model with
pretrained.pyto @@ -195,25 +195,25 @@ decode multiple sound files. Its output is given as follows for reference:When we publish the model, we always note down its WERs on some test dataset in
-RESULTS.md. This section describes how to use the pretrained model to reproduce the WER.cd egs/librispeech/ASR -git lfs install -git clone https://huggingface.co/csukuangfj/icefall-asr-librispeech-pruned-transducer-stateless3-2022-05-13 +cd egs/librispeech/ASR +git lfs install +git clone https://huggingface.co/csukuangfj/icefall-asr-librispeech-pruned-transducer-stateless3-2022-05-13 -cd icefall-asr-librispeech-pruned-transducer-stateless3-2022-05-13/exp -ln -s pretrained-iter-1224000-avg-14.pt epoch-9999.pt -cd ../.. +cd icefall-asr-librispeech-pruned-transducer-stateless3-2022-05-13/exp +ln -s pretrained-iter-1224000-avg-14.pt epoch-9999.pt +cd ../..We create a symlink with name
-epoch-9999.pttopretrained-iter-1224000-avg-14.pt, so that we can pass--epoch 9999 --avg 1todecode.pyin the following command:./pruned_transducer_stateless3/decode.py \ - --epoch 9999 \ - --avg 1 \ - --exp-dir ./icefall-asr-librispeech-pruned-transducer-stateless3-2022-05-13/exp \ - --lang-dir ./icefall-asr-librispeech-pruned-transducer-stateless3-2022-05-13/data/lang_bpe_500 \ - --max-duration 600 \ - --decoding-method greedy_search +./pruned_transducer_stateless3/decode.py \ + --epoch 9999 \ + --avg 1 \ + --exp-dir ./icefall-asr-librispeech-pruned-transducer-stateless3-2022-05-13/exp \ + --lang-dir ./icefall-asr-librispeech-pruned-transducer-stateless3-2022-05-13/data/lang_bpe_500 \ + --max-duration 600 \ + --decoding-method greedy_searchYou will find the decoding results in diff --git a/model-export/export-onnx.html b/model-export/export-onnx.html index a8362655f..9d3a8ef42 100644 --- a/model-export/export-onnx.html +++ b/model-export/export-onnx.html @@ -106,16 +106,16 @@ to run the pretrained model.
We use https://github.com/k2-fsa/icefall/tree/master/egs/librispeech/ASR/pruned_transducer_stateless3 as an example in the following.
-cd egs/librispeech/ASR +cd egs/librispeech/ASR epoch=14 avg=2 -./pruned_transducer_stateless3/export.py \ - --exp-dir ./pruned_transducer_stateless3/exp \ - --bpe-model data/lang_bpe_500/bpe.model \ - --epoch $epoch \ - --avg $avg \ - --onnx 1 +./pruned_transducer_stateless3/export.py \ + --exp-dir ./pruned_transducer_stateless3/exp \ + --bpe-model data/lang_bpe_500/bpe.model \ + --epoch $epoch \ + --avg $avg \ + --onnx 1It will generate the following files inside
@@ -130,16 +130,16 @@ as an example in the following.pruned_transducer_stateless3/exp:You can use
-./pruned_transducer_stateless3/exp/onnx_pretrained.pyto decode waves with the generated files:./pruned_transducer_stateless3/onnx_pretrained.py \ - --bpe-model ./data/lang_bpe_500/bpe.model \ - --encoder-model-filename ./pruned_transducer_stateless3/exp/encoder.onnx \ - --decoder-model-filename ./pruned_transducer_stateless3/exp/decoder.onnx \ - --joiner-model-filename ./pruned_transducer_stateless3/exp/joiner.onnx \ - --joiner-encoder-proj-model-filename ./pruned_transducer_stateless3/exp/joiner_encoder_proj.onnx \ - --joiner-decoder-proj-model-filename ./pruned_transducer_stateless3/exp/joiner_decoder_proj.onnx \ - /path/to/foo.wav \ - /path/to/bar.wav \ - /path/to/baz.wav +diff --git a/model-export/export-with-torch-jit-script.html b/model-export/export-with-torch-jit-script.html index 36714562b..ae0ee3fec 100644 --- a/model-export/export-with-torch-jit-script.html +++ b/model-export/export-with-torch-jit-script.html @@ -108,16 +108,16 @@ if you want to use./pruned_transducer_stateless3/onnx_pretrained.py \ + --bpe-model ./data/lang_bpe_500/bpe.model \ + --encoder-model-filename ./pruned_transducer_stateless3/exp/encoder.onnx \ + --decoder-model-filename ./pruned_transducer_stateless3/exp/decoder.onnx \ + --joiner-model-filename ./pruned_transducer_stateless3/exp/joiner.onnx \ + --joiner-encoder-proj-model-filename ./pruned_transducer_stateless3/exp/joiner_encoder_proj.onnx \ + --joiner-decoder-proj-model-filename ./pruned_transducer_stateless3/exp/joiner_decoder_proj.onnx \ + /path/to/foo.wav \ + /path/to/bar.wav \ + /path/to/baz.wavWe use https://github.com/k2-fsa/icefall/tree/master/egs/librispeech/ASR/pruned_transducer_stateless3 as an example in the following.
