diff --git a/_sources/decoding-with-langugage-models/rescoring.rst.txt b/_sources/decoding-with-langugage-models/rescoring.rst.txt index de7e700d0..ee2e2113c 100644 --- a/_sources/decoding-with-langugage-models/rescoring.rst.txt +++ b/_sources/decoding-with-langugage-models/rescoring.rst.txt @@ -4,11 +4,7 @@ LM rescoring for Transducer ================================= LM rescoring is a commonly used approach to incorporate external LM information. Unlike shallow-fusion-based -<<<<<<< HEAD methods (see :ref:`shallow_fusion`, :ref:`LODR`), rescoring is usually performed to re-rank the n-best hypotheses after beam search. -======= -methods (see :ref:`shallow-fusion`, :ref:`LODR`), rescoring is usually performed to re-rank the n-best hypotheses after beam search. ->>>>>>> 80d922c1583b9b7fb7e9b47008302cdc74ef58b7 Rescoring is usually more efficient than shallow fusion since less computation is performed on the external LM. In this tutorial, we will show you how to use external LM to rescore the n-best hypotheses decoded from neural transducer models in `icefall `__. diff --git a/_sources/installation/index.rst.txt b/_sources/installation/index.rst.txt index 738b24ab2..534b674f9 100644 --- a/_sources/installation/index.rst.txt +++ b/_sources/installation/index.rst.txt @@ -3,40 +3,23 @@ Installation ============ +.. hint:: + We have a colab notebook guiding you step by step to setup the environment. -``icefall`` depends on `k2 `_ and -`lhotse `_. + |yesno colab notebook| + + .. |yesno colab notebook| image:: https://colab.research.google.com/assets/colab-badge.svg + :target: https://colab.research.google.com/drive/1tIjjzaJc3IvGyKiMCDWO-TSnBgkcuN3B?usp=sharing + +`icefall`_ depends on `k2`_ and `lhotse`_. We recommend that you use the following steps to install the dependencies. - (0) Install CUDA toolkit and cuDNN -- (1) Install PyTorch and torchaudio -- (2) Install k2 -- (3) Install lhotse - -.. caution:: - - 99% users who have issues about the installation are using conda. - -.. caution:: - - 99% users who have issues about the installation are using conda. - -.. caution:: - - 99% users who have issues about the installation are using conda. - -.. hint:: - - We suggest that you use ``pip install`` to install PyTorch. - - You can use the following command to create a virutal environment in Python: - - .. code-block:: bash - - python3 -m venv ./my_env - source ./my_env/bin/activate +- (1) Install `torch`_ and `torchaudio`_ +- (2) Install `k2`_ +- (3) Install `lhotse`_ .. caution:: @@ -50,27 +33,20 @@ Please refer to to install CUDA and cuDNN. -(1) Install PyTorch and torchaudio ----------------------------------- +(1) Install torch and torchaudio +-------------------------------- -Please refer ``_ to install PyTorch -and torchaudio. - -.. hint:: - - You can also go to ``_ - to download pre-compiled wheels and install them. +Please refer ``_ to install `torch`_ and `torchaudio`_. .. caution:: Please install torch and torchaudio at the same time. - (2) Install k2 -------------- Please refer to ``_ -to install ``k2``. +to install `k2`_. .. caution:: @@ -78,21 +54,18 @@ to install ``k2``. .. note:: - We suggest that you install k2 from source by following - ``_ - or - ``_. + We suggest that you install k2 from pre-compiled wheels by following + ``_ .. hint:: - Please always install the latest version of k2. + Please always install the latest version of `k2`_. (3) Install lhotse ------------------ Please refer to ``_ -to install ``lhotse``. - +to install `lhotse`_. .. hint:: @@ -100,17 +73,16 @@ to install ``lhotse``. pip install git+https://github.com/lhotse-speech/lhotse - to install the latest version of lhotse. + to install the latest version of `lhotse`_. (4) Download icefall -------------------- -``icefall`` is a collection of Python scripts; what you need is to download it +`icefall`_ is a collection of Python scripts; what you need is to download it and set the environment variable ``PYTHONPATH`` to point to it. -Assume you want to place ``icefall`` in the folder ``/tmp``. The -following commands show you how to setup ``icefall``: - +Assume you want to place `icefall`_ in the folder ``/tmp``. The +following commands show you how to setup `icefall`_: .. code-block:: bash @@ -122,285 +94,334 @@ following commands show you how to setup ``icefall``: .. HINT:: - You can put several versions of ``icefall`` in the same virtual environment. - To switch among different versions of ``icefall``, just set ``PYTHONPATH`` + You can put several versions of `icefall`_ in the same virtual environment. + To switch among different versions of `icefall`_, just set ``PYTHONPATH`` to point to the version you want. - Installation example -------------------- The following shows an example about setting up the environment. - (1) Create a virtual environment ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~ .. code-block:: bash - $ virtualenv -p python3.8 test-icefall + kuangfangjun:~$ virtualenv -p python3.8 test-icefall + created virtual environment CPython3.8.0.final.0-64 in 9422ms + creator CPython3Posix(dest=/star-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=/star-fj/fangjun/.local/share/virtualenv) + added seed packages: pip==22.3.1, setuptools==65.6.3, wheel==0.38.4 + activators BashActivator,CShellActivator,FishActivator,NushellActivator,PowerShellActivator,PythonActivator - 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 + kuangfangjun:~$ source test-icefall/bin/activate + (test-icefall) kuangfangjun:~$ -(2) Activate your virtual environment -~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~ +(2) Install CUDA toolkit and cuDNN +~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~ + +You need to determine the version of CUDA toolkit to install. .. code-block:: bash - $ source test-icefall/bin/activate + (test-icefall) kuangfangjun:~$ nvidia-smi | head -n 4 -(3) Install k2 + Wed Jul 26 21:57:49 2023 + +-----------------------------------------------------------------------------+ + | NVIDIA-SMI 510.47.03 Driver Version: 510.47.03 CUDA Version: 11.6 | + |-------------------------------+----------------------+----------------------+ + +You can choose any CUDA version that is ``not`` greater than the version printed by ``nvidia-smi``. +In our case, we can choose any version ``<= 11.6``. + +We will use ``CUDA 11.6`` in this example. Please follow +``_ +to install CUDA toolkit and cuDNN if you have not done that before. + +After installing CUDA toolkit, you can use the following command to verify it: + +.. code-block:: bash + + (test-icefall) kuangfangjun:~$ nvcc --version + + nvcc: NVIDIA (R) Cuda compiler driver + Copyright (c) 2005-2019 NVIDIA Corporation + Built on Wed_Oct_23_19:24:38_PDT_2019 + Cuda compilation tools, release 10.2, V10.2.89 + +(3) Install torch and torchaudio +~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~ + +Since we have selected CUDA toolkit ``11.6``, we have to install a version of `torch`_ +that is compiled against CUDA ``11.6``. We select ``torch 1.13.0+cu116`` in this +example. + +After selecting the version of `torch`_ to install, we need to also install +a compatible version of `torchaudio`_, which is ``0.13.0+cu116`` in our case. + +Please refer to ``_ +to select an appropriate version of `torchaudio`_ to install if you use a different +version of `torch`_. + +.. code-block:: bash + + (test-icefall) kuangfangjun:~$ pip install torch==1.13.0+cu116 torchaudio==0.13.0+cu116 -f https://download.pytorch.org/whl/torch_stable.html + + Looking in links: https://download.pytorch.org/whl/torch_stable.html + Collecting torch==1.13.0+cu116 + Downloading https://download.pytorch.org/whl/cu116/torch-1.13.0%2Bcu116-cp38-cp38-linux_x86_64.whl (1983.0 MB) + ________________________________________ 2.0/2.0 GB 764.4 kB/s eta 0:00:00 + Collecting torchaudio==0.13.0+cu116 + Downloading https://download.pytorch.org/whl/cu116/torchaudio-0.13.0%2Bcu116-cp38-cp38-linux_x86_64.whl (4.2 MB) + ________________________________________ 4.2/4.2 MB 1.3 MB/s eta 0:00:00 + Requirement already satisfied: typing-extensions in /star-fj/fangjun/test-icefall/lib/python3.8/site-packages (from torch==1.13.0+cu116) (4.7.1) + Installing collected packages: torch, torchaudio + Successfully installed torch-1.13.0+cu116 torchaudio-0.13.0+cu116 + +Verify that `torch`_ and `torchaudio`_ are successfully installed: + +.. code-block:: bash + + (test-icefall) kuangfangjun:~$ python3 -c "import torch; print(torch.__version__)" + + 1.13.0+cu116 + + (test-icefall) kuangfangjun:~$ python3 -c "import torchaudio; print(torchaudio.__version__)" + + 0.13.0+cu116 + +(4) Install k2 ~~~~~~~~~~~~~~ +We will install `k2`_ from pre-compiled wheels by following +``_ + .. code-block:: bash - $ pip install k2==1.4.dev20210822+cpu.torch1.9.0 -f https://k2-fsa.org/nightly/index.html + (test-icefall) kuangfangjun:~$ pip install k2==1.24.3.dev20230725+cuda11.6.torch1.13.0 -f https://k2-fsa.github.io/k2/cuda.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 + Looking in indexes: https://pypi.tuna.tsinghua.edu.cn/simple + Looking in links: https://k2-fsa.github.io/k2/cuda.html + Collecting k2==1.24.3.dev20230725+cuda11.6.torch1.13.0 + Downloading https://huggingface.co/csukuangfj/k2/resolve/main/ubuntu-cuda/k2-1.24.3.dev20230725%2Bcuda11.6.torch1.13.0-cp38-cp38-manylinux_2_17_x86_64.manylinux2014_x86_64.whl (104.3 MB) + ________________________________________ 104.3/104.3 MB 5.1 MB/s eta 0:00:00 + Requirement already satisfied: torch==1.13.0 in /star-fj/fangjun/test-icefall/lib/python3.8/site-packages (from k2==1.24.3.dev20230725+cuda11.6.torch1.13.0) (1.13.0+cu116) 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 + Using cached https://pypi.tuna.tsinghua.edu.cn/packages/de/5e/fcbb22c68208d39edff467809d06c9d81d7d27426460ebc598e55130c1aa/graphviz-0.20.1-py3-none-any.whl (47 kB) + Requirement already satisfied: typing-extensions in /star-fj/fangjun/test-icefall/lib/python3.8/site-packages (from torch==1.13.0->k2==1.24.3.dev20230725+cuda11.6.torch1.13.0) (4.7.1) + Installing collected packages: graphviz, k2 + Successfully installed graphviz-0.20.1 k2-1.24.3.dev20230725+cuda11.6.torch1.13.0 -.. WARNING:: +.. hint:: - We choose to install a CPU version of k2 for testing. You would probably want to install - a CUDA version of k2. + Please refer to ``_ for the available + pre-compiled wheels about `k2`_. +Verify that `k2`_ has been installed successfully: -(4) Install lhotse +.. code-block:: bash + + (test-icefall) kuangfangjun:~$ python3 -m k2.version + + Collecting environment information... + + k2 version: 1.24.3 + Build type: Release + Git SHA1: 4c05309499a08454997adf500b56dcc629e35ae5 + Git date: Tue Jul 25 16:23:36 2023 + Cuda used to build k2: 11.6 + cuDNN used to build k2: 8.3.2 + Python version used to build k2: 3.8 + OS used to build k2: CentOS Linux release 7.9.2009 (Core) + CMake version: 3.27.0 + GCC version: 9.3.1 + CMAKE_CUDA_FLAGS: -Wno-deprecated-gpu-targets -lineinfo --expt-extended-lambda -use_fast_math -Xptxas=-w --expt-extended-lambda -gencode arch=compute_35,code=sm_35 -lineinfo --expt-extended-lambda -use_fast_math -Xptxas=-w --expt-extended-lambda -gencode arch=compute_50,code=sm_50 -lineinfo --expt-extended-lambda -use_fast_math -Xptxas=-w --expt-extended-lambda -gencode arch=compute_60,code=sm_60 -lineinfo --expt-extended-lambda -use_fast_math -Xptxas=-w --expt-extended-lambda -gencode arch=compute_61,code=sm_61 -lineinfo --expt-extended-lambda -use_fast_math -Xptxas=-w --expt-extended-lambda -gencode arch=compute_70,code=sm_70 -lineinfo --expt-extended-lambda -use_fast_math -Xptxas=-w --expt-extended-lambda -gencode arch=compute_75,code=sm_75 -lineinfo --expt-extended-lambda -use_fast_math -Xptxas=-w --expt-extended-lambda -gencode arch=compute_80,code=sm_80 -lineinfo --expt-extended-lambda -use_fast_math -Xptxas=-w --expt-extended-lambda -gencode arch=compute_86,code=sm_86 -DONNX_NAMESPACE=onnx_c2 -gencode arch=compute_35,code=sm_35 -gencode arch=compute_50,code=sm_50 -gencode arch=compute_52,code=sm_52 -gencode arch=compute_60,code=sm_60 -gencode arch=compute_61,code=sm_61 -gencode arch=compute_70,code=sm_70 -gencode arch=compute_75,code=sm_75 -gencode arch=compute_80,code=sm_80 -gencode arch=compute_86,code=sm_86 -gencode arch=compute_86,code=compute_86 -Xcudafe --diag_suppress=cc_clobber_ignored,--diag_suppress=integer_sign_change,--diag_suppress=useless_using_declaration,--diag_suppress=set_but_not_used,--diag_suppress=field_without_dll_interface,--diag_suppress=base_class_has_different_dll_interface,--diag_suppress=dll_interface_conflict_none_assumed,--diag_suppress=dll_interface_conflict_dllexport_assumed,--diag_suppress=implicit_return_from_non_void_function,--diag_suppress=unsigned_compare_with_zero,--diag_suppress=declared_but_not_referenced,--diag_suppress=bad_friend_decl --expt-relaxed-constexpr --expt-extended-lambda -D_GLIBCXX_USE_CXX11_ABI=0 --compiler-options -Wall --compiler-options -Wno-strict-overflow --compiler-options -Wno-unknown-pragmas + CMAKE_CXX_FLAGS: -D_GLIBCXX_USE_CXX11_ABI=0 -Wno-unused-variable -Wno-strict-overflow + PyTorch version used to build k2: 1.13.0+cu116 + PyTorch is using Cuda: 11.6 + NVTX enabled: True + With CUDA: True + Disable debug: True + Sync kernels : False + Disable checks: False + Max cpu memory allocate: 214748364800 bytes (or 200.0 GB) + k2 abort: False + __file__: /star-fj/fangjun/test-icefall/lib/python3.8/site-packages/k2/version/version.py + _k2.__file__: /star-fj/fangjun/test-icefall/lib/python3.8/site-packages/_k2.cpython-38-x86_64-linux-gnu.so + +(5) Install lhotse ~~~~~~~~~~~~~~~~~~ -.. code-block:: +.. code-block:: bash - $ pip install git+https://github.com/lhotse-speech/lhotse + (test-icefall) kuangfangjun:~$ pip install git+https://github.com/lhotse-speech/lhotse Collecting git+https://github.com/lhotse-speech/lhotse - Cloning https://github.com/lhotse-speech/lhotse to /tmp/pip-req-build-7b1b76ge - Running command git clone -q https://github.com/lhotse-speech/lhotse /tmp/pip-req-build-7b1b76ge - Collecting audioread>=2.1.9 - Using cached audioread-2.1.9-py3-none-any.whl - Collecting SoundFile>=0.10 - Using cached SoundFile-0.10.3.post1-py2.py3-none-any.whl (21 kB) - Collecting click>=7.1.1 - Using cached click-8.0.1-py3-none-any.whl (97 kB) + Cloning https://github.com/lhotse-speech/lhotse to /tmp/pip-req-build-vq12fd5i + Running command git clone --filter=blob:none --quiet https://github.com/lhotse-speech/lhotse /tmp/pip-req-build-vq12fd5i + Resolved https://github.com/lhotse-speech/lhotse to commit 7640d663469b22cd0b36f3246ee9b849cd25e3b7 + Installing build dependencies ... done + Getting requirements to build wheel ... done + Preparing metadata (pyproject.toml) ... done Collecting cytoolz>=0.10.1 - Using cached cytoolz-0.11.0-cp38-cp38-manylinux_2_17_x86_64.manylinux2014_x86_64.whl (1.9 MB) - Collecting dataclasses - Using cached dataclasses-0.6-py3-none-any.whl (14 kB) - Collecting h5py>=2.10.0 - Downloading h5py-3.4.0-cp38-cp38-manylinux_2_12_x86_64.manylinux2010_x86_64.whl (4.5 MB) - |________________________________| 4.5 MB 684 kB/s - Collecting intervaltree>=3.1.0 - Using cached intervaltree-3.1.0-py2.py3-none-any.whl - Collecting lilcom>=1.1.0 - Using cached lilcom-1.1.1-cp38-cp38-linux_x86_64.whl - Collecting numpy>=1.18.1 - Using cached numpy-1.21.2-cp38-cp38-manylinux_2_12_x86_64.manylinux2010_x86_64.whl (15.8 MB) - Collecting packaging - Using cached packaging-21.0-py3-none-any.whl (40 kB) + Downloading https://pypi.tuna.tsinghua.edu.cn/packages/1e/3b/a7828d575aa17fb7acaf1ced49a3655aa36dad7e16eb7e6a2e4df0dda76f/cytoolz-0.12.2-cp38-cp38- + manylinux_2_17_x86_64.manylinux2014_x86_64.whl (2.0 MB) + ________________________________________ 2.0/2.0 MB 33.2 MB/s eta 0:00:00 Collecting pyyaml>=5.3.1 - Using cached PyYAML-5.4.1-cp38-cp38-manylinux1_x86_64.whl (662 kB) + Downloading https://pypi.tuna.tsinghua.edu.cn/packages/c8/6b/6600ac24725c7388255b2f5add93f91e58a5d7efaf4af244fdbcc11a541b/PyYAML-6.0.1-cp38-cp38-ma + nylinux_2_17_x86_64.manylinux2014_x86_64.whl (736 kB) + ________________________________________ 736.6/736.6 kB 38.6 MB/s eta 0:00:00 + Collecting dataclasses + Downloading https://pypi.tuna.tsinghua.edu.cn/packages/26/2f/1095cdc2868052dd1e64520f7c0d5c8c550ad297e944e641dbf1ffbb9a5d/dataclasses-0.6-py3-none- + any.whl (14 kB) + Requirement already satisfied: torchaudio in ./test-icefall/lib/python3.8/site-packages (from lhotse==1.16.0.dev0+git.7640d66.clean) (0.13.0+cu116) + Collecting lilcom>=1.1.0 + Downloading https://pypi.tuna.tsinghua.edu.cn/packages/a8/65/df0a69c52bd085ca1ad4e5c4c1a5c680e25f9477d8e49316c4ff1e5084a4/lilcom-1.7-cp38-cp38-many + linux_2_17_x86_64.manylinux2014_x86_64.whl (87 kB) + ________________________________________ 87.1/87.1 kB 8.7 MB/s eta 0:00:00 Collecting tqdm - Downloading tqdm-4.62.1-py2.py3-none-any.whl (76 kB) - |________________________________| 76 kB 2.7 MB/s - Collecting torchaudio==0.9.0 - Downloading torchaudio-0.9.0-cp38-cp38-manylinux1_x86_64.whl (1.9 MB) - |________________________________| 1.9 MB 73.1 MB/s - Requirement already satisfied: torch==1.9.0 in ./test-icefall/lib/python3.8/site-packages (from torchaudio==0.9.0->lhotse===0.8.0.dev - -2a1410b-clean) (1.9.0) - Requirement already satisfied: typing-extensions in ./test-icefall/lib/python3.8/site-packages (from torch==1.9.0->torchaudio==0.9.0- - >lhotse===0.8.0.dev-2a1410b-clean) (3.10.0.0) + Using cached https://pypi.tuna.tsinghua.edu.cn/packages/e6/02/a2cff6306177ae6bc73bc0665065de51dfb3b9db7373e122e2735faf0d97/tqdm-4.65.0-py3-none-any + .whl (77 kB) + Requirement already satisfied: numpy>=1.18.1 in ./test-icefall/lib/python3.8/site-packages (from lhotse==1.16.0.dev0+git.7640d66.clean) (1.24.4) + Collecting audioread>=2.1.9 + Using cached https://pypi.tuna.tsinghua.edu.cn/packages/5d/cb/82a002441902dccbe427406785db07af10182245ee639ea9f4d92907c923/audioread-3.0.0.tar.gz ( + 377 kB) + Preparing metadata (setup.py) ... done + Collecting tabulate>=0.8.1 + Using cached https://pypi.tuna.tsinghua.edu.cn/packages/40/44/4a5f08c96eb108af5cb50b41f76142f0afa346dfa99d5296fe7202a11854/tabulate-0.9.0-py3-none- + any.whl (35 kB) + Collecting click>=7.1.1 + Downloading https://pypi.tuna.tsinghua.edu.cn/packages/1a/70/e63223f8116931d365993d4a6b7ef653a4d920b41d03de7c59499962821f/click-8.1.6-py3-none-any. + whl (97 kB) + ________________________________________ 97.9/97.9 kB 8.4 MB/s eta 0:00:00 + Collecting packaging + Using cached https://pypi.tuna.tsinghua.edu.cn/packages/ab/c3/57f0601a2d4fe15de7a553c00adbc901425661bf048f2a22dfc500caf121/packaging-23.1-py3-none- + any.whl (48 kB) + Collecting intervaltree>=3.1.0 + Downloading https://pypi.tuna.tsinghua.edu.cn/packages/50/fb/396d568039d21344639db96d940d40eb62befe704ef849b27949ded5c3bb/intervaltree-3.1.0.tar.gz + (32 kB) + Preparing metadata (setup.py) ... done + Requirement already satisfied: torch in ./test-icefall/lib/python3.8/site-packages (from lhotse==1.16.0.dev0+git.7640d66.clean) (1.13.0+cu116) + Collecting SoundFile>=0.10 + Downloading https://pypi.tuna.tsinghua.edu.cn/packages/ad/bd/0602167a213d9184fc688b1086dc6d374b7ae8c33eccf169f9b50ce6568c/soundfile-0.12.1-py2.py3- + none-manylinux_2_17_x86_64.whl (1.3 MB) + ________________________________________ 1.3/1.3 MB 46.5 MB/s eta 0:00:00 Collecting toolz>=0.8.0 - Using cached toolz-0.11.1-py3-none-any.whl (55 kB) + Using cached https://pypi.tuna.tsinghua.edu.cn/packages/7f/5c/922a3508f5bda2892be3df86c74f9cf1e01217c2b1f8a0ac4841d903e3e9/toolz-0.12.0-py3-none-any.whl (55 kB) Collecting sortedcontainers<3.0,>=2.0 - Using cached sortedcontainers-2.4.0-py2.py3-none-any.whl (29 kB) + Using cached https://pypi.tuna.tsinghua.edu.cn/packages/32/46/9cb0e58b2deb7f82b84065f37f3bffeb12413f947f9388e4cac22c4621ce/sortedcontainers-2.4.0-py2.py3-none-any.whl (29 kB) Collecting cffi>=1.0 - Using cached cffi-1.14.6-cp38-cp38-manylinux1_x86_64.whl (411 kB) + Using cached https://pypi.tuna.tsinghua.edu.cn/packages/b7/8b/06f30caa03b5b3ac006de4f93478dbd0239e2a16566d81a106c322dc4f79/cffi-1.15.1-cp38-cp38-manylinux_2_17_x86_64.manylinux2014_x86_64.whl (442 kB) + Requirement already satisfied: typing-extensions in ./test-icefall/lib/python3.8/site-packages (from torch->lhotse==1.16.0.dev0+git.7640d66.clean) (4.7.1) Collecting pycparser - Using cached pycparser-2.20-py2.py3-none-any.whl (112 kB) - Collecting pyparsing>=2.0.2 - Using cached pyparsing-2.4.7-py2.py3-none-any.whl (67 kB) - Building wheels for collected packages: lhotse - Building wheel for lhotse (setup.py) ... done - Created wheel for lhotse: filename=lhotse-0.8.0.dev_2a1410b_clean-py3-none-any.whl size=342242 sha256=f683444afa4dc0881133206b4646a - 9d0f774224cc84000f55d0a67f6e4a37997 - Stored in directory: /tmp/pip-ephem-wheel-cache-ftu0qysz/wheels/7f/7a/8e/a0bf241336e2e3cb573e1e21e5600952d49f5162454f2e612f - WARNING: Built wheel for lhotse is invalid: Metadata 1.2 mandates PEP 440 version, but '0.8.0.dev-2a1410b-clean' is not - Failed to build lhotse - Installing collected packages: pycparser, toolz, sortedcontainers, pyparsing, numpy, cffi, tqdm, torchaudio, SoundFile, pyyaml, packa - ging, lilcom, intervaltree, h5py, dataclasses, cytoolz, click, audioread, lhotse - Running setup.py install for lhotse ... done - DEPRECATION: lhotse was installed using the legacy 'setup.py install' method, because a wheel could not be built for it. A possible - replacement is to fix the wheel build issue reported above. You can find discussion regarding this at https://github.com/pypa/pip/is - sues/8368. - Successfully installed SoundFile-0.10.3.post1 audioread-2.1.9 cffi-1.14.6 click-8.0.1 cytoolz-0.11.0 dataclasses-0.6 h5py-3.4.0 inter - valtree-3.1.0 lhotse-0.8.0.dev-2a1410b-clean lilcom-1.1.1 numpy-1.21.2 packaging-21.0 pycparser-2.20 pyparsing-2.4.7 pyyaml-5.4.1 sor - tedcontainers-2.4.0 toolz-0.11.1 torchaudio-0.9.0 tqdm-4.62.1 + Using cached https://pypi.tuna.tsinghua.edu.cn/packages/62/d5/5f610ebe421e85889f2e55e33b7f9a6795bd982198517d912eb1c76e1a53/pycparser-2.21-py2.py3-none-any.whl (118 kB) + Building wheels for collected packages: lhotse, audioread, intervaltree + Building wheel for lhotse (pyproject.toml) ... done + Created wheel for lhotse: filename=lhotse-1.16.0.dev0+git.7640d66.clean-py3-none-any.whl size=687627 sha256=cbf0a4d2d0b639b33b91637a4175bc251d6a021a069644ecb1a9f2b3a83d072a + Stored in directory: /tmp/pip-ephem-wheel-cache-wwtk90_m/wheels/7f/7a/8e/a0bf241336e2e3cb573e1e21e5600952d49f5162454f2e612f + Building wheel for audioread (setup.py) ... done + Created wheel for audioread: filename=audioread-3.0.0-py3-none-any.whl size=23704 sha256=5e2d3537c96ce9cf0f645a654c671163707bf8cb8d9e358d0e2b0939a85ff4c2 + Stored in directory: /star-fj/fangjun/.cache/pip/wheels/e2/c3/9c/f19ae5a03f8862d9f0776b0c0570f1fdd60a119d90954e3f39 + Building wheel for intervaltree (setup.py) ... done + Created wheel for intervaltree: filename=intervaltree-3.1.0-py2.py3-none-any.whl size=26098 sha256=2604170976cfffe0d2f678cb1a6e5b525f561cd50babe53d631a186734fec9f9 + Stored in directory: /star-fj/fangjun/.cache/pip/wheels/f3/ed/2b/c179ebfad4e15452d6baef59737f27beb9bfb442e0620f7271 + Successfully built lhotse audioread intervaltree + Installing collected packages: sortedcontainers, dataclasses, tqdm, toolz, tabulate, pyyaml, pycparser, packaging, lilcom, intervaltree, click, audioread, cytoolz, cffi, SoundFile, lhotse + Successfully installed SoundFile-0.12.1 audioread-3.0.0 cffi-1.15.1 click-8.1.6 cytoolz-0.12.2 dataclasses-0.6 intervaltree-3.1.0 lhotse-1.16.0.dev0+git.7640d66.clean lilcom-1.7 packaging-23.1 pycparser-2.21 pyyaml-6.0.1 sortedcontainers-2.4.0 tabulate-0.9.0 toolz-0.12.0 tqdm-4.65.0 -(5) Download icefall + +Verify that `lhotse`_ has been installed successfully: + +.. code-block:: bash + + (test-icefall) kuangfangjun:~$ python3 -c "import lhotse; print(lhotse.__version__)" + + 1.16.0.dev+git.7640d66.clean + +(6) Download icefall ~~~~~~~~~~~~~~~~~~~~ -.. code-block:: +.. code-block:: bash - $ cd /tmp - $ git clone https://github.com/k2-fsa/icefall + (test-icefall) kuangfangjun:~$ cd /tmp/ + + (test-icefall) kuangfangjun:tmp$ git clone https://github.com/k2-fsa/icefall Cloning into 'icefall'... - remote: Enumerating objects: 500, done. - remote: Counting objects: 100% (500/500), done. - remote: Compressing objects: 100% (308/308), done. - remote: Total 500 (delta 263), reused 307 (delta 102), pack-reused 0 - Receiving objects: 100% (500/500), 172.49 KiB | 385.00 KiB/s, done. - Resolving deltas: 100% (263/263), done. + remote: Enumerating objects: 12942, done. + remote: Counting objects: 100% (67/67), done. + remote: Compressing objects: 100% (56/56), done. + remote: Total 12942 (delta 17), reused 35 (delta 6), pack-reused 12875 + Receiving objects: 100% (12942/12942), 14.77 MiB | 9.29 MiB/s, done. + Resolving deltas: 100% (8835/8835), done. - $ cd icefall - $ pip install -r requirements.txt - - Collecting kaldilm - Downloading kaldilm-1.8.tar.gz (48 kB) - |________________________________| 48 kB 574 kB/s - Collecting kaldialign - Using cached kaldialign-0.2-cp38-cp38-linux_x86_64.whl - Collecting sentencepiece>=0.1.96 - Using cached sentencepiece-0.1.96-cp38-cp38-manylinux_2_17_x86_64.manylinux2014_x86_64.whl (1.2 MB) - Collecting tensorboard - Using cached tensorboard-2.6.0-py3-none-any.whl (5.6 MB) - Requirement already satisfied: setuptools>=41.0.0 in /ceph-fj/fangjun/test-icefall/lib/python3.8/site-packages (from tensorboard->-r - requirements.txt (line 4)) (57.4.0) - Collecting absl-py>=0.4 - Using cached absl_py-0.13.0-py3-none-any.whl (132 kB) - Collecting google-auth-oauthlib<0.5,>=0.4.1 - Using cached google_auth_oauthlib-0.4.5-py2.py3-none-any.whl (18 kB) - Collecting grpcio>=1.24.3 - Using cached grpcio-1.39.0-cp38-cp38-manylinux2014_x86_64.whl (4.3 MB) - Requirement already satisfied: wheel>=0.26 in /ceph-fj/fangjun/test-icefall/lib/python3.8/site-packages (from tensorboard->-r require - ments.txt (line 4)) (0.36.2) - Requirement already satisfied: numpy>=1.12.0 in /ceph-fj/fangjun/test-icefall/lib/python3.8/site-packages (from tensorboard->-r requi - rements.txt (line 4)) (1.21.2) - Collecting protobuf>=3.6.0 - Using cached protobuf-3.17.3-cp38-cp38-manylinux_2_5_x86_64.manylinux1_x86_64.whl (1.0 MB) - Collecting werkzeug>=0.11.15 - Using cached Werkzeug-2.0.1-py3-none-any.whl (288 kB) - Collecting tensorboard-data-server<0.7.0,>=0.6.0 - Using cached tensorboard_data_server-0.6.1-py3-none-manylinux2010_x86_64.whl (4.9 MB) - Collecting google-auth<2,>=1.6.3 - Downloading google_auth-1.35.0-py2.py3-none-any.whl (152 kB) - |________________________________| 152 kB 1.4 MB/s - Collecting requests<3,>=2.21.0 - Using cached requests-2.26.0-py2.py3-none-any.whl (62 kB) - Collecting tensorboard-plugin-wit>=1.6.0 - Using cached tensorboard_plugin_wit-1.8.0-py3-none-any.whl (781 kB) - Collecting markdown>=2.6.8 - Using cached Markdown-3.3.4-py3-none-any.whl (97 kB) - Collecting six - Using cached six-1.16.0-py2.py3-none-any.whl (11 kB) - Collecting cachetools<5.0,>=2.0.0 - Using cached cachetools-4.2.2-py3-none-any.whl (11 kB) - Collecting rsa<5,>=3.1.4 - Using cached rsa-4.7.2-py3-none-any.whl (34 kB) - Collecting pyasn1-modules>=0.2.1 - Using cached pyasn1_modules-0.2.8-py2.py3-none-any.whl (155 kB) - Collecting requests-oauthlib>=0.7.0 - Using cached requests_oauthlib-1.3.0-py2.py3-none-any.whl (23 kB) - Collecting pyasn1<0.5.0,>=0.4.6 - Using cached pyasn1-0.4.8-py2.py3-none-any.whl (77 kB) - Collecting urllib3<1.27,>=1.21.1 - Using cached urllib3-1.26.6-py2.py3-none-any.whl (138 kB) - Collecting certifi>=2017.4.17 - Using cached certifi-2021.5.30-py2.py3-none-any.whl (145 kB) - Collecting charset-normalizer~=2.0.0 - Using cached charset_normalizer-2.0.4-py3-none-any.whl (36 kB) - Collecting idna<4,>=2.5 - Using cached idna-3.2-py3-none-any.whl (59 kB) - Collecting oauthlib>=3.0.0 - Using cached oauthlib-3.1.1-py2.py3-none-any.whl (146 kB) - Building wheels for collected packages: kaldilm - Building wheel for kaldilm (setup.py) ... done - Created wheel for kaldilm: filename=kaldilm-1.8-cp38-cp38-linux_x86_64.whl size=897233 sha256=eccb906cafcd45bf9a7e1a1718e4534254bfb - f4c0d0cbc66eee6c88d68a63862 - Stored in directory: /root/fangjun/.cache/pip/wheels/85/7d/63/f2dd586369b8797cb36d213bf3a84a789eeb92db93d2e723c9 - Successfully built kaldilm - Installing collected packages: urllib3, pyasn1, idna, charset-normalizer, certifi, six, rsa, requests, pyasn1-modules, oauthlib, cach - etools, requests-oauthlib, google-auth, werkzeug, tensorboard-plugin-wit, tensorboard-data-server, protobuf, markdown, grpcio, google - -auth-oauthlib, absl-py, tensorboard, sentencepiece, kaldilm, kaldialign - Successfully installed absl-py-0.13.0 cachetools-4.2.2 certifi-2021.5.30 charset-normalizer-2.0.4 google-auth-1.35.0 google-auth-oaut - hlib-0.4.5 grpcio-1.39.0 idna-3.2 kaldialign-0.2 kaldilm-1.8 markdown-3.3.4 oauthlib-3.1.1 protobuf-3.17.3 pyasn1-0.4.8 pyasn1-module - s-0.2.8 requests-2.26.0 requests-oauthlib-1.3.0 rsa-4.7.2 sentencepiece-0.1.96 six-1.16.0 tensorboard-2.6.0 tensorboard-data-server-0 - .6.1 tensorboard-plugin-wit-1.8.0 urllib3-1.26.6 werkzeug-2.0.1 + (test-icefall) kuangfangjun:tmp$ cd icefall/ + (test-icefall) kuangfangjun:icefall$ pip install -r ./requirements.txt Test Your Installation ---------------------- To test that your installation is successful, let us run the `yesno recipe `_ -on CPU. +on ``CPU``. Data preparation ~~~~~~~~~~~~~~~~ .. code-block:: bash - $ export PYTHONPATH=/tmp/icefall:$PYTHONPATH - $ cd /tmp/icefall - $ cd egs/yesno/ASR - $ ./prepare.sh + (test-icefall) kuangfangjun:icefall$ export PYTHONPATH=/tmp/icefall:$PYTHONPATH + + (test-icefall) kuangfangjun:icefall$ cd /tmp/icefall + + (test-icefall) kuangfangjun:icefall$ cd egs/yesno/ASR + + (test-icefall) kuangfangjun:ASR$ ./prepare.sh + The log of running ``./prepare.sh`` is: .. code-block:: - 2023-05-12 17:55:21 (prepare.sh:27:main) dl_dir: /tmp/icefall/egs/yesno/ASR/download - 2023-05-12 17:55:21 (prepare.sh:30:main) Stage 0: Download data - /tmp/icefall/egs/yesno/ASR/download/waves_yesno.tar.gz: 100%|_______________________________________________________________| 4.70M/4.70M [06:54<00:00, 11.4kB/s] - 2023-05-12 18:02:19 (prepare.sh:39:main) Stage 1: Prepare yesno manifest - 2023-05-12 18:02:21 (prepare.sh:45:main) Stage 2: Compute fbank for yesno - 2023-05-12 18:02:23,199 INFO [compute_fbank_yesno.py:65] Processing train - Extracting and storing features: 100%|_______________________________________________________________| 90/90 [00:00<00:00, 212.60it/s] - 2023-05-12 18:02:23,640 INFO [compute_fbank_yesno.py:65] Processing test - Extracting and storing features: 100%|_______________________________________________________________| 30/30 [00:00<00:00, 304.53it/s] - 2023-05-12 18:02:24 (prepare.sh:51:main) Stage 3: Prepare lang - 2023-05-12 18:02:26 (prepare.sh:66:main) Stage 4: Prepare G - /project/kaldilm/csrc/arpa_file_parser.cc:void kaldilm::ArpaFileParser::Read(std::istream&):79 - [I] Reading \data\ section. - /project/kaldilm/csrc/arpa_file_parser.cc:void kaldilm::ArpaFileParser::Read(std::istream&):140 - [I] Reading \1-grams: section. - 2023-05-12 18:02:26 (prepare.sh:92:main) Stage 5: Compile HLG - 2023-05-12 18:02:28,581 INFO [compile_hlg.py:124] Processing data/lang_phone - 2023-05-12 18:02:28,582 INFO [lexicon.py:171] Converting L.pt to Linv.pt - 2023-05-12 18:02:28,609 INFO [compile_hlg.py:48] Building ctc_topo. max_token_id: 3 - 2023-05-12 18:02:28,610 INFO [compile_hlg.py:52] Loading G.fst.txt - 2023-05-12 18:02:28,611 INFO [compile_hlg.py:62] Intersecting L and G - 2023-05-12 18:02:28,613 INFO [compile_hlg.py:64] LG shape: (4, None) - 2023-05-12 18:02:28,613 INFO [compile_hlg.py:66] Connecting LG - 2023-05-12 18:02:28,614 INFO [compile_hlg.py:68] LG shape after k2.connect: (4, None) - 2023-05-12 18:02:28,614 INFO [compile_hlg.py:70] - 2023-05-12 18:02:28,614 INFO [compile_hlg.py:71] Determinizing LG - 2023-05-12 18:02:28,615 INFO [compile_hlg.py:74] - 2023-05-12 18:02:28,615 INFO [compile_hlg.py:76] Connecting LG after k2.determinize - 2023-05-12 18:02:28,615 INFO [compile_hlg.py:79] Removing disambiguation symbols on LG - 2023-05-12 18:02:28,616 INFO [compile_hlg.py:91] LG shape after k2.remove_epsilon: (6, None) - 2023-05-12 18:02:28,617 INFO [compile_hlg.py:96] Arc sorting LG - 2023-05-12 18:02:28,617 INFO [compile_hlg.py:99] Composing H and LG - 2023-05-12 18:02:28,619 INFO [compile_hlg.py:106] Connecting LG - 2023-05-12 18:02:28,619 INFO [compile_hlg.py:109] Arc sorting LG - 2023-05-12 18:02:28,619 INFO [compile_hlg.py:111] HLG.shape: (8, None) - 2023-05-12 18:02:28,619 INFO [compile_hlg.py:127] Saving HLG.pt to data/lang_phone - + 2023-07-27 12:41:39 (prepare.sh:27:main) dl_dir: /tmp/icefall/egs/yesno/ASR/download + 2023-07-27 12:41:39 (prepare.sh:30:main) Stage 0: Download data + /tmp/icefall/egs/yesno/ASR/download/waves_yesno.tar.gz: 100%|___________________________________________________| 4.70M/4.70M [00:00<00:00, 11.1MB/s] + 2023-07-27 12:41:46 (prepare.sh:39:main) Stage 1: Prepare yesno manifest + 2023-07-27 12:41:50 (prepare.sh:45:main) Stage 2: Compute fbank for yesno + 2023-07-27 12:41:55,718 INFO [compute_fbank_yesno.py:65] Processing train + Extracting and storing features: 100%|_______________________________________________________________________________| 90/90 [00:01<00:00, 87.82it/s] + 2023-07-27 12:41:56,778 INFO [compute_fbank_yesno.py:65] Processing test + Extracting and storing features: 100%|______________________________________________________________________________| 30/30 [00:00<00:00, 256.92it/s] + 2023-07-27 12:41:57 (prepare.sh:51:main) Stage 3: Prepare lang + 2023-07-27 12:42:02 (prepare.sh:66:main) Stage 4: Prepare G + /project/kaldilm/csrc/arpa_file_parser.cc:void kaldilm::ArpaFileParser::Read(std::istream&):79 + [I] Reading \data\ section. + /project/kaldilm/csrc/arpa_file_parser.cc:void kaldilm::ArpaFileParser::Read(std::istream&):140 + [I] Reading \1-grams: section. + 2023-07-27 12:42:02 (prepare.sh:92:main) Stage 5: Compile HLG + 2023-07-27 12:42:07,275 INFO [compile_hlg.py:124] Processing data/lang_phone + 2023-07-27 12:42:07,276 INFO [lexicon.py:171] Converting L.pt to Linv.pt + 2023-07-27 12:42:07,309 INFO [compile_hlg.py:48] Building ctc_topo. max_token_id: 3 + 2023-07-27 12:42:07,310 INFO [compile_hlg.py:52] Loading G.fst.txt + 2023-07-27 12:42:07,314 INFO [compile_hlg.py:62] Intersecting L and G + 2023-07-27 12:42:07,323 INFO [compile_hlg.py:64] LG shape: (4, None) + 2023-07-27 12:42:07,323 INFO [compile_hlg.py:66] Connecting LG + 2023-07-27 12:42:07,323 INFO [compile_hlg.py:68] LG shape after k2.connect: (4, None) + 2023-07-27 12:42:07,323 INFO [compile_hlg.py:70] + 2023-07-27 12:42:07,323 INFO [compile_hlg.py:71] Determinizing LG + 2023-07-27 12:42:07,341 INFO [compile_hlg.py:74] + 2023-07-27 12:42:07,341 INFO [compile_hlg.py:76] Connecting LG after k2.determinize + 2023-07-27 12:42:07,341 INFO [compile_hlg.py:79] Removing disambiguation symbols on LG + 2023-07-27 12:42:07,354 INFO [compile_hlg.py:91] LG shape after k2.remove_epsilon: (6, None) + 2023-07-27 12:42:07,445 INFO [compile_hlg.py:96] Arc sorting LG + 2023-07-27 12:42:07,445 INFO [compile_hlg.py:99] Composing H and LG + 2023-07-27 12:42:07,446 INFO [compile_hlg.py:106] Connecting LG + 2023-07-27 12:42:07,446 INFO [compile_hlg.py:109] Arc sorting LG + 2023-07-27 12:42:07,447 INFO [compile_hlg.py:111] HLG.shape: (8, None) + 2023-07-27 12:42:07,447 INFO [compile_hlg.py:127] Saving HLG.pt to data/lang_phone Training ~~~~~~~~ @@ -409,12 +430,13 @@ Now let us run the training part: .. code-block:: - $ export CUDA_VISIBLE_DEVICES="" - $ ./tdnn/train.py + (test-icefall) kuangfangjun:ASR$ export CUDA_VISIBLE_DEVICES="" + + (test-icefall) kuangfangjun:ASR$ ./tdnn/train.py .. CAUTION:: - We use ``export