-cd egs/librispeech/ASR +cd egs/librispeech/ASR epoch=14 avg=1 -./pruned_transducer_stateless3/export.py \ - --exp-dir ./pruned_transducer_stateless3/exp \ - --bpe-model data/lang_bpe_500/bpe.model \ - --epoch $epoch \ - --avg $avg \ - --jit 1 +./pruned_transducer_stateless3/export.py \ + --exp-dir ./pruned_transducer_stateless3/exp \ + --bpe-model data/lang_bpe_500/bpe.model \ + --epoch $epoch \ + --avg $avg \ + --jit 1It will generate a file
diff --git a/model-export/export-with-torch-jit-trace.html b/model-export/export-with-torch-jit-trace.html index 4216fe87b..9215ca393 100644 --- a/model-export/export-with-torch-jit-trace.html +++ b/model-export/export-with-torch-jit-trace.html @@ -111,14 +111,14 @@ as an example in the following.cpu_jit.ptinpruned_transducer_stateless3/exp.iter=468000 avg=16 -cd egs/librispeech/ASR +cd egs/librispeech/ASR -./lstm_transducer_stateless2/export.py \ - --exp-dir ./lstm_transducer_stateless2/exp \ - --bpe-model data/lang_bpe_500/bpe.model \ - --iter $iter \ - --avg $avg \ - --jit-trace 1 +./lstm_transducer_stateless2/export.py \ + --exp-dir ./lstm_transducer_stateless2/exp \ + --bpe-model data/lang_bpe_500/bpe.model \ + --iter $iter \ + --avg $avg \ + --jit-trace 1It will generate three files inside
@@ -132,15 +132,15 @@ as an example in the following.lstm_transducer_stateless2/exp:You can use https://github.com/k2-fsa/icefall/blob/master/egs/librispeech/ASR/lstm_transducer_stateless2/jit_pretrained.py to decode sound files with the following commands:
-cd egs/librispeech/ASR -./lstm_transducer_stateless2/jit_pretrained.py \ - --bpe-model ./data/lang_bpe_500/bpe.model \ - --encoder-model-filename ./lstm_transducer_stateless2/exp/encoder_jit_trace.pt \ - --decoder-model-filename ./lstm_transducer_stateless2/exp/decoder_jit_trace.pt \ - --joiner-model-filename ./lstm_transducer_stateless2/exp/joiner_jit_trace.pt \ - /path/to/foo.wav \ - /path/to/bar.wav \ - /path/to/baz.wav +diff --git a/recipes/Non-streaming-ASR/aishell/conformer_ctc.html b/recipes/Non-streaming-ASR/aishell/conformer_ctc.html index f736b40f7..1689ac219 100644 --- a/recipes/Non-streaming-ASR/aishell/conformer_ctc.html +++ b/recipes/Non-streaming-ASR/aishell/conformer_ctc.html @@ -130,8 +130,8 @@ the environment forcd egs/librispeech/ASR +./lstm_transducer_stateless2/jit_pretrained.py \ + --bpe-model ./data/lang_bpe_500/bpe.model \ + --encoder-model-filename ./lstm_transducer_stateless2/exp/encoder_jit_trace.pt \ + --decoder-model-filename ./lstm_transducer_stateless2/exp/decoder_jit_trace.pt \ + --joiner-model-filename ./lstm_transducer_stateless2/exp/joiner_jit_trace.pt \ + /path/to/foo.wav \ + /path/to/bar.wav \ + /path/to/baz.wavData preparation
-$ cd egs/aishell/ASR -$ ./prepare.sh +$ cd egs/aishell/ASR +$ ./prepare.shThe script
./prepare.shhandles the data preparation for you, automagically. @@ -146,13 +146,13 @@ options:to control which stage(s) should be run. By default, all stages are executed.
For example,
-$ cd egs/aishell/ASR -$ ./prepare.sh --stage 0 --stop-stage 0 +$ cd egs/aishell/ASR +$ ./prepare.sh --stage 0 --stop-stage 0means to run only stage 0.
To run stage 2 to stage 5, use:
-$ ./prepare.sh --stage 2 --stop-stage 5 +$ ./prepare.sh --stage 2 --stop-stage 5@@ -167,8 +167,8 @@ thedl_dirHintA 3-gram language model will be downloaded from huggingface, we assume you have intalled and initialized
-git-lfs. If not, you could installgit-lfsby$ sudo apt-get install git-lfs -$ git-lfs install +$ sudo apt-get install git-lfs +$ git-lfs installIf you don’t have the
sudopermission, you could download the @@ -184,8 +184,8 @@ are saved in./dataTraining
Configurable options
-$ cd egs/aishell/ASR -$ ./conformer_ctc/train.py --help +$ cd egs/aishell/ASR +$ ./conformer_ctc/train.py --helpshows you the training options that can be passed from the commandline. @@ -227,26 +227,26 @@ training from epoch 10, based on the state from epoch 9.
Use case 1: You have 4 GPUs, but you only want to use GPU 0 and GPU 2 for training. You can do the following:
-$ cd egs/aishell/ASR -$ export CUDA_VISIBLE_DEVICES="0,2" -$ ./conformer_ctc/train.py --world-size 2 +$ cd egs/aishell/ASR +$ export CUDA_VISIBLE_DEVICES="0,2" +$ ./conformer_ctc/train.py --world-size 2Use case 2: You have 4 GPUs and you want to use all of them for training. You can do the following:
-$ cd egs/aishell/ASR -$ ./conformer_ctc/train.py --world-size 4 +$ cd egs/aishell/ASR +$ ./conformer_ctc/train.py --world-size 4Use case 3: You have 4 GPUs but you only want to use GPU 3 for training. You can do the following:
-$ cd egs/aishell/ASR -$ export CUDA_VISIBLE_DEVICES="3" -$ ./conformer_ctc/train.py --world-size 1 +@@ -299,7 +299,7 @@ Each epoch actually processes$ cd egs/aishell/ASR +$ export CUDA_VISIBLE_DEVICES="3" +$ ./conformer_ctc/train.py --world-size 1These are checkpoint files, containing modelstate_dictand optimizerstate_dict. To resume training from some checkpoint, sayepoch-10.pt, you can use:-$ ./conformer_ctc/train.py --start-epoch 11 +@@ -308,8 +308,8 @@ To resume training from some checkpoint, say$ ./conformer_ctc/train.py --start-epoch 11The following shows typical use cases:
Case 1
-$ cd egs/aishell/ASR -$ ./conformer_ctc/train.py --max-duration 200 +$ cd egs/aishell/ASR +$ ./conformer_ctc/train.py --max-duration 200It uses
--max-durationof 200 to avoid OOM.Case 2
-$ cd egs/aishell/ASR -$ export CUDA_VISIBLE_DEVICES="0,3" -$ ./conformer_ctc/train.py --world-size 2 +$ cd egs/aishell/ASR +$ export CUDA_VISIBLE_DEVICES="0,3" +$ ./conformer_ctc/train.py --world-size 2It uses GPU 0 and GPU 3 for DDP training.
Case 3
-$ cd egs/aishell/ASR -$ ./conformer_ctc/train.py --num-epochs 10 --start-epoch 3 +$ cd egs/aishell/ASR +$ ./conformer_ctc/train.py --num-epochs 10 --start-epoch 3It loads checkpoint
./conformer_ctc/exp/epoch-2.ptand starts @@ -381,8 +381,8 @@ training from epoch 3. Also, it trains for 10 epochs.Decoding
The decoding part uses checkpoints saved by the training part, so you have to run the training part first.
-$ cd egs/aishell/ASR -$ ./conformer_ctc/decode.py --help +$ cd egs/aishell/ASR +$ ./conformer_ctc/decode.py --helpshows the options for decoding.
@@ -440,27 +440,27 @@ $ git clone https://huggingface.co/pkufool/icefall_asr_aishell_conformer_ctcIn order to use this pre-trained model, your k2 version has to be v1.7 or later.
After downloading, you will have the following files:
-$ cd egs/aishell/ASR -$ tree tmp +$ cd egs/aishell/ASR +$ tree tmptmp/ -`-- icefall_asr_aishell_conformer_ctc - |-- README.md - |-- data - | `-- lang_char - | |-- HLG.pt - | |-- tokens.txt - | `-- words.txt - |-- exp - | `-- pretrained.pt - `-- test_waves - |-- BAC009S0764W0121.wav - |-- BAC009S0764W0122.wav - |-- BAC009S0764W0123.wav - `-- trans.txt +`-- icefall_asr_aishell_conformer_ctc + |-- README.md + |-- data + | `-- lang_char + | |-- HLG.pt + | |-- tokens.txt + | `-- words.txt + |-- exp + | `-- pretrained.pt + `-- test_waves + |-- BAC009S0764W0121.wav + |-- BAC009S0764W0122.wav + |-- BAC009S0764W0123.wav + `-- trans.txt -5 directories, 9 files +5 directories, 9 filesFile descriptions:
@@ -502,38 +502,38 @@ Note: We have removed optimizerThe information of the test sound files is listed below:
-$ soxi tmp/icefall_asr_aishell_conformer_ctc/test_waves/*.wav +@@ -556,14 +556,14 @@ $ ./conformer_ctc/pretrained.py --help$ soxi tmp/icefall_asr_aishell_conformer_ctc/test_waves/*.wav -Input File : 'tmp/icefall_asr_aishell_conformer_ctc/test_waves/BAC009S0764W0121.wav' -Channels : 1 -Sample Rate : 16000 -Precision : 16-bit -Duration : 00:00:04.20 = 67263 samples ~ 315.295 CDDA sectors -File Size : 135k -Bit Rate : 256k -Sample Encoding: 16-bit Signed Integer PCM +Input File : 'tmp/icefall_asr_aishell_conformer_ctc/test_waves/BAC009S0764W0121.wav' +Channels : 1 +Sample Rate : 16000 +Precision : 16-bit +Duration : 