CUDA_VISIBLE_DEVICES=""`` so that ``icefall`` uses CPU + We use ``export CUDA_VISIBLE_DEVICES=""`` so that `icefall`_ uses CPU even if there are GPUs available. .. hint:: @@ -432,53 +454,52 @@ The training log is given below: .. code-block:: - 2023-05-12 18:04:59,759 INFO [train.py:481] Training started - 2023-05-12 18:04:59,759 INFO [train.py:482] {'exp_dir': PosixPath('tdnn/exp'), 'lang_dir': PosixPath('data/lang_phone'), 'lr': 0.01, 'feature_dim': 23, 'weight_decay': 1e-06, 'start_epoch': 0, - 'best_train_loss': inf, 'best_valid_loss': inf, 'best_train_epoch': -1, 'best_valid_epoch': -1, 'batch_idx_train': 0, 'log_interval': 10, 'reset_interval': 20, 'valid_interval': 10, 'beam_size': 10, - 'reduction': 'sum', 'use_double_scores': True, 'world_size': 1, 'master_port': 12354, 'tensorboard': True, 'num_epochs': 15, 'seed': 42, 'feature_dir': PosixPath('data/fbank'), 'max_duration': 30.0, - 'bucketing_sampler': False, 'num_buckets': 10, 'concatenate_cuts': False, 'duration_factor': 1.0, 'gap': 1.0, 'on_the_fly_feats': False, 'shuffle': False, 'return_cuts': True, 'num_workers': 2, - 'env_info': {'k2-version': '1.24.3', 'k2-build-type': 'Release', 'k2-with-cuda': True, 'k2-git-sha1': '3b7f09fa35e72589914f67089c0da9f196a92ca4', 'k2-git-date': 'Mon May 8 22:58:45 2023', - 'lhotse-version': '1.15.0.dev+git.6fcfced.clean', 'torch-version': '2.0.0+cu118', 'torch-cuda-available': False, 'torch-cuda-version': '11.8', 'python-version': '3.1', 'icefall-git-branch': 'master', - 'icefall-git-sha1': '30bde4b-clean', 'icefall-git-date': 'Thu May 11 17:37:47 2023', 'icefall-path': '/tmp/icefall', - 'k2-path': 'tmp/lib/python3.10/site-packages/k2-1.24.3.dev20230512+cuda11.8.torch2.0.0-py3.10-linux-x86_64.egg/k2/__init__.py', - 'lhotse-path': 'tmp/lib/python3.10/site-packages/lhotse/__init__.py', 'hostname': 'host', 'IP address': '0.0.0.0'}} - 2023-05-12 18:04:59,761 INFO [lexicon.py:168] Loading pre-compiled data/lang_phone/Linv.pt - 2023-05-12 18:04:59,764 INFO [train.py:495] device: cpu - 2023-05-12 18:04:59,791 INFO [asr_datamodule.py:146] About to get train cuts - 2023-05-12 18:04:59,791 INFO [asr_datamodule.py:244] About to get train cuts - 2023-05-12 18:04:59,852 INFO [asr_datamodule.py:149] About to create train dataset - 2023-05-12 18:04:59,852 INFO [asr_datamodule.py:199] Using SingleCutSampler. - 2023-05-12 18:04:59,852 INFO [asr_datamodule.py:205] About to create train dataloader - 2023-05-12 18:04:59,853 INFO [asr_datamodule.py:218] About to get test cuts - 2023-05-12 18:04:59,853 INFO [asr_datamodule.py:252] About to get test cuts - 2023-05-12 18:04:59,986 INFO [train.py:422] Epoch 0, batch 0, loss[loss=1.065, over 2436.00 frames. ], tot_loss[loss=1.065, over 2436.00 frames. ], batch size: 4 - 2023-05-12 18:05:00,352 INFO [train.py:422] Epoch 0, batch 10, loss[loss=0.4561, over 2828.00 frames. ], tot_loss[loss=0.7076, over 22192.90 frames. ], batch size: 4 - 2023-05-12 18:05:00,691 INFO [train.py:444] Epoch 0, validation loss=0.9002, over 18067.00 frames. - 2023-05-12 18:05:00,996 INFO [train.py:422] Epoch 0, batch 20, loss[loss=0.2555, over 2695.00 frames. ], tot_loss[loss=0.484, over 34971.47 frames. ], batch size: 5 - 2023-05-12 18:05:01,217 INFO [train.py:444] Epoch 0, validation loss=0.4688, over 18067.00 frames. - 2023-05-12 18:05:01,251 INFO [checkpoint.py:75] Saving checkpoint to tdnn/exp/epoch-0.pt - 2023-05-12 18:05:01,389 INFO [train.py:422] Epoch 1, batch 0, loss[loss=0.2532, over 2436.00 frames. ], tot_loss[loss=0.2532, over 2436.00 frames. ], batch size: 4 - 2023-05-12 18:05:01,637 INFO [train.py:422] Epoch 1, batch 10, loss[loss=0.1139, over 2828.00 frames. ], tot_loss[loss=0.1592, over 22192.90 frames. ], batch size: 4 - 2023-05-12 18:05:01,859 INFO [train.py:444] Epoch 1, validation loss=0.1629, over 18067.00 frames. - 2023-05-12 18:05:02,094 INFO [train.py:422] Epoch 1, batch 20, loss[loss=0.0767, over 2695.00 frames. ], tot_loss[loss=0.118, over 34971.47 frames. ], batch size: 5 - 2023-05-12 18:05:02,350 INFO [train.py:444] Epoch 1, validation loss=0.06778, over 18067.00 frames. - 2023-05-12 18:05:02,395 INFO [checkpoint.py:75] Saving checkpoint to tdnn/exp/epoch-1.pt + 2023-07-27 12:50:51,936 INFO [train.py:481] Training started + 2023-07-27 12:50:51,936 INFO [train.py:482] {'exp_dir': PosixPath('tdnn/exp'), 'lang_dir': PosixPath('data/lang_phone'), 'lr': 0.01, 'feature_dim': 23, 'weight_decay': 1e-06, 'start_epoch': 0, 'best_train_loss': inf, 'best_valid_loss': inf, 'best_train_epoch': -1, 'best_valid_epoch': -1, 'batch_idx_train': 0, 'log_interval': 10, 'reset_interval': 20, 'valid_interval': 10, 'beam_size': 10, 'reduction': 'sum', 'use_double_scores': True, 'world_size': 1, 'master_port': 12354, 'tensorboard': True, 'num_epochs': 15, 'seed': 42, 'feature_dir': PosixPath('data/fbank'), 'max_duration': 30.0, 'bucketing_sampler': False, 'num_buckets': 10, 'concatenate_cuts': False, 'duration_factor': 1.0, 'gap': 1.0, 'on_the_fly_feats': False, 'shuffle': False, 'return_cuts': True, 'num_workers': 2, 'env_info': {'k2-version': '1.24.3', 'k2-build-type': 'Release', 'k2-with-cuda': True, 'k2-git-sha1': '4c05309499a08454997adf500b56dcc629e35ae5', 'k2-git-date': 'Tue Jul 25 16:23:36 2023', 'lhotse-version': '1.16.0.dev+git.7640d66.clean', 'torch-version': '1.13.0+cu116', 'torch-cuda-available': False, 'torch-cuda-version': '11.6', 'python-version': '3.8', 'icefall-git-branch': 'master', 'icefall-git-sha1': '3fb0a43-clean', 'icefall-git-date': 'Thu Jul 27 12:36:05 2023', 'icefall-path': '/tmp/icefall', 'k2-path': '/star-fj/fangjun/test-icefall/lib/python3.8/site-packages/k2/__init__.py', 'lhotse-path': '/star-fj/fangjun/test-icefall/lib/python3.8/site-packages/lhotse/__init__.py', 'hostname': 'de-74279-k2-train-1-1220091118-57c4d55446-sph26', 'IP address': '10.177.77.20'}} + 2023-07-27 12:50:51,941 INFO [lexicon.py:168] Loading pre-compiled data/lang_phone/Linv.pt + 2023-07-27 12:50:51,949 INFO [train.py:495] device: cpu + 2023-07-27 12:50:51,965 INFO [asr_datamodule.py:146] About to get train cuts + 2023-07-27 12:50:51,965 INFO [asr_datamodule.py:244] About to get train cuts + 2023-07-27 12:50:51,967 INFO [asr_datamodule.py:149] About to create train dataset + 2023-07-27 12:50:51,967 INFO [asr_datamodule.py:199] Using SingleCutSampler. + 2023-07-27 12:50:51,967 INFO [asr_datamodule.py:205] About to create train dataloader + 2023-07-27 12:50:51,968 INFO [asr_datamodule.py:218] About to get test cuts + 2023-07-27 12:50:51,968 INFO [asr_datamodule.py:252] About to get test cuts + 2023-07-27 12:50:52,565 INFO [train.py:422] Epoch 0, batch 0, loss[loss=1.065, over 2436.00 frames. ], tot_loss[loss=1.065, over 2436.00 frames. ], batch size: 4 + 2023-07-27 12:50:53,681 INFO [train.py:422] Epoch 0, batch 10, loss[loss=0.4561, over 2828.00 frames. ], tot_loss[loss=0.7076, over 22192.90 frames.], batch size: 4 + 2023-07-27 12:50:54,167 INFO [train.py:444] Epoch 0, validation loss=0.9002, over 18067.00 frames. + 2023-07-27 12:50:55,011 INFO [train.py:422] Epoch 0, batch 20, loss[loss=0.2555, over 2695.00 frames. ], tot_loss[loss=0.484, over 34971.47 frames. ], batch size: 5 + 2023-07-27 12:50:55,331 INFO [train.py:444] Epoch 0, validation loss=0.4688, over 18067.00 frames. + 2023-07-27 12:50:55,368 INFO [checkpoint.py:75] Saving checkpoint to tdnn/exp/epoch-0.pt + 2023-07-27 12:50:55,633 INFO [train.py:422] Epoch 1, batch 0, loss[loss=0.2532, over 2436.00 frames. ], tot_loss[loss=0.2532, over 2436.00 frames. ], + batch size: 4 + 2023-07-27 12:50:56,242 INFO [train.py:422] Epoch 1, batch 10, loss[loss=0.1139, over 2828.00 frames. ], tot_loss[loss=0.1592, over 22192.90 frames.], batch size: 4 + 2023-07-27 12:50:56,522 INFO [train.py:444] Epoch 1, validation loss=0.1627, over 18067.00 frames. + 2023-07-27 12:50:57,209 INFO [train.py:422] Epoch 1, batch 20, loss[loss=0.07055, over 2695.00 frames. ], tot_loss[loss=0.1175, over 34971.47 frames.], batch size: 5 + 2023-07-27 12:50:57,600 INFO [train.py:444] Epoch 1, validation loss=0.07091, over 18067.00 frames. + 2023-07-27 12:50:57,640 INFO [checkpoint.py:75] Saving checkpoint to tdnn/exp/epoch-1.pt + 2023-07-27 12:50:57,847 INFO [train.py:422] Epoch 2, batch 0, loss[loss=0.07731, over 2436.00 frames. ], tot_loss[loss=0.07731, over 2436.00 frames.], batch size: 4 + 2023-07-27 12:50:58,427 INFO [train.py:422] Epoch 2, batch 10, loss[loss=0.04391, over 2828.00 frames. ], tot_loss[loss=0.05341, over 22192.90 frames. ], batch size: 4 + 2023-07-27 12:50:58,884 INFO [train.py:444] Epoch 2, validation loss=0.04384, over 18067.00 frames. + 2023-07-27 12:50:59,387 INFO [train.py:422] Epoch 2, batch 20, loss[loss=0.03458, over 2695.00 frames. ], tot_loss[loss=0.04616, over 34971.47 frames. ], batch size: 5 + 2023-07-27 12:50:59,707 INFO [train.py:444] Epoch 2, validation loss=0.03379, over 18067.00 frames. + 2023-07-27 12:50:59,758 INFO [checkpoint.py:75] Saving checkpoint to tdnn/exp/epoch-2.pt - ... ... + ... ... - 2023-05-12 18:05:14,789 INFO [train.py:422] Epoch 13, batch 0, loss[loss=0.01056, over 2436.00 frames. ], tot_loss[loss=0.01056, over 2436.00 frames. ], batch size: 4 - 2023-05-12 18:05:15,016 INFO [train.py:422] Epoch 13, batch 10, loss[loss=0.009022, over 2828.00 frames. ], tot_loss[loss=0.009985, over 22192.90 frames. ], batch size: 4 - 2023-05-12 18:05:15,271 INFO [train.py:444] Epoch 13, validation loss=0.01088, over 18067.00 frames. - 2023-05-12 18:05:15,497 INFO [train.py:422] Epoch 13, batch 20, loss[loss=0.01174, over 2695.00 frames. ], tot_loss[loss=0.01077, over 34971.47 frames. ], batch size: 5 - 2023-05-12 18:05:15,747 INFO [train.py:444] Epoch 13, validation loss=0.01087, over 18067.00 frames. - 2023-05-12 18:05:15,783 INFO [checkpoint.py:75] Saving checkpoint to tdnn/exp/epoch-13.pt - 2023-05-12 18:05:15,921 INFO [train.py:422] Epoch 14, batch 0, loss[loss=0.01045, over 2436.00 frames. ], tot_loss[loss=0.01045, over 2436.00 frames. ], batch size: 4 - 2023-05-12 18:05:16,146 INFO [train.py:422] Epoch 14, batch 10, loss[loss=0.008957, over 2828.00 frames. ], tot_loss[loss=0.009903, over 22192.90 frames. ], batch size: 4 - 2023-05-12 18:05:16,374 INFO [train.py:444] Epoch 14, validation loss=0.01092, over 18067.00 frames. - 2023-05-12 18:05:16,598 INFO [train.py:422] Epoch 14, batch 20, loss[loss=0.01169, over 2695.00 frames. ], tot_loss[loss=0.01065, over 34971.47 frames. ], batch size: 5 - 2023-05-12 18:05:16,824 INFO [train.py:444] Epoch 14, validation loss=0.01077, over 18067.00 frames. - 2023-05-12 18:05:16,862 INFO [checkpoint.py:75] Saving checkpoint to tdnn/exp/epoch-14.pt - 2023-05-12 18:05:16,865 INFO [train.py:555] Done! + 2023-07-27 12:51:23,433 INFO [train.py:422] Epoch 13, batch 0, loss[loss=0.01054, over 2436.00 frames. ], tot_loss[loss=0.01054, over 2436.00 frames. ], batch size: 4 + 2023-07-27 12:51:23,980 INFO [train.py:422] Epoch 13, batch 10, loss[loss=0.009014, over 2828.00 frames. ], tot_loss[loss=0.009974, over 22192.90 frames. ], batch size: 4 + 2023-07-27 12:51:24,489 INFO [train.py:444] Epoch 13, validation loss=0.01085, over 18067.00 frames. + 2023-07-27 12:51:25,258 INFO [train.py:422] Epoch 13, batch 20, loss[loss=0.01172, over 2695.00 frames. ], tot_loss[loss=0.01055, over 34971.47 frames. ], batch size: 5 + 2023-07-27 12:51:25,621 INFO [train.py:444] Epoch 13, validation loss=0.01074, over 18067.00 frames. + 2023-07-27 12:51:25,699 INFO [checkpoint.py:75] Saving checkpoint to tdnn/exp/epoch-13.pt + 2023-07-27 12:51:25,866 INFO [train.py:422] Epoch 14, batch 0, loss[loss=0.01044, over 2436.00 frames. ], tot_loss[loss=0.01044, over 2436.00 frames. ], batch size: 4 + 2023-07-27 12:51:26,844 INFO [train.py:422] Epoch 14, batch 10, loss[loss=0.008942, over 2828.00 frames. ], tot_loss[loss=0.01, over 22192.90 frames. ], batch size: 4 + 2023-07-27 12:51:27,221 INFO [train.py:444] Epoch 14, validation loss=0.01082, over 18067.00 frames. + 2023-07-27 12:51:27,970 INFO [train.py:422] Epoch 14, batch 20, loss[loss=0.01169, over 2695.00 frames. ], tot_loss[loss=0.01054, over 34971.47 frames. ], batch size: 5 + 2023-07-27 12:51:28,247 INFO [train.py:444] Epoch 14, validation loss=0.01073, over 18067.00 frames. + 2023-07-27 12:51:28,323 INFO [checkpoint.py:75] Saving checkpoint to tdnn/exp/epoch-14.pt + 2023-07-27 12:51:28,326 INFO [train.py:555] Done! Decoding ~~~~~~~~ @@ -487,42 +508,32 @@ Let us use the trained model to decode the test set: .. code-block:: - $ ./tdnn/decode.py + (test-icefall) kuangfangjun:ASR$ ./tdnn/decode.py -The decoding log is: + 2023-07-27 12:55:12,840 INFO [decode.py:263] Decoding started + 2023-07-27 12:55:12,840 INFO [decode.py:264] {'exp_dir': PosixPath('tdnn/exp'), 'lang_dir': PosixPath('data/lang_phone'), 'lm_dir': PosixPath('data/lm'), 'feature_dim': 23, 'search_beam': 20, 'output_beam': 8, 'min_active_states': 30, 'max_active_states': 10000, 'use_double_scores': True, 'epoch': 14, 'avg': 2, 'export': False, 'feature_dir': PosixPath('data/fbank'), 'max_duration': 30.0, 'bucketing_sampler': False, 'num_buckets': 10, 'concatenate_cuts': False, 'duration_factor': 1.0, 'gap': 1.0, 'on_the_fly_feats': False, 'shuffle': False, 'return_cuts': True, 'num_workers': 2, 'env_info': {'k2-version': '1.24.3', 'k2-build-type': 'Release', 'k2-with-cuda': True, 'k2-git-sha1': '4c05309499a08454997adf500b56dcc629e35ae5', 'k2-git-date': 'Tue Jul 25 16:23:36 2023', 'lhotse-version': '1.16.0.dev+git.7640d66.clean', 'torch-version': '1.13.0+cu116', 'torch-cuda-available': False, 'torch-cuda-version': '11.6', 'python-version': '3.8', 'icefall-git-branch': 'master', 'icefall-git-sha1': '3fb0a43-clean', 'icefall-git-date': 'Thu Jul 27 12:36:05 2023', 'icefall-path': '/tmp/icefall', 'k2-path': '/star-fj/fangjun/test-icefall/lib/python3.8/site-packages/k2/__init__.py', 'lhotse-path': '/star-fj/fangjun/test-icefall/lib/python3.8/site-packages/lhotse/__init__.py', 'hostname': 'de-74279-k2-train-1-1220091118-57c4d55446-sph26', 'IP address': '10.177.77.20'}} + 2023-07-27 12:55:12,841 INFO [lexicon.py:168] Loading pre-compiled data/lang_phone/Linv.pt + 2023-07-27 12:55:12,855 INFO [decode.py:273] device: cpu + 2023-07-27 12:55:12,868 INFO [decode.py:291] averaging ['tdnn/exp/epoch-13.pt', 'tdnn/exp/epoch-14.pt'] + 2023-07-27 12:55:12,882 INFO [asr_datamodule.py:218] About to get test cuts + 2023-07-27 12:55:12,883 INFO [asr_datamodule.py:252] About to get test cuts + 2023-07-27 12:55:13,157 INFO [decode.py:204] batch 0/?, cuts processed until now is 4 + 2023-07-27 12:55:13,701 INFO [decode.py:241] The transcripts are stored in tdnn/exp/recogs-test_set.txt + 2023-07-27 12:55:13,702 INFO [utils.py:564] [test_set] %WER 0.42% [1 / 240, 0 ins, 1 del, 0 sub ] + 2023-07-27 12:55:13,704 INFO [decode.py:249] Wrote detailed error stats to tdnn/exp/errs-test_set.txt + 2023-07-27 12:55:13,704 INFO [decode.py:316] Done! -.. code-block:: - 2023-05-12 18:08:30,482 INFO [decode.py:263] Decoding started - 2023-05-12 18:08:30,483 INFO [decode.py:264] {'exp_dir': PosixPath('tdnn/exp'), 'lang_dir': PosixPath('data/lang_phone'), 'lm_dir': PosixPath('data/lm'), 'feature_dim': 23, - 'search_beam': 20, 'output_beam': 8, 'min_active_states': 30, 'max_active_states': 10000, 'use_double_scores': True, 'epoch': 14, 'avg': 2, 'export': False, 'feature_dir': PosixPath('data/fbank'), - 'max_duration': 30.0, 'bucketing_sampler': False, 'num_buckets': 10, 'concatenate_cuts': False, 'duration_factor': 1.0, 'gap': 1.0, 'on_the_fly_feats': False, 'shuffle': False, 'return_cuts': True, - 'num_workers': 2, 'env_info': {'k2-version': '1.24.3', 'k2-build-type': 'Release', 'k2-with-cuda': True, 'k2-git-sha1': '3b7f09fa35e72589914f67089c0da9f196a92ca4', 'k2-git-date': 'Mon May 8 22:58:45 2023', - 'lhotse-version': '1.15.0.dev+git.6fcfced.clean', 'torch-version': '2.0.0+cu118', 'torch-cuda-available': False, 'torch-cuda-version': '11.8', 'python-version': '3.1', 'icefall-git-branch': 'master', - 'icefall-git-sha1': '30bde4b-clean', 'icefall-git-date': 'Thu May 11 17:37:47 2023', 'icefall-path': '/tmp/icefall', - 'k2-path': '/tmp/lib/python3.10/site-packages/k2-1.24.3.dev20230512+cuda11.8.torch2.0.0-py3.10-linux-x86_64.egg/k2/__init__.py', - 'lhotse-path': '/tmp/lib/python3.10/site-packages/lhotse/__init__.py', 'hostname': 'host', 'IP address': '0.0.0.0'}} - 2023-05-12 18:08:30,483 INFO [lexicon.py:168] Loading pre-compiled data/lang_phone/Linv.pt - 2023-05-12 18:08:30,487 INFO [decode.py:273] device: cpu - 2023-05-12 18:08:30,513 INFO [decode.py:291] averaging ['tdnn/exp/epoch-13.pt', 'tdnn/exp/epoch-14.pt'] - 2023-05-12 18:08:30,521 INFO [asr_datamodule.py:218] About to get test cuts - 2023-05-12 18:08:30,521 INFO [asr_datamodule.py:252] About to get test cuts - 2023-05-12 18:08:30,675 INFO [decode.py:204] batch 0/?, cuts processed until now is 4 - 2023-05-12 18:08:30,923 INFO [decode.py:241] The transcripts are stored in tdnn/exp/recogs-test_set.txt - 2023-05-12 18:08:30,924 INFO [utils.py:558] [test_set] %WER 0.42% [1 / 240, 0 ins, 1 del, 0 sub ] - 2023-05-12 18:08:30,925 INFO [decode.py:249] Wrote detailed error stats to tdnn/exp/errs-test_set.txt - 2023-05-12 18:08:30,925 INFO [decode.py:316] Done! - -**Congratulations!** You have successfully setup the environment and have run the first recipe in ``icefall``. +**Congratulations!** You have successfully setup the environment and have run the first recipe in `icefall`_. Have fun with ``icefall``! YouTube Video ------------- -We provide the following YouTube video showing how to install ``icefall``. +We provide the following YouTube video showing how to install `icefall`_. It also shows how to debug various problems that you may encounter while -using ``icefall``. +using `icefall`_. .. note:: diff --git a/decoding-with-langugage-models/rescoring.html b/decoding-with-langugage-models/rescoring.html index 0c5978979..0fdbccb63 100644 --- a/decoding-with-langugage-models/rescoring.html +++ b/decoding-with-langugage-models/rescoring.html @@ -91,11 +91,7 @@

LM rescoring for Transducer

LM rescoring is a commonly used approach to incorporate external LM information. Unlike shallow-fusion-based -<<<<<<< HEAD methods (see Shallow fusion for Transducer, LODR for RNN Transducer), rescoring is usually performed to re-rank the n-best hypotheses after beam search. -======= -methods (see shallow-fusion, LODR for RNN Transducer), rescoring is usually performed to re-rank the n-best hypotheses after beam search. ->>>>>>> 80d922c1583b9b7fb7e9b47008302cdc74ef58b7 Rescoring is usually more efficient than shallow fusion since less computation is performed on the external LM. In this tutorial, we will show you how to use external LM to rescore the n-best hypotheses decoded from neural transducer models in icefall.

diff --git a/index.html b/index.html index 93c9b2ca5..5999e90c4 100644 --- a/index.html +++ b/index.html @@ -92,7 +92,7 @@ speech recognition recipes using Installation
  • (0) Install CUDA toolkit and cuDNN
    • -
  • (1) Install PyTorch and torchaudio
    • +
  • (1) Install torch and torchaudio
  • (2) Install k2
  • (3) Install lhotse
  • (4) Download icefall
    • diff --git a/installation/index.html b/installation/index.html index a7e1bd057..e0caf99a2 100644 --- a/installation/index.html +++ b/installation/index.html @@ -46,16 +46,17 @@
    • Installation
      • (0) Install CUDA toolkit and cuDNN
        • -
      • (1) Install PyTorch and torchaudio
        • +
      • (1) Install torch and torchaudio
      • (2) Install k2
      • (3) Install lhotse
      • (4) Download icefall
      • Installation example
      • Test Your Installation
          @@ -107,8 +108,12 @@

          Installation

          -

          icefall depends on k2 and -lhotse.

          +
          +

          Hint

          +

          We have a colab notebook guiding you step by step to setup the environment.

          +

          yesno colab notebook

          +
          +

          icefall depends on k2 and lhotse.

          We recommend that you use the following steps to install the dependencies.

            • @@ -116,43 +121,20 @@
            • -

            1. Install PyTorch and torchaudio

              2. +

            2. Install torch and torchaudio

            • -

            1. Install k2

              2. +

            2. Install k2

            • -

            1. Install lhotse

              2. +

            2. Install lhotse

          Caution

          -

          99% users who have issues about the installation are using conda.

          -
          -
          -

          Caution

          -

          99% users who have issues about the installation are using conda.

          -
          -
          -

          Caution

          -

          99% users who have issues about the installation are using conda.

          -
          -
          -

          Hint

          -

          We suggest that you use pip install to install PyTorch.

          -

          You can use the following command to create a virutal environment in Python:

          -
          -
          python3 -m venv ./my_env
-source ./my_env/bin/activate
-
          -
          -
          -
          -
          -

          Caution

          Installation order matters.

          @@ -161,15 +143,9 @@ https://k2-fsa.github.io/k2/installation/cuda-cudnn.html to install CUDA and cuDNN.

          -
          -

          (1) Install PyTorch and torchaudio

          -

          Please refer https://pytorch.org/ to install PyTorch -and torchaudio.

          -
          -

          Hint

          -

          You can also go to https://download.pytorch.org/whl/torch_stable.html -to download pre-compiled wheels and install them.

          -
          +
          +

          (1) Install torch and torchaudio

          +

          Please refer https://pytorch.org/ to install torch and torchaudio.

          Caution

          Please install torch and torchaudio at the same time.

          @@ -178,42 +154,40 @@ to download pre-compiled wheels and install them.