00:00:04.20 = 67263 samples ~ 315.295 CDDA sectors +File Size : 135k +Bit Rate : 256k +Sample Encoding: 16-bit Signed Integer PCM -Input File : 'tmp/icefall_asr_aishell_conformer_ctc/test_waves/BAC009S0764W0122.wav' -Channels : 1 -Sample Rate : 16000 -Precision : 16-bit -Duration : 00:00:04.12 = 65840 samples ~ 308.625 CDDA sectors -File Size : 132k -Bit Rate : 256k -Sample Encoding: 16-bit Signed Integer PCM +Input File : 'tmp/icefall_asr_aishell_conformer_ctc/test_waves/BAC009S0764W0122.wav' +Channels : 1 +Sample Rate : 16000 +Precision : 16-bit +Duration : 00:00:04.12 = 65840 samples ~ 308.625 CDDA sectors +File Size : 132k +Bit Rate : 256k +Sample Encoding: 16-bit Signed Integer PCM -Input File : 'tmp/icefall_asr_aishell_conformer_ctc/test_waves/BAC009S0764W0123.wav' -Channels : 1 -Sample Rate : 16000 -Precision : 16-bit -Duration : 00:00:04.00 = 64000 samples ~ 300 CDDA sectors -File Size : 128k -Bit Rate : 256k -Sample Encoding: 16-bit Signed Integer PCM +Input File : 'tmp/icefall_asr_aishell_conformer_ctc/test_waves/BAC009S0764W0123.wav' +Channels : 1 +Sample Rate : 16000 +Precision : 16-bit +Duration : 00:00:04.00 = 64000 samples ~ 300 CDDA sectors +File Size : 128k +Bit Rate : 256k +Sample Encoding: 16-bit Signed Integer PCM -Total Duration of 3 files: 00:00:12.32 +Total Duration of 3 files: 00:00:12.32CTC decoding
CTC decoding only uses the ctc topology for decoding without a lexicon and language model
The command to run CTC decoding is:
-$ cd egs/aishell/ASR -$ ./conformer_ctc/pretrained.py \ - --checkpoint ./tmp/icefall_asr_aishell_conformer_ctc/exp/pretrained.pt \ - --tokens-file ./tmp/icefall_asr_aishell_conformer_ctc/data/lang_char/tokens.txt \ - --method ctc-decoding \ - ./tmp/icefall_asr_aishell_conformer_ctc/test_waves/BAC009S0764W0121.wav \ - ./tmp/icefall_asr_aishell_conformer_ctc/test_waves/BAC009S0764W0122.wav \ - ./tmp/icefall_asr_aishell_conformer_ctc/test_waves/BAC009S0764W0123.wav +$ cd egs/aishell/ASR +$ ./conformer_ctc/pretrained.py \ + --checkpoint ./tmp/icefall_asr_aishell_conformer_ctc/exp/pretrained.pt \ + --tokens-file ./tmp/icefall_asr_aishell_conformer_ctc/data/lang_char/tokens.txt \ + --method ctc-decoding \ + ./tmp/icefall_asr_aishell_conformer_ctc/test_waves/BAC009S0764W0121.wav \ + ./tmp/icefall_asr_aishell_conformer_ctc/test_waves/BAC009S0764W0122.wav \ + ./tmp/icefall_asr_aishell_conformer_ctc/test_waves/BAC009S0764W0123.wavThe output is given below:
@@ -593,15 +593,15 @@ $ ./conformer_ctc/pretrained.py \HLG decoding
HLG decoding uses the best path of the decoding lattice as the decoding result.
The command to run HLG decoding is:
-$ cd egs/aishell/ASR -$ ./conformer_ctc/pretrained.py \ - --checkpoint ./tmp/icefall_asr_aishell_conformer_ctc/exp/pretrained.pt \ - --words-file ./tmp/icefall_asr_aishell_conformer_ctc/data/lang_char/words.txt \ - --HLG ./tmp/icefall_asr_aishell_conformer_ctc/data/lang_char/HLG.pt \ - --method 1best \ - ./tmp/icefall_asr_aishell_conformer_ctc/test_waves/BAC009S0764W0121.wav \ - ./tmp/icefall_asr_aishell_conformer_ctc/test_waves/BAC009S0764W0122.wav \ - ./tmp/icefall_asr_aishell_conformer_ctc/test_waves/BAC009S0764W0123.wav +$ cd egs/aishell/ASR +$ ./conformer_ctc/pretrained.py \ + --checkpoint ./tmp/icefall_asr_aishell_conformer_ctc/exp/pretrained.pt \ + --words-file ./tmp/icefall_asr_aishell_conformer_ctc/data/lang_char/words.txt \ + --HLG ./tmp/icefall_asr_aishell_conformer_ctc/data/lang_char/HLG.pt \ + --method 1best \ + ./tmp/icefall_asr_aishell_conformer_ctc/test_waves/BAC009S0764W0121.wav \ + ./tmp/icefall_asr_aishell_conformer_ctc/test_waves/BAC009S0764W0122.wav \ + ./tmp/icefall_asr_aishell_conformer_ctc/test_waves/BAC009S0764W0123.wavThe output is given below:
@@ -633,15 +633,15 @@ $ ./conformer_ctc/pretrained.py \It extracts n paths from the lattice, recores the extracted paths with an attention decoder. The path with the highest score is the decoding result.