          (2) Install k2

          Please refer to https://k2-fsa.github.io/k2/installation/index.html -to install k2.

          +to install k2.

          Caution

          Please don’t change your installed PyTorch after you have installed k2.

          Note

          -

          We suggest that you install k2 from source by following -https://k2-fsa.github.io/k2/installation/from_source.html -or -https://k2-fsa.github.io/k2/installation/for_developers.html.

          +

          We suggest that you install k2 from pre-compiled wheels by following +https://k2-fsa.github.io/k2/installation/from_wheels.html

          Hint

          -

          Please always install the latest version of k2.

          +

          Please always install the latest version of k2.

          (3) Install lhotse

          Please refer to https://lhotse.readthedocs.io/en/latest/getting-started.html#installation -to install lhotse.

          +to install lhotse.

          Hint

          We strongly recommend you to use:

          pip install git+https://github.com/lhotse-speech/lhotse
          -

          to install the latest version of lhotse.

          +

          to install the latest version of lhotse.

          (4) Download icefall

          -

          icefall is a collection of Python scripts; what you need is to download it +

          icefall is a collection of Python scripts; what you need is to download it and set the environment variable PYTHONPATH to point to it.

          -

          Assume you want to place icefall in the folder /tmp. The -following commands show you how to setup icefall:

          +

          Assume you want to place icefall in the folder /tmp. The +following commands show you how to setup icefall:

          cd /tmp
 git clone https://github.com/k2-fsa/icefall
 cd icefall
@@ -223,8 +197,8 @@ pip install -r 
 
          
 
          Hint
          
-
          You can put several versions of icefall in the same virtual environment.
-To switch among different versions of icefall, just set PYTHONPATH
+
          You can put several versions of icefall in the same virtual environment.
+To switch among different versions of icefall, just set PYTHONPATH
 to point to the version you want.
          @@ -233,203 +207,237 @@ 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
+
          kuangfangjun:~$ virtualenv -p python3.8 test-icefall
+created virtual environment CPython3.8.0.final.0-64 in 9422ms
+  creator CPython3Posix(dest=/star-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=/star-fj/fangjun/.local/share/virtualenv)
+    added seed packages: pip==22.3.1, setuptools==65.6.3, wheel==0.38.4
+  activators BashActivator,CShellActivator,FishActivator,NushellActivator,PowerShellActivator,PythonActivator
 
-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
-
          
-
          
-
          -
          -

          (2) Activate your virtual environment

          -
          $ source test-icefall/bin/activate
+kuangfangjun:~$ source test-icefall/bin/activate
+
+(test-icefall) kuangfangjun:~$
          -

          (3) Install k2

          -
          $ pip install k2==1.4.dev20210822+cpu.torch1.9.0 -f https://k2-fsa.org/nightly/index.html
+
          (2) Install CUDA toolkit and cuDNN
          
+
          You need to determine the version of CUDA toolkit to install.
          
+
          (test-icefall) kuangfangjun:~$ nvidia-smi | head -n 4
 
-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
+Wed Jul 26 21:57:49 2023
++-----------------------------------------------------------------------------+
+| NVIDIA-SMI 510.47.03    Driver Version: 510.47.03    CUDA Version: 11.6     |
+|-------------------------------+----------------------+----------------------+
 
          
 
          
-
          
-
          Warning
          
-
          We choose to install a CPU version of k2 for testing. You would probably want to install
-a CUDA version of k2.
          
+
          You can choose any CUDA version that is not greater than the version printed by nvidia-smi.
+In our case, we can choose any version <= 11.6.
          
+
          We will use CUDA 11.6 in this example. Please follow
+https://k2-fsa.github.io/k2/installation/cuda-cudnn.html#cuda-11-6
+to install CUDA toolkit and cuDNN if you have not done that before.
          
+
          After installing CUDA toolkit, you can use the following command to verify it:
          
+
          (test-icefall) kuangfangjun:~$ nvcc --version
+
+nvcc: NVIDIA (R) Cuda compiler driver
+Copyright (c) 2005-2019 NVIDIA Corporation
+Built on Wed_Oct_23_19:24:38_PDT_2019
+Cuda compilation tools, release 10.2, V10.2.89
+
          -

          (4) Install lhotse

          -
          $ pip install git+https://github.com/lhotse-speech/lhotse
+
          (3) Install torch and torchaudio
          
+
          Since we have selected CUDA toolkit 11.6, we have to install a version of torch
+that is compiled against CUDA 11.6. We select torch 1.13.0+cu116 in this
+example.
          
+
          After selecting the version of torch to install, we need to also install
+a compatible version of torchaudio, which is 0.13.0+cu116 in our case.
          
+
          Please refer to https://pytorch.org/audio/stable/installation.html#compatibility-matrix
+to select an appropriate version of torchaudio to install if you use a different
+version of torch.
          
+
          (test-icefall) kuangfangjun:~$ pip install torch==1.13.0+cu116 torchaudio==0.13.0+cu116 -f https://download.pytorch.org/whl/torch_stable.html
 
-Collecting git+https://github.com/lhotse-speech/lhotse
-  Cloning https://github.com/lhotse-speech/lhotse to /tmp/pip-req-build-7b1b76ge
-  Running command git clone -q https://github.com/lhotse-speech/lhotse /tmp/pip-req-build-7b1b76ge
-Collecting audioread>=2.1.9
-  Using cached audioread-2.1.9-py3-none-any.whl
-Collecting SoundFile>=0.10
-  Using cached SoundFile-0.10.3.post1-py2.py3-none-any.whl (21 kB)
-Collecting click>=7.1.1
-  Using cached click-8.0.1-py3-none-any.whl (97 kB)
-Collecting cytoolz>=0.10.1
-  Using cached cytoolz-0.11.0-cp38-cp38-manylinux_2_17_x86_64.manylinux2014_x86_64.whl (1.9 MB)
-Collecting dataclasses
-  Using cached dataclasses-0.6-py3-none-any.whl (14 kB)
-Collecting h5py>=2.10.0
-  Downloading h5py-3.4.0-cp38-cp38-manylinux_2_12_x86_64.manylinux2010_x86_64.whl (4.5 MB)
-     |________________________________| 4.5 MB 684 kB/s
-Collecting intervaltree>=3.1.0
-  Using cached intervaltree-3.1.0-py2.py3-none-any.whl
-Collecting lilcom>=1.1.0
-  Using cached lilcom-1.1.1-cp38-cp38-linux_x86_64.whl
-Collecting numpy>=1.18.1
-  Using cached numpy-1.21.2-cp38-cp38-manylinux_2_12_x86_64.manylinux2010_x86_64.whl (15.8 MB)
-Collecting packaging
-  Using cached packaging-21.0-py3-none-any.whl (40 kB)
-Collecting pyyaml>=5.3.1
-  Using cached PyYAML-5.4.1-cp38-cp38-manylinux1_x86_64.whl (662 kB)
-Collecting tqdm
-  Downloading tqdm-4.62.1-py2.py3-none-any.whl (76 kB)
-     |________________________________| 76 kB 2.7 MB/s
-Collecting torchaudio==0.9.0
-  Downloading torchaudio-0.9.0-cp38-cp38-manylinux1_x86_64.whl (1.9 MB)
-     |________________________________| 1.9 MB 73.1 MB/s
-Requirement already satisfied: torch==1.9.0 in ./test-icefall/lib/python3.8/site-packages (from torchaudio==0.9.0->lhotse===0.8.0.dev
--2a1410b-clean) (1.9.0)
-Requirement already satisfied: typing-extensions in ./test-icefall/lib/python3.8/site-packages (from torch==1.9.0->torchaudio==0.9.0-
->lhotse===0.8.0.dev-2a1410b-clean) (3.10.0.0)
-Collecting toolz>=0.8.0
-  Using cached toolz-0.11.1-py3-none-any.whl (55 kB)
-Collecting sortedcontainers<3.0,>=2.0
-  Using cached sortedcontainers-2.4.0-py2.py3-none-any.whl (29 kB)
-Collecting cffi>=1.0
-  Using cached cffi-1.14.6-cp38-cp38-manylinux1_x86_64.whl (411 kB)
-Collecting pycparser
-  Using cached pycparser-2.20-py2.py3-none-any.whl (112 kB)
-Collecting pyparsing>=2.0.2
-  Using cached pyparsing-2.4.7-py2.py3-none-any.whl (67 kB)
-Building wheels for collected packages: lhotse
-  Building wheel for lhotse (setup.py) ... done
-  Created wheel for lhotse: filename=lhotse-0.8.0.dev_2a1410b_clean-py3-none-any.whl size=342242 sha256=f683444afa4dc0881133206b4646a
-9d0f774224cc84000f55d0a67f6e4a37997
-  Stored in directory: /tmp/pip-ephem-wheel-cache-ftu0qysz/wheels/7f/7a/8e/a0bf241336e2e3cb573e1e21e5600952d49f5162454f2e612f
-  WARNING: Built wheel for lhotse is invalid: Metadata 1.2 mandates PEP 440 version, but '0.8.0.dev-2a1410b-clean' is not
-Failed to build lhotse
-Installing collected packages: pycparser, toolz, sortedcontainers, pyparsing, numpy, cffi, tqdm, torchaudio, SoundFile, pyyaml, packa
-ging, lilcom, intervaltree, h5py, dataclasses, cytoolz, click, audioread, lhotse
-    Running setup.py install for lhotse ... done
-  DEPRECATION: lhotse was installed using the legacy 'setup.py install' method, because a wheel could not be built for it. A possible
- replacement is to fix the wheel build issue reported above. You can find discussion regarding this at https://github.com/pypa/pip/is
-sues/8368.
-Successfully installed SoundFile-0.10.3.post1 audioread-2.1.9 cffi-1.14.6 click-8.0.1 cytoolz-0.11.0 dataclasses-0.6 h5py-3.4.0 inter
-valtree-3.1.0 lhotse-0.8.0.dev-2a1410b-clean lilcom-1.1.1 numpy-1.21.2 packaging-21.0 pycparser-2.20 pyparsing-2.4.7 pyyaml-5.4.1 sor
-tedcontainers-2.4.0 toolz-0.11.1 torchaudio-0.9.0 tqdm-4.62.1
+Looking in links: https://download.pytorch.org/whl/torch_stable.html
+Collecting torch==1.13.0+cu116
+  Downloading https://download.pytorch.org/whl/cu116/torch-1.13.0%2Bcu116-cp38-cp38-linux_x86_64.whl (1983.0 MB)
+     ________________________________________ 2.0/2.0 GB 764.4 kB/s eta 0:00:00
+Collecting torchaudio==0.13.0+cu116
+  Downloading https://download.pytorch.org/whl/cu116/torchaudio-0.13.0%2Bcu116-cp38-cp38-linux_x86_64.whl (4.2 MB)
+     ________________________________________ 4.2/4.2 MB 1.3 MB/s eta 0:00:00
+Requirement already satisfied: typing-extensions in /star-fj/fangjun/test-icefall/lib/python3.8/site-packages (from torch==1.13.0+cu116) (4.7.1)
+Installing collected packages: torch, torchaudio
+Successfully installed torch-1.13.0+cu116 torchaudio-0.13.0+cu116
+
          
+
          
+
          Verify that torch and torchaudio are successfully installed:
          
+
          (test-icefall) kuangfangjun:~$ python3 -c "import torch; print(torch.__version__)"
+
+1.13.0+cu116
+
+(test-icefall) kuangfangjun:~$ python3 -c "import torchaudio; print(torchaudio.__version__)"
+
+0.13.0+cu116
          -

          (5) Download icefall

          -
          $ cd /tmp
-$ git clone https://github.com/k2-fsa/icefall
+
          (4) Install k2
          
+
          We will install k2 from pre-compiled wheels by following
+https://k2-fsa.github.io/k2/installation/from_wheels.html
          
+
          (test-icefall) kuangfangjun:~$ pip install k2==1.24.3.dev20230725+cuda11.6.torch1.13.0 -f https://k2-fsa.github.io/k2/cuda.html
 
-Cloning into 'icefall'...
-remote: Enumerating objects: 500, done.
-remote: Counting objects: 100% (500/500), done.
-remote: Compressing objects: 100% (308/308), done.
-remote: Total 500 (delta 263), reused 307 (delta 102), pack-reused 0
-Receiving objects: 100% (500/500), 172.49 KiB | 385.00 KiB/s, done.
-Resolving deltas: 100% (263/263), done.
+Looking in indexes: https://pypi.tuna.tsinghua.edu.cn/simple
+Looking in links: https://k2-fsa.github.io/k2/cuda.html
+Collecting k2==1.24.3.dev20230725+cuda11.6.torch1.13.0
+  Downloading https://huggingface.co/csukuangfj/k2/resolve/main/ubuntu-cuda/k2-1.24.3.dev20230725%2Bcuda11.6.torch1.13.0-cp38-cp38-manylinux_2_17_x86_64.manylinux2014_x86_64.whl (104.3 MB)
+     ________________________________________ 104.3/104.3 MB 5.1 MB/s eta 0:00:00
+Requirement already satisfied: torch==1.13.0 in /star-fj/fangjun/test-icefall/lib/python3.8/site-packages (from k2==1.24.3.dev20230725+cuda11.6.torch1.13.0) (1.13.0+cu116)
+Collecting graphviz
+  Using cached https://pypi.tuna.tsinghua.edu.cn/packages/de/5e/fcbb22c68208d39edff467809d06c9d81d7d27426460ebc598e55130c1aa/graphviz-0.20.1-py3-none-any.whl (47 kB)
+Requirement already satisfied: typing-extensions in /star-fj/fangjun/test-icefall/lib/python3.8/site-packages (from torch==1.13.0->k2==1.24.3.dev20230725+cuda11.6.torch1.13.0) (4.7.1)
+Installing collected packages: graphviz, k2
+Successfully installed graphviz-0.20.1 k2-1.24.3.dev20230725+cuda11.6.torch1.13.0
+
          
+
          
+
          
+
          Hint
          
+
          Please refer to https://k2-fsa.github.io/k2/cuda.html for the available
+pre-compiled wheels about k2.
          
+
          
+
          Verify that k2 has been installed successfully:
          
+
          (test-icefall) kuangfangjun:~$ python3 -m k2.version
 
-$ cd icefall
-$ pip install -r requirements.txt
+Collecting environment information...
 
-Collecting kaldilm
-  Downloading kaldilm-1.8.tar.gz (48 kB)
-     |________________________________| 48 kB 574 kB/s
-Collecting kaldialign
-  Using cached kaldialign-0.2-cp38-cp38-linux_x86_64.whl
-Collecting sentencepiece>=0.1.96
-  Using cached sentencepiece-0.1.96-cp38-cp38-manylinux_2_17_x86_64.manylinux2014_x86_64.whl (1.2 MB)
-Collecting tensorboard
-  Using cached tensorboard-2.6.0-py3-none-any.whl (5.6 MB)
-Requirement already satisfied: setuptools>=41.0.0 in /ceph-fj/fangjun/test-icefall/lib/python3.8/site-packages (from tensorboard->-r
-requirements.txt (line 4)) (57.4.0)
-Collecting absl-py>=0.4
-  Using cached absl_py-0.13.0-py3-none-any.whl (132 kB)
-Collecting google-auth-oauthlib<0.5,>=0.4.1
-  Using cached google_auth_oauthlib-0.4.5-py2.py3-none-any.whl (18 kB)
-Collecting grpcio>=1.24.3
-  Using cached grpcio-1.39.0-cp38-cp38-manylinux2014_x86_64.whl (4.3 MB)
-Requirement already satisfied: wheel>=0.26 in /ceph-fj/fangjun/test-icefall/lib/python3.8/site-packages (from tensorboard->-r require
-ments.txt (line 4)) (0.36.2)
-Requirement already satisfied: numpy>=1.12.0 in /ceph-fj/fangjun/test-icefall/lib/python3.8/site-packages (from tensorboard->-r requi
-rements.txt (line 4)) (1.21.2)
-Collecting protobuf>=3.6.0
-  Using cached protobuf-3.17.3-cp38-cp38-manylinux_2_5_x86_64.manylinux1_x86_64.whl (1.0 MB)
-Collecting werkzeug>=0.11.15
-  Using cached Werkzeug-2.0.1-py3-none-any.whl (288 kB)
-Collecting tensorboard-data-server<0.7.0,>=0.6.0
-  Using cached tensorboard_data_server-0.6.1-py3-none-manylinux2010_x86_64.whl (4.9 MB)
-Collecting google-auth<2,>=1.6.3
-  Downloading google_auth-1.35.0-py2.py3-none-any.whl (152 kB)
-     |________________________________| 152 kB 1.4 MB/s
-Collecting requests<3,>=2.21.0
-  Using cached requests-2.26.0-py2.py3-none-any.whl (62 kB)
-Collecting tensorboard-plugin-wit>=1.6.0
-  Using cached tensorboard_plugin_wit-1.8.0-py3-none-any.whl (781 kB)
-Collecting markdown>=2.6.8
-  Using cached Markdown-3.3.4-py3-none-any.whl (97 kB)
-Collecting six
-  Using cached six-1.16.0-py2.py3-none-any.whl (11 kB)
-Collecting cachetools<5.0,>=2.0.0
-  Using cached cachetools-4.2.2-py3-none-any.whl (11 kB)
-Collecting rsa<5,>=3.1.4
-  Using cached rsa-4.7.2-py3-none-any.whl (34 kB)
-Collecting pyasn1-modules>=0.2.1
-  Using cached pyasn1_modules-0.2.8-py2.py3-none-any.whl (155 kB)
-Collecting requests-oauthlib>=0.7.0
-  Using cached requests_oauthlib-1.3.0-py2.py3-none-any.whl (23 kB)
-Collecting pyasn1<0.5.0,>=0.4.6
-  Using cached pyasn1-0.4.8-py2.py3-none-any.whl (77 kB)
-Collecting urllib3<1.27,>=1.21.1
-  Using cached urllib3-1.26.6-py2.py3-none-any.whl (138 kB)
-Collecting certifi>=2017.4.17
-  Using cached certifi-2021.5.30-py2.py3-none-any.whl (145 kB)
-Collecting charset-normalizer~=2.0.0
-  Using cached charset_normalizer-2.0.4-py3-none-any.whl (36 kB)
-Collecting idna<4,>=2.5
-  Using cached idna-3.2-py3-none-any.whl (59 kB)
-Collecting oauthlib>=3.0.0
-  Using cached oauthlib-3.1.1-py2.py3-none-any.whl (146 kB)
-Building wheels for collected packages: kaldilm
-  Building wheel for kaldilm (setup.py) ... done
-  Created wheel for kaldilm: filename=kaldilm-1.8-cp38-cp38-linux_x86_64.whl size=897233 sha256=eccb906cafcd45bf9a7e1a1718e4534254bfb
-f4c0d0cbc66eee6c88d68a63862
-  Stored in directory: /root/fangjun/.cache/pip/wheels/85/7d/63/f2dd586369b8797cb36d213bf3a84a789eeb92db93d2e723c9
-Successfully built kaldilm
-Installing collected packages: urllib3, pyasn1, idna, charset-normalizer, certifi, six, rsa, requests, pyasn1-modules, oauthlib, cach
-etools, requests-oauthlib, google-auth, werkzeug, tensorboard-plugin-wit, tensorboard-data-server, protobuf, markdown, grpcio, google
--auth-oauthlib, absl-py, tensorboard, sentencepiece, kaldilm, kaldialign
-Successfully installed absl-py-0.13.0 cachetools-4.2.2 certifi-2021.5.30 charset-normalizer-2.0.4 google-auth-1.35.0 google-auth-oaut
-hlib-0.4.5 grpcio-1.39.0 idna-3.2 kaldialign-0.2 kaldilm-1.8 markdown-3.3.4 oauthlib-3.1.1 protobuf-3.17.3 pyasn1-0.4.8 pyasn1-module
-s-0.2.8 requests-2.26.0 requests-oauthlib-1.3.0 rsa-4.7.2 sentencepiece-0.1.96 six-1.16.0 tensorboard-2.6.0 tensorboard-data-server-0
-.6.1 tensorboard-plugin-wit-1.8.0 urllib3-1.26.6 werkzeug-2.0.1
+k2 version: 1.24.3
+Build type: Release
+Git SHA1: 4c05309499a08454997adf500b56dcc629e35ae5
+Git date: Tue Jul 25 16:23:36 2023
+Cuda used to build k2: 11.6
+cuDNN used to build k2: 8.3.2
+Python version used to build k2: 3.8
+OS used to build k2: CentOS Linux release 7.9.2009 (Core)
+CMake version: 3.27.0
+GCC version: 9.3.1
+CMAKE_CUDA_FLAGS:  -Wno-deprecated-gpu-targets   -lineinfo --expt-extended-lambda -use_fast_math -Xptxas=-w  --expt-extended-lambda -gencode arch=compute_35,code=sm_35  -lineinfo --expt-extended-lambda -use_fast_math -Xptxas=-w  --expt-extended-lambda -gencode arch=compute_50,code=sm_50  -lineinfo --expt-extended-lambda -use_fast_math -Xptxas=-w  --expt-extended-lambda -gencode arch=compute_60,code=sm_60  -lineinfo --expt-extended-lambda -use_fast_math -Xptxas=-w  --expt-extended-lambda -gencode arch=compute_61,code=sm_61  -lineinfo --expt-extended-lambda -use_fast_math -Xptxas=-w  --expt-extended-lambda -gencode arch=compute_70,code=sm_70  -lineinfo --expt-extended-lambda -use_fast_math -Xptxas=-w  --expt-extended-lambda -gencode arch=compute_75,code=sm_75  -lineinfo --expt-extended-lambda -use_fast_math -Xptxas=-w  --expt-extended-lambda -gencode arch=compute_80,code=sm_80  -lineinfo --expt-extended-lambda -use_fast_math -Xptxas=-w  --expt-extended-lambda -gencode arch=compute_86,code=sm_86 -DONNX_NAMESPACE=onnx_c2 -gencode arch=compute_35,code=sm_35 -gencode arch=compute_50,code=sm_50 -gencode arch=compute_52,code=sm_52 -gencode arch=compute_60,code=sm_60 -gencode arch=compute_61,code=sm_61 -gencode arch=compute_70,code=sm_70 -gencode arch=compute_75,code=sm_75 -gencode arch=compute_80,code=sm_80 -gencode arch=compute_86,code=sm_86 -gencode arch=compute_86,code=compute_86 -Xcudafe --diag_suppress=cc_clobber_ignored,--diag_suppress=integer_sign_change,--diag_suppress=useless_using_declaration,--diag_suppress=set_but_not_used,--diag_suppress=field_without_dll_interface,--diag_suppress=base_class_has_different_dll_interface,--diag_suppress=dll_interface_conflict_none_assumed,--diag_suppress=dll_interface_conflict_dllexport_assumed,--diag_suppress=implicit_return_from_non_void_function,--diag_suppress=unsigned_compare_with_zero,--diag_suppress=declared_but_not_referenced,--diag_suppress=bad_friend_decl --expt-relaxed-constexpr --expt-extended-lambda -D_GLIBCXX_USE_CXX11_ABI=0 --compiler-options -Wall  --compiler-options -Wno-strict-overflow  --compiler-options -Wno-unknown-pragmas
+CMAKE_CXX_FLAGS:  -D_GLIBCXX_USE_CXX11_ABI=0 -Wno-unused-variable  -Wno-strict-overflow
+PyTorch version used to build k2: 1.13.0+cu116
+PyTorch is using Cuda: 11.6
+NVTX enabled: True
+With CUDA: True
+Disable debug: True
+Sync kernels : False
+Disable checks: False
+Max cpu memory allocate: 214748364800 bytes (or 200.0 GB)
+k2 abort: False
+__file__: /star-fj/fangjun/test-icefall/lib/python3.8/site-packages/k2/version/version.py
+_k2.__file__: /star-fj/fangjun/test-icefall/lib/python3.8/site-packages/_k2.cpython-38-x86_64-linux-gnu.so
+
          
+
          
+
          +
          +

          (5) Install lhotse

          +
          (test-icefall) kuangfangjun:~$ pip install git+https://github.com/lhotse-speech/lhotse
+
+Collecting git+https://github.com/lhotse-speech/lhotse
+  Cloning https://github.com/lhotse-speech/lhotse to /tmp/pip-req-build-vq12fd5i
+  Running command git clone --filter=blob:none --quiet https://github.com/lhotse-speech/lhotse /tmp/pip-req-build-vq12fd5i
+  Resolved https://github.com/lhotse-speech/lhotse to commit 7640d663469b22cd0b36f3246ee9b849cd25e3b7
+  Installing build dependencies ... done
+  Getting requirements to build wheel ... done
+  Preparing metadata (pyproject.toml) ... done
+Collecting cytoolz>=0.10.1
+  Downloading https://pypi.tuna.tsinghua.edu.cn/packages/1e/3b/a7828d575aa17fb7acaf1ced49a3655aa36dad7e16eb7e6a2e4df0dda76f/cytoolz-0.12.2-cp38-cp38-
+manylinux_2_17_x86_64.manylinux2014_x86_64.whl (2.0 MB)
+     ________________________________________ 2.0/2.0 MB 33.2 MB/s eta 0:00:00
+Collecting pyyaml>=5.3.1
+  Downloading https://pypi.tuna.tsinghua.edu.cn/packages/c8/6b/6600ac24725c7388255b2f5add93f91e58a5d7efaf4af244fdbcc11a541b/PyYAML-6.0.1-cp38-cp38-ma
+nylinux_2_17_x86_64.manylinux2014_x86_64.whl (736 kB)
+     ________________________________________ 736.6/736.6 kB 38.6 MB/s eta 0:00:00
+Collecting dataclasses
+  Downloading https://pypi.tuna.tsinghua.edu.cn/packages/26/2f/1095cdc2868052dd1e64520f7c0d5c8c550ad297e944e641dbf1ffbb9a5d/dataclasses-0.6-py3-none-
+any.whl (14 kB)
+Requirement already satisfied: torchaudio in ./test-icefall/lib/python3.8/site-packages (from lhotse==1.16.0.dev0+git.7640d66.clean) (0.13.0+cu116)
+Collecting lilcom>=1.1.0
+  Downloading https://pypi.tuna.tsinghua.edu.cn/packages/a8/65/df0a69c52bd085ca1ad4e5c4c1a5c680e25f9477d8e49316c4ff1e5084a4/lilcom-1.7-cp38-cp38-many
+linux_2_17_x86_64.manylinux2014_x86_64.whl (87 kB)
+     ________________________________________ 87.1/87.1 kB 8.7 MB/s eta 0:00:00
+Collecting tqdm
+  Using cached https://pypi.tuna.tsinghua.edu.cn/packages/e6/02/a2cff6306177ae6bc73bc0665065de51dfb3b9db7373e122e2735faf0d97/tqdm-4.65.0-py3-none-any
+.whl (77 kB)
+Requirement already satisfied: numpy>=1.18.1 in ./test-icefall/lib/python3.8/site-packages (from lhotse==1.16.0.dev0+git.7640d66.clean) (1.24.4)
+Collecting audioread>=2.1.9
+  Using cached https://pypi.tuna.tsinghua.edu.cn/packages/5d/cb/82a002441902dccbe427406785db07af10182245ee639ea9f4d92907c923/audioread-3.0.0.tar.gz (
+377 kB)
+  Preparing metadata (setup.py) ... done
+Collecting tabulate>=0.8.1
+  Using cached https://pypi.tuna.tsinghua.edu.cn/packages/40/44/4a5f08c96eb108af5cb50b41f76142f0afa346dfa99d5296fe7202a11854/tabulate-0.9.0-py3-none-
+any.whl (35 kB)
+Collecting click>=7.1.1
+  Downloading https://pypi.tuna.tsinghua.edu.cn/packages/1a/70/e63223f8116931d365993d4a6b7ef653a4d920b41d03de7c59499962821f/click-8.1.6-py3-none-any.
+whl (97 kB)
+     ________________________________________ 97.9/97.9 kB 8.4 MB/s eta 0:00:00
+Collecting packaging
+  Using cached https://pypi.tuna.tsinghua.edu.cn/packages/ab/c3/57f0601a2d4fe15de7a553c00adbc901425661bf048f2a22dfc500caf121/packaging-23.1-py3-none-
+any.whl (48 kB)
+Collecting intervaltree>=3.1.0
+  Downloading https://pypi.tuna.tsinghua.edu.cn/packages/50/fb/396d568039d21344639db96d940d40eb62befe704ef849b27949ded5c3bb/intervaltree-3.1.0.tar.gz
+ (32 kB)
+  Preparing metadata (setup.py) ... done
+Requirement already satisfied: torch in ./test-icefall/lib/python3.8/site-packages (from lhotse==1.16.0.dev0+git.7640d66.clean) (1.13.0+cu116)
+Collecting SoundFile>=0.10
+  Downloading https://pypi.tuna.tsinghua.edu.cn/packages/ad/bd/0602167a213d9184fc688b1086dc6d374b7ae8c33eccf169f9b50ce6568c/soundfile-0.12.1-py2.py3-
+none-manylinux_2_17_x86_64.whl (1.3 MB)
+     ________________________________________ 1.3/1.3 MB 46.5 MB/s eta 0:00:00
+Collecting toolz>=0.8.0
+  Using cached https://pypi.tuna.tsinghua.edu.cn/packages/7f/5c/922a3508f5bda2892be3df86c74f9cf1e01217c2b1f8a0ac4841d903e3e9/toolz-0.12.0-py3-none-any.whl (55 kB)
+Collecting sortedcontainers<3.0,>=2.0
+  Using cached https://pypi.tuna.tsinghua.edu.cn/packages/32/46/9cb0e58b2deb7f82b84065f37f3bffeb12413f947f9388e4cac22c4621ce/sortedcontainers-2.4.0-py2.py3-none-any.whl (29 kB)
+Collecting cffi>=1.0
+  Using cached https://pypi.tuna.tsinghua.edu.cn/packages/b7/8b/06f30caa03b5b3ac006de4f93478dbd0239e2a16566d81a106c322dc4f79/cffi-1.15.1-cp38-cp38-manylinux_2_17_x86_64.manylinux2014_x86_64.whl (442 kB)
+Requirement already satisfied: typing-extensions in ./test-icefall/lib/python3.8/site-packages (from torch->lhotse==1.16.0.dev0+git.7640d66.clean) (4.7.1)
+Collecting pycparser
+  Using cached https://pypi.tuna.tsinghua.edu.cn/packages/62/d5/5f610ebe421e85889f2e55e33b7f9a6795bd982198517d912eb1c76e1a53/pycparser-2.21-py2.py3-none-any.whl (118 kB)
+Building wheels for collected packages: lhotse, audioread, intervaltree
+  Building wheel for lhotse (pyproject.toml) ... done
+  Created wheel for lhotse: filename=lhotse-1.16.0.dev0+git.7640d66.clean-py3-none-any.whl size=687627 sha256=cbf0a4d2d0b639b33b91637a4175bc251d6a021a069644ecb1a9f2b3a83d072a
+  Stored in directory: /tmp/pip-ephem-wheel-cache-wwtk90_m/wheels/7f/7a/8e/a0bf241336e2e3cb573e1e21e5600952d49f5162454f2e612f
+  Building wheel for audioread (setup.py) ... done
+  Created wheel for audioread: filename=audioread-3.0.0-py3-none-any.whl size=23704 sha256=5e2d3537c96ce9cf0f645a654c671163707bf8cb8d9e358d0e2b0939a85ff4c2
+  Stored in directory: /star-fj/fangjun/.cache/pip/wheels/e2/c3/9c/f19ae5a03f8862d9f0776b0c0570f1fdd60a119d90954e3f39
+  Building wheel for intervaltree (setup.py) ... done
+  Created wheel for intervaltree: filename=intervaltree-3.1.0-py2.py3-none-any.whl size=26098 sha256=2604170976cfffe0d2f678cb1a6e5b525f561cd50babe53d631a186734fec9f9
+  Stored in directory: /star-fj/fangjun/.cache/pip/wheels/f3/ed/2b/c179ebfad4e15452d6baef59737f27beb9bfb442e0620f7271
+Successfully built lhotse audioread intervaltree
+Installing collected packages: sortedcontainers, dataclasses, tqdm, toolz, tabulate, pyyaml, pycparser, packaging, lilcom, intervaltree, click, audioread, cytoolz, cffi, SoundFile, lhotse
+Successfully installed SoundFile-0.12.1 audioread-3.0.0 cffi-1.15.1 click-8.1.6 cytoolz-0.12.2 dataclasses-0.6 intervaltree-3.1.0 lhotse-1.16.0.dev0+git.7640d66.clean lilcom-1.7 packaging-23.1 pycparser-2.21 pyyaml-6.0.1 sortedcontainers-2.4.0 tabulate-0.9.0 toolz-0.12.0 tqdm-4.65.0
+
          +
          +

          Verify that lhotse has been installed successfully:

          +
          (test-icefall) kuangfangjun:~$ python3 -c "import lhotse; print(lhotse.__version__)"
+
+1.16.0.dev+git.7640d66.clean
+
          +
          +
          +
          +

          (6) Download icefall

          +
          (test-icefall) kuangfangjun:~$ cd /tmp/
+
+(test-icefall) kuangfangjun:tmp$ git clone https://github.com/k2-fsa/icefall
+
+Cloning into 'icefall'...
+remote: Enumerating objects: 12942, done.
+remote: Counting objects: 100% (67/67), done.
+remote: Compressing objects: 100% (56/56), done.
+remote: Total 12942 (delta 17), reused 35 (delta 6), pack-reused 12875
+Receiving objects: 100% (12942/12942), 14.77 MiB | 9.29 MiB/s, done.
+Resolving deltas: 100% (8835/8835), done.
+
+(test-icefall) kuangfangjun:tmp$ cd icefall/
+
+(test-icefall) kuangfangjun:icefall$ pip install -r ./requirements.txt
          @@ -438,65 +446,69 @@ s-0.2.8 requests-2.26.0 requests-oauthlib-1.3.0 rsa-4.7.2 sentencepiece-0.1.96 s

          Test Your Installation

          To test that your installation is successful, let us run the yesno recipe -on CPU.

          +on CPU.