The command to run HLG decoding + attention decoder rescoring is:
-$ cd egs/aishell/ASR -$ ./conformer_ctc/pretrained.py \ - --checkpoint ./tmp/icefall_asr_aishell_conformer_ctc/exp/pretrained.pt \ - --words-file ./tmp/icefall_asr_aishell_conformer_ctc/data/lang_char/words.txt \ - --HLG ./tmp/icefall_asr_aishell_conformer_ctc/data/lang_char/HLG.pt \ - --method attention-decoder \ - ./tmp/icefall_asr_aishell_conformer_ctc/test_waves/BAC009S0764W0121.wav \ - ./tmp/icefall_asr_aishell_conformer_ctc/test_waves/BAC009S0764W0122.wav \ - ./tmp/icefall_asr_aishell_conformer_ctc/test_waves/BAC009S0764W0123.wav +$ cd egs/aishell/ASR +$ ./conformer_ctc/pretrained.py \ + --checkpoint ./tmp/icefall_asr_aishell_conformer_ctc/exp/pretrained.pt \ + --words-file ./tmp/icefall_asr_aishell_conformer_ctc/data/lang_char/words.txt \ + --HLG ./tmp/icefall_asr_aishell_conformer_ctc/data/lang_char/HLG.pt \ + --method attention-decoder \ + ./tmp/icefall_asr_aishell_conformer_ctc/test_waves/BAC009S0764W0121.wav \ + ./tmp/icefall_asr_aishell_conformer_ctc/test_waves/BAC009S0764W0122.wav \ + ./tmp/icefall_asr_aishell_conformer_ctc/test_waves/BAC009S0764W0123.wavThe output is below:
@@ -693,20 +693,20 @@ Python dependencies.At present, it does NOT support streaming decoding.
First, let us compile k2 from source:
-$ cd $HOME -$ git clone https://github.com/k2-fsa/k2 -$ cd k2 -$ git checkout v2.0-pre +$ cd $HOME +$ git clone https://github.com/k2-fsa/k2 +$ cd k2 +$ git checkout v2.0-pre-Caution
You have to switch to the branch
v2.0-pre!$ mkdir build-release -$ cd build-release -$ cmake -DCMAKE_BUILD_TYPE=Release .. -$ make -j hlg_decode +@@ -714,8 +714,8 @@ $ make -j hlg_decode$ mkdir build-release +$ cd build-release +$ cmake -DCMAKE_BUILD_TYPE=Release .. +$ make -j hlg_decode # You will find four binaries in `./bin`, i.e. ./bin/hlg_decode,Now you are ready to go!
Assume you have run:
-$ cd k2/build-release -$ ln -s /path/to/icefall-asr-aishell-conformer-ctc ./ +@@ -724,40 +724,40 @@ $ ln -s /path/to/icefall-asr-aishell-conformer-ctc ./$ cd k2/build-release +$ ln -s /path/to/icefall-asr-aishell-conformer-ctc ./It will show you the following message:
-Please provide --nn_model +Please provide --nn_model -This file implements decoding with an HLG decoding graph. +This file implements decoding with an HLG decoding graph. Usage: - ./bin/hlg_decode \ - --use_gpu true \ - --nn_model <path to torch scripted pt file> \ - --hlg <path to HLG.pt> \ - --word_table <path to words.txt> \ - <path to foo.wav> \ - <path to bar.wav> \ - <more waves if any> + ./bin/hlg_decode \ + --use_gpu true \ + --nn_model <path to torch scripted pt file> \ + --hlg <path to HLG.pt> \ + --word_table <path to words.txt> \ + <path to foo.wav> \ + <path to bar.wav> \ + <more waves if any> -To see all possible options, use - ./bin/hlg_decode --help +To see all possible options, use + ./bin/hlg_decode --help Caution: - - Only sound files (*.wav) with single channel are supported. - - It assumes the model is conformer_ctc/transformer.py from icefall. - If you use a different model, you have to change the code - related to `model.forward` in this file. + - Only sound files (*.wav) with single channel are supported. + - It assumes the model is conformer_ctc/transformer.py from icefall. + If you use a different model, you have to change the code + related to `model.forward` in this file.HLG decoding
-./bin/hlg_decode \ - --use_gpu true \ - --nn_model icefall_asr_aishell_conformer_ctc/exp/cpu_jit.pt \ - --hlg icefall_asr_aishell_conformer_ctc/data/lang_char/HLG.pt \ - --word_table icefall_asr_aishell_conformer_ctc/data/lang_char/words.txt \ - icefall_asr_aishell_conformer_ctc/test_waves/BAC009S0764W0121.wav \ - icefall_asr_aishell_conformer_ctc/test_waves/BAC009S0764W0122.wav \ - icefall_asr_aishell_conformer_ctc/test_waves/BAC009S0764W0123.wav +./bin/hlg_decode \ + --use_gpu true \ + --nn_model icefall_asr_aishell_conformer_ctc/exp/cpu_jit.pt \ + --hlg icefall_asr_aishell_conformer_ctc/data/lang_char/HLG.pt \ + --word_table icefall_asr_aishell_conformer_ctc/data/lang_char/words.txt \ + icefall_asr_aishell_conformer_ctc/test_waves/BAC009S0764W0121.wav \ + icefall_asr_aishell_conformer_ctc/test_waves/BAC009S0764W0122.wav \ + icefall_asr_aishell_conformer_ctc/test_waves/BAC009S0764W0123.wavThe output is:
diff --git a/recipes/Non-streaming-ASR/aishell/stateless_transducer.html b/recipes/Non-streaming-ASR/aishell/stateless_transducer.html index f3c8456d9..8f6aaa22a 100644 --- a/recipes/Non-streaming-ASR/aishell/stateless_transducer.html +++ b/recipes/Non-streaming-ASR/aishell/stateless_transducer.html @@ -211,9 +211,9 @@ alternatives.Data Preparation
To prepare the data for training, please use the following commands:
-cd egs/aishell/ASR -./prepare.sh --stop-stage 4 -./prepare.sh --stage 6 --stop-stage 6 +cd egs/aishell/ASR +./prepare.sh --stop-stage 4 +./prepare.sh --stage 6 --stop-stage 6@@ -231,8 +231,8 @@ are not used in transducer training.Training
-cd egs/aishell/ASR -./transducer_stateless_modified/train.py --help +cd egs/aishell/ASR +./transducer_stateless_modified/train.py --helpshows you the training options that can be passed from the commandline. @@ -274,26 +274,26 @@ training from epoch 10, based on the state from epoch 9.