          Data preparation

          -
          $ export PYTHONPATH=/tmp/icefall:$PYTHONPATH
-$ cd /tmp/icefall
-$ cd egs/yesno/ASR
-$ ./prepare.sh
+
          (test-icefall) kuangfangjun:icefall$ export PYTHONPATH=/tmp/icefall:$PYTHONPATH
+
+(test-icefall) kuangfangjun:icefall$ cd /tmp/icefall
+
+(test-icefall) kuangfangjun:icefall$ cd egs/yesno/ASR
+
+(test-icefall) kuangfangjun:ASR$ ./prepare.sh
 
          
 
          
 
          The log of running ./prepare.sh is:
          
-
          2023-05-12 17:55:21 (prepare.sh:27:main) dl_dir: /tmp/icefall/egs/yesno/ASR/download
-2023-05-12 17:55:21 (prepare.sh:30:main) Stage 0: Download data
-/tmp/icefall/egs/yesno/ASR/download/waves_yesno.tar.gz: 100%|_______________________________________________________________| 4.70M/4.70M [06:54<00:00, 11.4kB/s]
-2023-05-12 18:02:19 (prepare.sh:39:main) Stage 1: Prepare yesno manifest
-2023-05-12 18:02:21 (prepare.sh:45:main) Stage 2: Compute fbank for yesno
-2023-05-12 18:02:23,199 INFO [compute_fbank_yesno.py:65] Processing train
-Extracting and storing features: 100%|_______________________________________________________________| 90/90 [00:00<00:00, 212.60it/s]
-2023-05-12 18:02:23,640 INFO [compute_fbank_yesno.py:65] Processing test
-Extracting and storing features: 100%|_______________________________________________________________| 30/30 [00:00<00:00, 304.53it/s]
-2023-05-12 18:02:24 (prepare.sh:51:main) Stage 3: Prepare lang
-2023-05-12 18:02:26 (prepare.sh:66:main) Stage 4: Prepare G
+
          2023-07-27 12:41:39 (prepare.sh:27:main) dl_dir: /tmp/icefall/egs/yesno/ASR/download
+2023-07-27 12:41:39 (prepare.sh:30:main) Stage 0: Download data
+/tmp/icefall/egs/yesno/ASR/download/waves_yesno.tar.gz: 100%|___________________________________________________| 4.70M/4.70M [00:00<00:00, 11.1MB/s]
+2023-07-27 12:41:46 (prepare.sh:39:main) Stage 1: Prepare yesno manifest
+2023-07-27 12:41:50 (prepare.sh:45:main) Stage 2: Compute fbank for yesno
+2023-07-27 12:41:55,718 INFO [compute_fbank_yesno.py:65] Processing train
+Extracting and storing features: 100%|_______________________________________________________________________________| 90/90 [00:01<00:00, 87.82it/s]
+2023-07-27 12:41:56,778 INFO [compute_fbank_yesno.py:65] Processing test
+Extracting and storing features: 100%|______________________________________________________________________________| 30/30 [00:00<00:00, 256.92it/s]
+2023-07-27 12:41:57 (prepare.sh:51:main) Stage 3: Prepare lang
+2023-07-27 12:42:02 (prepare.sh:66:main) Stage 4: Prepare G
 /project/kaldilm/csrc/arpa_file_parser.cc:void kaldilm::ArpaFileParser::Read(std::istream&):79
 [I] Reading \data\ section.
 /project/kaldilm/csrc/arpa_file_parser.cc:void kaldilm::ArpaFileParser::Read(std::istream&):140
 [I] Reading \1-grams: section.
-2023-05-12 18:02:26 (prepare.sh:92:main) Stage 5: Compile HLG
-2023-05-12 18:02:28,581 INFO [compile_hlg.py:124] Processing data/lang_phone
-2023-05-12 18:02:28,582 INFO [lexicon.py:171] Converting L.pt to Linv.pt
-2023-05-12 18:02:28,609 INFO [compile_hlg.py:48] Building ctc_topo. max_token_id: 3
-2023-05-12 18:02:28,610 INFO [compile_hlg.py:52] Loading G.fst.txt
-2023-05-12 18:02:28,611 INFO [compile_hlg.py:62] Intersecting L and G
-2023-05-12 18:02:28,613 INFO [compile_hlg.py:64] LG shape: (4, None)
-2023-05-12 18:02:28,613 INFO [compile_hlg.py:66] Connecting LG
-2023-05-12 18:02:28,614 INFO [compile_hlg.py:68] LG shape after k2.connect: (4, None)
-2023-05-12 18:02:28,614 INFO [compile_hlg.py:70] <class 'torch.Tensor'>
-2023-05-12 18:02:28,614 INFO [compile_hlg.py:71] Determinizing LG
-2023-05-12 18:02:28,615 INFO [compile_hlg.py:74] <class '_k2.ragged.RaggedTensor'>
-2023-05-12 18:02:28,615 INFO [compile_hlg.py:76] Connecting LG after k2.determinize
-2023-05-12 18:02:28,615 INFO [compile_hlg.py:79] Removing disambiguation symbols on LG
-2023-05-12 18:02:28,616 INFO [compile_hlg.py:91] LG shape after k2.remove_epsilon: (6, None)
-2023-05-12 18:02:28,617 INFO [compile_hlg.py:96] Arc sorting LG
-2023-05-12 18:02:28,617 INFO [compile_hlg.py:99] Composing H and LG
-2023-05-12 18:02:28,619 INFO [compile_hlg.py:106] Connecting LG
-2023-05-12 18:02:28,619 INFO [compile_hlg.py:109] Arc sorting LG
-2023-05-12 18:02:28,619 INFO [compile_hlg.py:111] HLG.shape: (8, None)
-2023-05-12 18:02:28,619 INFO [compile_hlg.py:127] Saving HLG.pt to data/lang_phone
+2023-07-27 12:42:02 (prepare.sh:92:main) Stage 5: Compile HLG
+2023-07-27 12:42:07,275 INFO [compile_hlg.py:124] Processing data/lang_phone
+2023-07-27 12:42:07,276 INFO [lexicon.py:171] Converting L.pt to Linv.pt
+2023-07-27 12:42:07,309 INFO [compile_hlg.py:48] Building ctc_topo. max_token_id: 3
+2023-07-27 12:42:07,310 INFO [compile_hlg.py:52] Loading G.fst.txt
+2023-07-27 12:42:07,314 INFO [compile_hlg.py:62] Intersecting L and G
+2023-07-27 12:42:07,323 INFO [compile_hlg.py:64] LG shape: (4, None)
+2023-07-27 12:42:07,323 INFO [compile_hlg.py:66] Connecting LG
+2023-07-27 12:42:07,323 INFO [compile_hlg.py:68] LG shape after k2.connect: (4, None)
+2023-07-27 12:42:07,323 INFO [compile_hlg.py:70] <class 'torch.Tensor'>
+2023-07-27 12:42:07,323 INFO [compile_hlg.py:71] Determinizing LG
+2023-07-27 12:42:07,341 INFO [compile_hlg.py:74] <class '_k2.ragged.RaggedTensor'>
+2023-07-27 12:42:07,341 INFO [compile_hlg.py:76] Connecting LG after k2.determinize
+2023-07-27 12:42:07,341 INFO [compile_hlg.py:79] Removing disambiguation symbols on LG
+2023-07-27 12:42:07,354 INFO [compile_hlg.py:91] LG shape after k2.remove_epsilon: (6, None)
+2023-07-27 12:42:07,445 INFO [compile_hlg.py:96] Arc sorting LG
+2023-07-27 12:42:07,445 INFO [compile_hlg.py:99] Composing H and LG
+2023-07-27 12:42:07,446 INFO [compile_hlg.py:106] Connecting LG
+2023-07-27 12:42:07,446 INFO [compile_hlg.py:109] Arc sorting LG
+2023-07-27 12:42:07,447 INFO [compile_hlg.py:111] HLG.shape: (8, None)
+2023-07-27 12:42:07,447 INFO [compile_hlg.py:127] Saving HLG.pt to data/lang_phone

          Training

          Now let us run the training part:

          -
          $ export CUDA_VISIBLE_DEVICES=""
-$ ./tdnn/train.py
+
          (test-icefall) kuangfangjun:ASR$ export CUDA_VISIBLE_DEVICES=""
+
+(test-icefall) kuangfangjun:ASR$ ./tdnn/train.py
 
          
 
          
 
          
 
          Caution
          
-
          We use export CUDA_VISIBLE_DEVICES="" so that icefall uses CPU
+
          We use export CUDA_VISIBLE_DEVICES="" so that icefall uses CPU
 even if there are GPUs available.
          
 
          
 
          
@@ -511,93 +523,83 @@ even if there are GPUs available.
          
 interested.
          
 
          
 
          The training log is given below:
          
-
           2023-05-12 18:04:59,759 INFO [train.py:481] Training started
- 2023-05-12 18:04:59,759 INFO [train.py:482] {'exp_dir': PosixPath('tdnn/exp'), 'lang_dir': PosixPath('data/lang_phone'), 'lr': 0.01, 'feature_dim': 23, 'weight_decay': 1e-06, 'start_epoch': 0,
- 'best_train_loss': inf, 'best_valid_loss': inf, 'best_train_epoch': -1, 'best_valid_epoch': -1, 'batch_idx_train': 0, 'log_interval': 10, 'reset_interval': 20, 'valid_interval': 10, 'beam_size': 10,
- 'reduction': 'sum', 'use_double_scores': True, 'world_size': 1, 'master_port': 12354, 'tensorboard': True, 'num_epochs': 15, 'seed': 42, 'feature_dir': PosixPath('data/fbank'), 'max_duration': 30.0,
- 'bucketing_sampler': False, 'num_buckets': 10, 'concatenate_cuts': False, 'duration_factor': 1.0, 'gap': 1.0, 'on_the_fly_feats': False, 'shuffle': False, 'return_cuts': True, 'num_workers': 2,
- 'env_info': {'k2-version': '1.24.3', 'k2-build-type': 'Release', 'k2-with-cuda': True, 'k2-git-sha1': '3b7f09fa35e72589914f67089c0da9f196a92ca4', 'k2-git-date': 'Mon May 8 22:58:45 2023',
- 'lhotse-version': '1.15.0.dev+git.6fcfced.clean', 'torch-version': '2.0.0+cu118', 'torch-cuda-available': False, 'torch-cuda-version': '11.8', 'python-version': '3.1', 'icefall-git-branch': 'master',
- 'icefall-git-sha1': '30bde4b-clean', 'icefall-git-date': 'Thu May 11 17:37:47 2023', 'icefall-path': '/tmp/icefall',
- 'k2-path': 'tmp/lib/python3.10/site-packages/k2-1.24.3.dev20230512+cuda11.8.torch2.0.0-py3.10-linux-x86_64.egg/k2/__init__.py',
- 'lhotse-path': 'tmp/lib/python3.10/site-packages/lhotse/__init__.py', 'hostname': 'host', 'IP address': '0.0.0.0'}}
- 2023-05-12 18:04:59,761 INFO [lexicon.py:168] Loading pre-compiled data/lang_phone/Linv.pt
- 2023-05-12 18:04:59,764 INFO [train.py:495] device: cpu
- 2023-05-12 18:04:59,791 INFO [asr_datamodule.py:146] About to get train cuts
- 2023-05-12 18:04:59,791 INFO [asr_datamodule.py:244] About to get train cuts
- 2023-05-12 18:04:59,852 INFO [asr_datamodule.py:149] About to create train dataset
- 2023-05-12 18:04:59,852 INFO [asr_datamodule.py:199] Using SingleCutSampler.
- 2023-05-12 18:04:59,852 INFO [asr_datamodule.py:205] About to create train dataloader
- 2023-05-12 18:04:59,853 INFO [asr_datamodule.py:218] About to get test cuts
- 2023-05-12 18:04:59,853 INFO [asr_datamodule.py:252] About to get test cuts
- 2023-05-12 18:04:59,986 INFO [train.py:422] Epoch 0, batch 0, loss[loss=1.065, over 2436.00 frames. ], tot_loss[loss=1.065, over 2436.00 frames. ], batch size: 4
- 2023-05-12 18:05:00,352 INFO [train.py:422] Epoch 0, batch 10, loss[loss=0.4561, over 2828.00 frames. ], tot_loss[loss=0.7076, over 22192.90 frames. ], batch size: 4
- 2023-05-12 18:05:00,691 INFO [train.py:444] Epoch 0, validation loss=0.9002, over 18067.00 frames.
- 2023-05-12 18:05:00,996 INFO [train.py:422] Epoch 0, batch 20, loss[loss=0.2555, over 2695.00 frames. ], tot_loss[loss=0.484, over 34971.47 frames. ], batch size: 5
- 2023-05-12 18:05:01,217 INFO [train.py:444] Epoch 0, validation loss=0.4688, over 18067.00 frames.
- 2023-05-12 18:05:01,251 INFO [checkpoint.py:75] Saving checkpoint to tdnn/exp/epoch-0.pt
- 2023-05-12 18:05:01,389 INFO [train.py:422] Epoch 1, batch 0, loss[loss=0.2532, over 2436.00 frames. ], tot_loss[loss=0.2532, over 2436.00 frames. ], batch size: 4
- 2023-05-12 18:05:01,637 INFO [train.py:422] Epoch 1, batch 10, loss[loss=0.1139, over 2828.00 frames. ], tot_loss[loss=0.1592, over 22192.90 frames. ], batch size: 4
- 2023-05-12 18:05:01,859 INFO [train.py:444] Epoch 1, validation loss=0.1629, over 18067.00 frames.
- 2023-05-12 18:05:02,094 INFO [train.py:422] Epoch 1, batch 20, loss[loss=0.0767, over 2695.00 frames. ], tot_loss[loss=0.118, over 34971.47 frames. ], batch size: 5
- 2023-05-12 18:05:02,350 INFO [train.py:444] Epoch 1, validation loss=0.06778, over 18067.00 frames.
- 2023-05-12 18:05:02,395 INFO [checkpoint.py:75] Saving checkpoint to tdnn/exp/epoch-1.pt
+
          2023-07-27 12:50:51,936 INFO [train.py:481] Training started
+2023-07-27 12:50:51,936 INFO [train.py:482] {'exp_dir': PosixPath('tdnn/exp'), 'lang_dir': PosixPath('data/lang_phone'), 'lr': 0.01, 'feature_dim': 23, 'weight_decay': 1e-06, 'start_epoch': 0, 'best_train_loss': inf, 'best_valid_loss': inf, 'best_train_epoch': -1, 'best_valid_epoch': -1, 'batch_idx_train': 0, 'log_interval': 10, 'reset_interval': 20, 'valid_interval': 10, 'beam_size': 10, 'reduction': 'sum', 'use_double_scores': True, 'world_size': 1, 'master_port': 12354, 'tensorboard': True, 'num_epochs': 15, 'seed': 42, 'feature_dir': PosixPath('data/fbank'), 'max_duration': 30.0, 'bucketing_sampler': False, 'num_buckets': 10, 'concatenate_cuts': False, 'duration_factor': 1.0, 'gap': 1.0, 'on_the_fly_feats': False, 'shuffle': False, 'return_cuts': True, 'num_workers': 2, 'env_info': {'k2-version': '1.24.3', 'k2-build-type': 'Release', 'k2-with-cuda': True, 'k2-git-sha1': '4c05309499a08454997adf500b56dcc629e35ae5', 'k2-git-date': 'Tue Jul 25 16:23:36 2023', 'lhotse-version': '1.16.0.dev+git.7640d66.clean', 'torch-version': '1.13.0+cu116', 'torch-cuda-available': False, 'torch-cuda-version': '11.6', 'python-version': '3.8', 'icefall-git-branch': 'master', 'icefall-git-sha1': '3fb0a43-clean', 'icefall-git-date': 'Thu Jul 27 12:36:05 2023', 'icefall-path': '/tmp/icefall', 'k2-path': '/star-fj/fangjun/test-icefall/lib/python3.8/site-packages/k2/__init__.py', 'lhotse-path': '/star-fj/fangjun/test-icefall/lib/python3.8/site-packages/lhotse/__init__.py', 'hostname': 'de-74279-k2-train-1-1220091118-57c4d55446-sph26', 'IP address': '10.177.77.20'}}
+2023-07-27 12:50:51,941 INFO [lexicon.py:168] Loading pre-compiled data/lang_phone/Linv.pt
+2023-07-27 12:50:51,949 INFO [train.py:495] device: cpu
+2023-07-27 12:50:51,965 INFO [asr_datamodule.py:146] About to get train cuts
+2023-07-27 12:50:51,965 INFO [asr_datamodule.py:244] About to get train cuts
+2023-07-27 12:50:51,967 INFO [asr_datamodule.py:149] About to create train dataset
+2023-07-27 12:50:51,967 INFO [asr_datamodule.py:199] Using SingleCutSampler.
+2023-07-27 12:50:51,967 INFO [asr_datamodule.py:205] About to create train dataloader
+2023-07-27 12:50:51,968 INFO [asr_datamodule.py:218] About to get test cuts
+2023-07-27 12:50:51,968 INFO [asr_datamodule.py:252] About to get test cuts
+2023-07-27 12:50:52,565 INFO [train.py:422] Epoch 0, batch 0, loss[loss=1.065, over 2436.00 frames. ], tot_loss[loss=1.065, over 2436.00 frames. ], batch size: 4
+2023-07-27 12:50:53,681 INFO [train.py:422] Epoch 0, batch 10, loss[loss=0.4561, over 2828.00 frames. ], tot_loss[loss=0.7076, over 22192.90 frames.], batch size: 4
+2023-07-27 12:50:54,167 INFO [train.py:444] Epoch 0, validation loss=0.9002, over 18067.00 frames.
+2023-07-27 12:50:55,011 INFO [train.py:422] Epoch 0, batch 20, loss[loss=0.2555, over 2695.00 frames. ], tot_loss[loss=0.484, over 34971.47 frames. ], batch size: 5
+2023-07-27 12:50:55,331 INFO [train.py:444] Epoch 0, validation loss=0.4688, over 18067.00 frames.
+2023-07-27 12:50:55,368 INFO [checkpoint.py:75] Saving checkpoint to tdnn/exp/epoch-0.pt
+2023-07-27 12:50:55,633 INFO [train.py:422] Epoch 1, batch 0, loss[loss=0.2532, over 2436.00 frames. ], tot_loss[loss=0.2532, over 2436.00 frames. ],
+ batch size: 4
+2023-07-27 12:50:56,242 INFO [train.py:422] Epoch 1, batch 10, loss[loss=0.1139, over 2828.00 frames. ], tot_loss[loss=0.1592, over 22192.90 frames.], batch size: 4
+2023-07-27 12:50:56,522 INFO [train.py:444] Epoch 1, validation loss=0.1627, over 18067.00 frames.
+2023-07-27 12:50:57,209 INFO [train.py:422] Epoch 1, batch 20, loss[loss=0.07055, over 2695.00 frames. ], tot_loss[loss=0.1175, over 34971.47 frames.], batch size: 5
+2023-07-27 12:50:57,600 INFO [train.py:444] Epoch 1, validation loss=0.07091, over 18067.00 frames.
+2023-07-27 12:50:57,640 INFO [checkpoint.py:75] Saving checkpoint to tdnn/exp/epoch-1.pt
+2023-07-27 12:50:57,847 INFO [train.py:422] Epoch 2, batch 0, loss[loss=0.07731, over 2436.00 frames. ], tot_loss[loss=0.07731, over 2436.00 frames.], batch size: 4
+2023-07-27 12:50:58,427 INFO [train.py:422] Epoch 2, batch 10, loss[loss=0.04391, over 2828.00 frames. ], tot_loss[loss=0.05341, over 22192.90 frames. ], batch size: 4
+2023-07-27 12:50:58,884 INFO [train.py:444] Epoch 2, validation loss=0.04384, over 18067.00 frames.
+2023-07-27 12:50:59,387 INFO [train.py:422] Epoch 2, batch 20, loss[loss=0.03458, over 2695.00 frames. ], tot_loss[loss=0.04616, over 34971.47 frames. ], batch size: 5
+2023-07-27 12:50:59,707 INFO [train.py:444] Epoch 2, validation loss=0.03379, over 18067.00 frames.
+2023-07-27 12:50:59,758 INFO [checkpoint.py:75] Saving checkpoint to tdnn/exp/epoch-2.pt
 
-... ...
+  ... ...
 