Use case 1: You have 4 GPUs, but you only want to use GPU 0 and GPU 2 for training. You can do the following:
-$ cd egs/aishell/ASR -$ export CUDA_VISIBLE_DEVICES="0,2" -$ ./transducer_stateless_modified/train.py --world-size 2 +$ cd egs/aishell/ASR +$ export CUDA_VISIBLE_DEVICES="0,2" +$ ./transducer_stateless_modified/train.py --world-size 2Use case 2: You have 4 GPUs and you want to use all of them for training. You can do the following:
-$ cd egs/aishell/ASR -$ ./transducer_stateless_modified/train.py --world-size 4 +$ cd egs/aishell/ASR +$ ./transducer_stateless_modified/train.py --world-size 4Use case 3: You have 4 GPUs but you only want to use GPU 3 for training. You can do the following:
-$ cd egs/aishell/ASR -$ export CUDA_VISIBLE_DEVICES="3" -$ ./transducer_stateless_modified/train.py --world-size 1 +@@ -358,7 +358,7 @@ Each epoch actually processes$ cd egs/aishell/ASR +$ export CUDA_VISIBLE_DEVICES="3" +$ ./transducer_stateless_modified/train.py --world-size 1These are checkpoint files, containing modelstate_dictand optimizerstate_dict. To resume training from some checkpoint, sayepoch-10.pt, you can use:-$ ./transducer_stateless_modified/train.py --start-epoch 11 +@@ -367,8 +367,8 @@ To resume training from some checkpoint, say$ ./transducer_stateless_modified/train.py --start-epoch 11The following shows typical use cases:
Case 1
-$ cd egs/aishell/ASR -$ ./transducer_stateless_modified/train.py --max-duration 250 +$ cd egs/aishell/ASR +$ ./transducer_stateless_modified/train.py --max-duration 250It uses
--max-durationof 250 to avoid OOM.Case 2
-$ cd egs/aishell/ASR -$ export CUDA_VISIBLE_DEVICES="0,3" -$ ./transducer_stateless_modified/train.py --world-size 2 +$ cd egs/aishell/ASR +$ export CUDA_VISIBLE_DEVICES="0,3" +$ ./transducer_stateless_modified/train.py --world-size 2It uses GPU 0 and GPU 3 for DDP training.
Case 3
-$ cd egs/aishell/ASR -$ ./transducer_stateless_modified/train.py --num-epochs 10 --start-epoch 3 +$ cd egs/aishell/ASR +$ ./transducer_stateless_modified/train.py --num-epochs 10 --start-epoch 3It loads checkpoint
./transducer_stateless_modified/exp/epoch-2.ptand starts @@ -440,8 +440,8 @@ training from epoch 3. Also, it trains for 10 epochs.Decoding
The decoding part uses checkpoints saved by the training part, so you have to run the training part first.
-$ cd egs/aishell/ASR -$ ./transducer_stateless_modified/decode.py --help +$ cd egs/aishell/ASR +$ ./transducer_stateless_modified/decode.py --helpshows the options for decoding.
@@ -539,34 +539,34 @@ $ git clone https://huggingface.co/csukuangfj/icefall-aishell-transducer-stateleYou have to use
git lfsto download the pre-trained model.After downloading, you will have the following files:
-$ cd egs/aishell/ASR -$ tree tmp/icefall-aishell-transducer-stateless-modified-2022-03-01 +$ cd egs/aishell/ASR +$ tree tmp/icefall-aishell-transducer-stateless-modified-2022-03-01tmp/icefall-aishell-transducer-stateless-modified-2022-03-01/ -|-- README.md -|-- data -| `-- lang_char -| |-- L.pt -| |-- lexicon.txt -| |-- tokens.txt -| `-- words.txt -|-- exp -| `-- pretrained.pt -|-- log -| |-- errs-test-beam_4-epoch-64-avg-33-beam-4.txt -| |-- errs-test-greedy_search-epoch-64-avg-33-context-2-max-sym-per-frame-1.txt -| |-- log-decode-epoch-64-avg-33-beam-4-2022-03-02-12-05-03 -| |-- log-decode-epoch-64-avg-33-context-2-max-sym-per-frame-1-2022-02-28-18-13-07 -| |-- recogs-test-beam_4-epoch-64-avg-33-beam-4.txt -| `-- recogs-test-greedy_search-epoch-64-avg-33-context-2-max-sym-per-frame-1.txt -`-- test_wavs - |-- BAC009S0764W0121.wav - |-- BAC009S0764W0122.wav - |-- BAC009S0764W0123.wav - `-- transcript.txt +|-- README.md +|-- data +| `-- lang_char +| |-- L.pt +| |-- lexicon.txt +| |-- tokens.txt +| `-- words.txt +|-- exp +| `-- pretrained.pt +|-- log +| |-- errs-test-beam_4-epoch-64-avg-33-beam-4.txt +| |-- errs-test-greedy_search-epoch-64-avg-33-context-2-max-sym-per-frame-1.txt +| |-- log-decode-epoch-64-avg-33-beam-4-2022-03-02-12-05-03 +| |-- log-decode-epoch-64-avg-33-context-2-max-sym-per-frame-1-2022-02-28-18-13-07 +| |-- recogs-test-beam_4-epoch-64-avg-33-beam-4.txt +| `-- recogs-test-greedy_search-epoch-64-avg-33-context-2-max-sym-per-frame-1.txt +`-- test_wavs + |-- BAC009S0764W0121.wav + |-- BAC009S0764W0122.wav + |-- BAC009S0764W0123.wav + `-- transcript.txt -5 directories, 16 files +5 directories, 16 filesFile descriptions:
@@ -595,38 +595,38 @@ Note: We have removed optimizerThe information of the test sound files is listed below:
-$ soxi tmp/icefall-aishell-transducer-stateless-modified-2022-03-01/test_wavs/*.wav +@@ -655,14 +655,14 @@ it may give you poor results.$ soxi tmp/icefall-aishell-transducer-stateless-modified-2022-03-01/test_wavs/*.wav -Input File : 'tmp/icefall-aishell-transducer-stateless-modified-2022-03-01/test_wavs/BAC009S0764W0121.wav' -Channels : 1 -Sample Rate : 16000 -Precision : 16-bit -Duration : 00:00:04.20 = 67263 samples ~ 315.295 CDDA sectors -File Size : 135k -Bit Rate : 256k -Sample Encoding: 16-bit Signed Integer PCM +Input File : 'tmp/icefall-aishell-transducer-stateless-modified-2022-03-01/test_wavs/BAC009S0764W0121.wav' +Channels : 1 +Sample Rate : 16000 +Precision : 16-bit +Duration : 00:00:04.20 = 67263 samples ~ 315.295 CDDA sectors +File Size : 135k +Bit Rate : 256k +Sample Encoding: 16-bit Signed Integer PCM -Input File : 'tmp/icefall-aishell-transducer-stateless-modified-2022-03-01/test_wavs/BAC009S0764W0122.wav' -Channels : 1 -Sample Rate : 16000 -Precision : 16-bit -Duration : 00:00:04.12 = 65840 samples ~ 308.625 CDDA sectors -File Size : 132k -Bit Rate : 256k -Sample Encoding: 16-bit Signed Integer PCM +Input File : 'tmp/icefall-aishell-transducer-stateless-modified-2022-03-01/test_wavs/BAC009S0764W0122.wav' +Channels : 1 +Sample Rate : 16000 +Precision : 16-bit +Duration : 00:00:04.12 = 65840 samples ~ 308.625 CDDA sectors +File Size : 132k +Bit Rate : 256k +Sample Encoding: 16-bit Signed Integer PCM -Input File : 'tmp/icefall-aishell-transducer-stateless-modified-2022-03-01/test_wavs/BAC009S0764W0123.wav' -Channels : 1 -Sample Rate : 16000 -Precision : 16-bit -Duration : 00:00:04.00 = 64000 samples ~ 300 CDDA sectors -File Size : 128k -Bit Rate : 256k -Sample Encoding: 16-bit Signed Integer PCM +Input File : 'tmp/icefall-aishell-transducer-stateless-modified-2022-03-01/test_wavs/BAC009S0764W0123.wav' +Channels : 1 +Sample Rate : 16000 +Precision : 16-bit +Duration : 00:00:04.00 = 64000 samples ~ 300 CDDA sectors +File Size : 128k +Bit Rate : 256k +Sample Encoding: 16-bit Signed Integer PCM -Total Duration of 3 files: 00:00:12.32 +Total Duration of 3 files: 00:00:12.32Greedy search
The command to run greedy search is given below:
-$ cd egs/aishell/ASR -$ ./transducer_stateless_modified/pretrained.py \ - --checkpoint ./tmp/icefall-aishell-transducer-stateless-modified-2022-03-01/exp/pretrained.pt \ - --lang-dir ./tmp/icefall-aishell-transducer-stateless-modified-2022-03-01/data/lang_char \ - --method greedy_search \ - ./tmp/icefall-aishell-transducer-stateless-modified-2022-03-01/test_wavs/BAC009S0764W0121.wav \ - ./tmp/icefall-aishell-transducer-stateless-modified-2022-03-01/test_wavs/BAC009S0764W0122.wav \ - ./tmp/icefall-aishell-transducer-stateless-modified-2022-03-01/test_wavs/BAC009S0764W0123.wav +$ cd egs/aishell/ASR +$ ./transducer_stateless_modified/pretrained.py \ + --checkpoint ./tmp/icefall-aishell-transducer-stateless-modified-2022-03-01/exp/pretrained.pt \ + --lang-dir ./tmp/icefall-aishell-transducer-stateless-modified-2022-03-01/data/lang_char \ + --method greedy_search \ + ./tmp/icefall-aishell-transducer-stateless-modified-2022-03-01/test_wavs/BAC009S0764W0121.wav \ + ./tmp/icefall-aishell-transducer-stateless-modified-2022-03-01/test_wavs/BAC009S0764W0122.wav \ + ./tmp/icefall-aishell-transducer-stateless-modified-2022-03-01/test_wavs/BAC009S0764W0123.wavThe output is as follows:
@@ -692,16 +692,16 @@ $ ./transducer_stateless_modified/pretrained.py \Beam search
The command to run beam search is given below:
-$ cd egs/aishell/ASR +$ cd egs/aishell/ASR -$ ./transducer_stateless_modified/pretrained.py \ - --checkpoint ./tmp/icefall-aishell-transducer-stateless-modified-2022-03-01/exp/pretrained.pt \ - --lang-dir ./tmp/icefall-aishell-transducer-stateless-modified-2022-03-01/data/lang_char \ - --method beam_search \ - --beam-size 4 \ - ./tmp/icefall-aishell-transducer-stateless-modified-2022-03-01/test_wavs/BAC009S0764W0121.wav \ - ./tmp/icefall-aishell-transducer-stateless-modified-2022-03-01/test_wavs/BAC009S0764W0122.wav \ - ./tmp/icefall-aishell-transducer-stateless-modified-2022-03-01/test_wavs/BAC009S0764W0123.wav +$ ./transducer_stateless_modified/pretrained.py \ + --checkpoint ./tmp/icefall-aishell-transducer-stateless-modified-2022-03-01/exp/pretrained.pt \ + --lang-dir ./tmp/icefall-aishell-transducer-stateless-modified-2022-03-01/data/lang_char \ + --method beam_search \ + --beam-size 4 \ + ./tmp/icefall-aishell-transducer-stateless-modified-2022-03-01/test_wavs/BAC009S0764W0121.wav \ + ./tmp/icefall-aishell-transducer-stateless-modified-2022-03-01/test_wavs/BAC009S0764W0122.wav \ + ./tmp/icefall-aishell-transducer-stateless-modified-2022-03-01/test_wavs/BAC009S0764W0123.wavThe output is as follows:
@@ -731,16 +731,16 @@ $ ./transducer_stateless_modified/pretrained.py \Modified Beam search
The command to run modified beam search is given below:
-$ cd egs/aishell/ASR +$ cd egs/aishell/ASR -$ ./transducer_stateless_modified/pretrained.py \ - --checkpoint ./tmp/icefall-aishell-transducer-stateless-modified-2022-03-01/exp/pretrained.pt \ - --lang-dir ./tmp/icefall-aishell-transducer-stateless-modified-2022-03-01/data/lang_char \ - --method modified_beam_search \ - --beam-size 4 \ - ./tmp/icefall-aishell-transducer-stateless-modified-2022-03-01/test_wavs/BAC009S0764W0121.wav \ - ./tmp/icefall-aishell-transducer-stateless-modified-2022-03-01/test_wavs/BAC009S0764W0122.wav \ - ./tmp/icefall-aishell-transducer-stateless-modified-2022-03-01/test_wavs/BAC009S0764W0123.wav +$ ./transducer_stateless_modified/pretrained.py \ + --checkpoint ./tmp/icefall-aishell-transducer-stateless-modified-2022-03-01/exp/pretrained.pt \ + --lang-dir ./tmp/icefall-aishell-transducer-stateless-modified-2022-03-01/data/lang_char \ + --method modified_beam_search \ + --beam-size 4 \ + ./tmp/icefall-aishell-transducer-stateless-modified-2022-03-01/test_wavs/BAC009S0764W0121.wav \ + ./tmp/icefall-aishell-transducer-stateless-modified-2022-03-01/test_wavs/BAC009S0764W0122.wav \ + ./tmp/icefall-aishell-transducer-stateless-modified-2022-03-01/test_wavs/BAC009S0764W0123.wavThe output is as follows:
diff --git a/recipes/Non-streaming-ASR/aishell/tdnn_lstm_ctc.html b/recipes/Non-streaming-ASR/aishell/tdnn_lstm_ctc.html index 868290366..d3e88a5d0 100644 --- a/recipes/Non-streaming-ASR/aishell/tdnn_lstm_ctc.html +++ b/recipes/Non-streaming-ASR/aishell/tdnn_lstm_ctc.html @@ -130,8 +130,8 @@ the environment forData preparation
-$ cd egs/aishell/ASR -$ ./prepare.sh +$ cd egs/aishell/ASR +$ ./prepare.shThe script
./prepare.shhandles the data preparation for you, automagically. @@ -146,13 +146,13 @@ options:to control which stage(s) should be run. By default, all stages are executed.
For example,
-$ cd egs/aishell/ASR -$ ./prepare.sh --stage 0 --stop-stage 0 +$ cd egs/aishell/ASR +$ ./prepare.sh --stage 0 --stop-stage 0means to run only stage 0.
To run stage 2 to stage 5, use:
-$ ./prepare.sh --stage 2 --stop-stage 5 +$ ./prepare.sh --stage 2 --stop-stage 5@@ -167,8 +167,8 @@ thedl_dirHintA 3-gram language model will be downloaded from huggingface, we assume you have intalled and initialized
-git-lfs. If not, you could installgit-lfsby