- 2023-05-12 18:05:14,789 INFO [train.py:422] Epoch 13, batch 0, loss[loss=0.01056, over 2436.00 frames. ], tot_loss[loss=0.01056, over 2436.00 frames. ], batch size: 4
- 2023-05-12 18:05:15,016 INFO [train.py:422] Epoch 13, batch 10, loss[loss=0.009022, over 2828.00 frames. ], tot_loss[loss=0.009985, over 22192.90 frames. ], batch size: 4
- 2023-05-12 18:05:15,271 INFO [train.py:444] Epoch 13, validation loss=0.01088, over 18067.00 frames.
- 2023-05-12 18:05:15,497 INFO [train.py:422] Epoch 13, batch 20, loss[loss=0.01174, over 2695.00 frames. ], tot_loss[loss=0.01077, over 34971.47 frames. ], batch size: 5
- 2023-05-12 18:05:15,747 INFO [train.py:444] Epoch 13, validation loss=0.01087, over 18067.00 frames.
- 2023-05-12 18:05:15,783 INFO [checkpoint.py:75] Saving checkpoint to tdnn/exp/epoch-13.pt
- 2023-05-12 18:05:15,921 INFO [train.py:422] Epoch 14, batch 0, loss[loss=0.01045, over 2436.00 frames. ], tot_loss[loss=0.01045, over 2436.00 frames. ], batch size: 4
- 2023-05-12 18:05:16,146 INFO [train.py:422] Epoch 14, batch 10, loss[loss=0.008957, over 2828.00 frames. ], tot_loss[loss=0.009903, over 22192.90 frames. ], batch size: 4
- 2023-05-12 18:05:16,374 INFO [train.py:444] Epoch 14, validation loss=0.01092, over 18067.00 frames.
- 2023-05-12 18:05:16,598 INFO [train.py:422] Epoch 14, batch 20, loss[loss=0.01169, over 2695.00 frames. ], tot_loss[loss=0.01065, over 34971.47 frames. ], batch size: 5
- 2023-05-12 18:05:16,824 INFO [train.py:444] Epoch 14, validation loss=0.01077, over 18067.00 frames.
- 2023-05-12 18:05:16,862 INFO [checkpoint.py:75] Saving checkpoint to tdnn/exp/epoch-14.pt
- 2023-05-12 18:05:16,865 INFO [train.py:555] Done!
+2023-07-27 12:51:23,433 INFO [train.py:422] Epoch 13, batch 0, loss[loss=0.01054, over 2436.00 frames. ], tot_loss[loss=0.01054, over 2436.00 frames. ], batch size: 4
+2023-07-27 12:51:23,980 INFO [train.py:422] Epoch 13, batch 10, loss[loss=0.009014, over 2828.00 frames. ], tot_loss[loss=0.009974, over 22192.90 frames. ], batch size: 4
+2023-07-27 12:51:24,489 INFO [train.py:444] Epoch 13, validation loss=0.01085, over 18067.00 frames.
+2023-07-27 12:51:25,258 INFO [train.py:422] Epoch 13, batch 20, loss[loss=0.01172, over 2695.00 frames. ], tot_loss[loss=0.01055, over 34971.47 frames. ], batch size: 5
+2023-07-27 12:51:25,621 INFO [train.py:444] Epoch 13, validation loss=0.01074, over 18067.00 frames.
+2023-07-27 12:51:25,699 INFO [checkpoint.py:75] Saving checkpoint to tdnn/exp/epoch-13.pt
+2023-07-27 12:51:25,866 INFO [train.py:422] Epoch 14, batch 0, loss[loss=0.01044, over 2436.00 frames. ], tot_loss[loss=0.01044, over 2436.00 frames. ], batch size: 4
+2023-07-27 12:51:26,844 INFO [train.py:422] Epoch 14, batch 10, loss[loss=0.008942, over 2828.00 frames. ], tot_loss[loss=0.01, over 22192.90 frames. ], batch size: 4
+2023-07-27 12:51:27,221 INFO [train.py:444] Epoch 14, validation loss=0.01082, over 18067.00 frames.
+2023-07-27 12:51:27,970 INFO [train.py:422] Epoch 14, batch 20, loss[loss=0.01169, over 2695.00 frames. ], tot_loss[loss=0.01054, over 34971.47 frames. ], batch size: 5
+2023-07-27 12:51:28,247 INFO [train.py:444] Epoch 14, validation loss=0.01073, over 18067.00 frames.
+2023-07-27 12:51:28,323 INFO [checkpoint.py:75] Saving checkpoint to tdnn/exp/epoch-14.pt
+2023-07-27 12:51:28,326 INFO [train.py:555] Done!

          Decoding

          Let us use the trained model to decode the test set:

          -
          $ ./tdnn/decode.py
+
          (test-icefall) kuangfangjun:ASR$ ./tdnn/decode.py
+
+2023-07-27 12:55:12,840 INFO [decode.py:263] Decoding started
+2023-07-27 12:55:12,840 INFO [decode.py:264] {'exp_dir': PosixPath('tdnn/exp'), 'lang_dir': PosixPath('data/lang_phone'), 'lm_dir': PosixPath('data/lm'), 'feature_dim': 23, 'search_beam': 20, 'output_beam': 8, 'min_active_states': 30, 'max_active_states': 10000, 'use_double_scores': True, 'epoch': 14, 'avg': 2, 'export': False, 'feature_dir': PosixPath('data/fbank'), 'max_duration': 30.0, 'bucketing_sampler': False, 'num_buckets': 10, 'concatenate_cuts': False, 'duration_factor': 1.0, 'gap': 1.0, 'on_the_fly_feats': False, 'shuffle': False, 'return_cuts': True, 'num_workers': 2, 'env_info': {'k2-version': '1.24.3', 'k2-build-type': 'Release', 'k2-with-cuda': True, 'k2-git-sha1': '4c05309499a08454997adf500b56dcc629e35ae5', 'k2-git-date': 'Tue Jul 25 16:23:36 2023', 'lhotse-version': '1.16.0.dev+git.7640d66.clean', 'torch-version': '1.13.0+cu116', 'torch-cuda-available': False, 'torch-cuda-version': '11.6', 'python-version': '3.8', 'icefall-git-branch': 'master', 'icefall-git-sha1': '3fb0a43-clean', 'icefall-git-date': 'Thu Jul 27 12:36:05 2023', 'icefall-path': '/tmp/icefall', 'k2-path': '/star-fj/fangjun/test-icefall/lib/python3.8/site-packages/k2/__init__.py', 'lhotse-path': '/star-fj/fangjun/test-icefall/lib/python3.8/site-packages/lhotse/__init__.py', 'hostname': 'de-74279-k2-train-1-1220091118-57c4d55446-sph26', 'IP address': '10.177.77.20'}}
+2023-07-27 12:55:12,841 INFO [lexicon.py:168] Loading pre-compiled data/lang_phone/Linv.pt
+2023-07-27 12:55:12,855 INFO [decode.py:273] device: cpu
+2023-07-27 12:55:12,868 INFO [decode.py:291] averaging ['tdnn/exp/epoch-13.pt', 'tdnn/exp/epoch-14.pt']
+2023-07-27 12:55:12,882 INFO [asr_datamodule.py:218] About to get test cuts
+2023-07-27 12:55:12,883 INFO [asr_datamodule.py:252] About to get test cuts
+2023-07-27 12:55:13,157 INFO [decode.py:204] batch 0/?, cuts processed until now is 4
+2023-07-27 12:55:13,701 INFO [decode.py:241] The transcripts are stored in tdnn/exp/recogs-test_set.txt
+2023-07-27 12:55:13,702 INFO [utils.py:564] [test_set] %WER 0.42% [1 / 240, 0 ins, 1 del, 0 sub ]
+2023-07-27 12:55:13,704 INFO [decode.py:249] Wrote detailed error stats to tdnn/exp/errs-test_set.txt
+2023-07-27 12:55:13,704 INFO [decode.py:316] Done!
 
          
 
          
-
          The decoding log is:
          
-
          2023-05-12 18:08:30,482 INFO [decode.py:263] Decoding started
-2023-05-12 18:08:30,483 INFO [decode.py:264] {'exp_dir': PosixPath('tdnn/exp'), 'lang_dir': PosixPath('data/lang_phone'), 'lm_dir': PosixPath('data/lm'), 'feature_dim': 23,
-'search_beam': 20, 'output_beam': 8, 'min_active_states': 30, 'max_active_states': 10000, 'use_double_scores': True, 'epoch': 14, 'avg': 2, 'export': False, 'feature_dir': PosixPath('data/fbank'),
-'max_duration': 30.0, 'bucketing_sampler': False, 'num_buckets': 10, 'concatenate_cuts': False, 'duration_factor': 1.0, 'gap': 1.0, 'on_the_fly_feats': False, 'shuffle': False, 'return_cuts': True,
-'num_workers': 2, 'env_info': {'k2-version': '1.24.3', 'k2-build-type': 'Release', 'k2-with-cuda': True, 'k2-git-sha1': '3b7f09fa35e72589914f67089c0da9f196a92ca4', 'k2-git-date': 'Mon May 8 22:58:45 2023',
-'lhotse-version': '1.15.0.dev+git.6fcfced.clean', 'torch-version': '2.0.0+cu118', 'torch-cuda-available': False, 'torch-cuda-version': '11.8', 'python-version': '3.1', 'icefall-git-branch': 'master',
-'icefall-git-sha1': '30bde4b-clean', 'icefall-git-date': 'Thu May 11 17:37:47 2023', 'icefall-path': '/tmp/icefall',
-'k2-path': '/tmp/lib/python3.10/site-packages/k2-1.24.3.dev20230512+cuda11.8.torch2.0.0-py3.10-linux-x86_64.egg/k2/__init__.py',
-'lhotse-path': '/tmp/lib/python3.10/site-packages/lhotse/__init__.py', 'hostname': 'host', 'IP address': '0.0.0.0'}}
-2023-05-12 18:08:30,483 INFO [lexicon.py:168] Loading pre-compiled data/lang_phone/Linv.pt
-2023-05-12 18:08:30,487 INFO [decode.py:273] device: cpu
-2023-05-12 18:08:30,513 INFO [decode.py:291] averaging ['tdnn/exp/epoch-13.pt', 'tdnn/exp/epoch-14.pt']
-2023-05-12 18:08:30,521 INFO [asr_datamodule.py:218] About to get test cuts
-2023-05-12 18:08:30,521 INFO [asr_datamodule.py:252] About to get test cuts
-2023-05-12 18:08:30,675 INFO [decode.py:204] batch 0/?, cuts processed until now is 4
-2023-05-12 18:08:30,923 INFO [decode.py:241] The transcripts are stored in tdnn/exp/recogs-test_set.txt
-2023-05-12 18:08:30,924 INFO [utils.py:558] [test_set] %WER 0.42% [1 / 240, 0 ins, 1 del, 0 sub ]
-2023-05-12 18:08:30,925 INFO [decode.py:249] Wrote detailed error stats to tdnn/exp/errs-test_set.txt
-2023-05-12 18:08:30,925 INFO [decode.py:316] Done!
-
          
-
          
-
          Congratulations! You have successfully setup the environment and have run the first recipe in icefall.
          
+
          Congratulations! You have successfully setup the environment and have run the first recipe in icefall.
          
 
          Have fun with icefall!

          YouTube Video

          -

          We provide the following YouTube video showing how to install icefall. +

          We provide the following YouTube video showing how to install icefall. It also shows how to debug various problems that you may encounter while -using icefall.

          +using icefall.

          Note

          To get the latest news of next-gen Kaldi, please subscribe diff --git a/recipes/Non-streaming-ASR/yesno/tdnn.html b/recipes/Non-streaming-ASR/yesno/tdnn.html index c977d762c..1cc657757 100644 --- a/recipes/Non-streaming-ASR/yesno/tdnn.html +++ b/recipes/Non-streaming-ASR/yesno/tdnn.html @@ -289,10 +289,10 @@ $ tensorboard dev

          Note there is a URL in the above output, click it and you will see the following screenshot:

          -
          +
          TensorBoard screenshot
          -

          Fig. 8 TensorBoard screenshot.

          +

          Fig. 8 TensorBoard screenshot.

          diff --git a/searchindex.js b/searchindex.js index e6d0ecd3e..53721443e 100644 --- a/searchindex.js +++ b/searchindex.js @@ -1 +1 @@ -Search.setIndex({"docnames": ["contributing/code-style", "contributing/doc", "contributing/how-to-create-a-recipe", "contributing/index", "decoding-with-langugage-models/LODR", "decoding-with-langugage-models/index", "decoding-with-langugage-models/rescoring", "decoding-with-langugage-models/shallow-fusion", "faqs", "huggingface/index", "huggingface/pretrained-models", "huggingface/spaces", "index", "installation/index", "model-export/export-model-state-dict", "model-export/export-ncnn", "model-export/export-ncnn-conv-emformer", "model-export/export-ncnn-lstm", "model-export/export-ncnn-zipformer", "model-export/export-onnx", "model-export/export-with-torch-jit-script", "model-export/export-with-torch-jit-trace", "model-export/index", "recipes/Non-streaming-ASR/aishell/conformer_ctc", "recipes/Non-streaming-ASR/aishell/index", "recipes/Non-streaming-ASR/aishell/stateless_transducer", "recipes/Non-streaming-ASR/aishell/tdnn_lstm_ctc", "recipes/Non-streaming-ASR/index", "recipes/Non-streaming-ASR/librispeech/conformer_ctc", "recipes/Non-streaming-ASR/librispeech/distillation", "recipes/Non-streaming-ASR/librispeech/index", "recipes/Non-streaming-ASR/librispeech/pruned_transducer_stateless", "recipes/Non-streaming-ASR/librispeech/tdnn_lstm_ctc", "recipes/Non-streaming-ASR/librispeech/zipformer_ctc_blankskip", "recipes/Non-streaming-ASR/librispeech/zipformer_mmi", "recipes/Non-streaming-ASR/timit/index", "recipes/Non-streaming-ASR/timit/tdnn_ligru_ctc", "recipes/Non-streaming-ASR/timit/tdnn_lstm_ctc", "recipes/Non-streaming-ASR/yesno/index", "recipes/Non-streaming-ASR/yesno/tdnn", "recipes/Streaming-ASR/index", "recipes/Streaming-ASR/introduction", "recipes/Streaming-ASR/librispeech/index", "recipes/Streaming-ASR/librispeech/lstm_pruned_stateless_transducer", "recipes/Streaming-ASR/librispeech/pruned_transducer_stateless", "recipes/Streaming-ASR/librispeech/zipformer_transducer", "recipes/index"], "filenames": ["contributing/code-style.rst", "contributing/doc.rst", "contributing/how-to-create-a-recipe.rst", "contributing/index.rst", "decoding-with-langugage-models/LODR.rst", "decoding-with-langugage-models/index.rst", "decoding-with-langugage-models/rescoring.rst", "decoding-with-langugage-models/shallow-fusion.rst", "faqs.rst", "huggingface/index.rst", "huggingface/pretrained-models.rst", "huggingface/spaces.rst", "index.rst", "installation/index.rst", "model-export/export-model-state-dict.rst", "model-export/export-ncnn.rst", "model-export/export-ncnn-conv-emformer.rst", "model-export/export-ncnn-lstm.rst", "model-export/export-ncnn-zipformer.rst", "model-export/export-onnx.rst", "model-export/export-with-torch-jit-script.rst", "model-export/export-with-torch-jit-trace.rst", "model-export/index.rst", "recipes/Non-streaming-ASR/aishell/conformer_ctc.rst", "recipes/Non-streaming-ASR/aishell/index.rst", "recipes/Non-streaming-ASR/aishell/stateless_transducer.rst", "recipes/Non-streaming-ASR/aishell/tdnn_lstm_ctc.rst", "recipes/Non-streaming-ASR/index.rst", "recipes/Non-streaming-ASR/librispeech/conformer_ctc.rst", "recipes/Non-streaming-ASR/librispeech/distillation.rst", "recipes/Non-streaming-ASR/librispeech/index.rst", "recipes/Non-streaming-ASR/librispeech/pruned_transducer_stateless.rst", "recipes/Non-streaming-ASR/librispeech/tdnn_lstm_ctc.rst", "recipes/Non-streaming-ASR/librispeech/zipformer_ctc_blankskip.rst", "recipes/Non-streaming-ASR/librispeech/zipformer_mmi.rst", "recipes/Non-streaming-ASR/timit/index.rst", "recipes/Non-streaming-ASR/timit/tdnn_ligru_ctc.rst", "recipes/Non-streaming-ASR/timit/tdnn_lstm_ctc.rst", "recipes/Non-streaming-ASR/yesno/index.rst", "recipes/Non-streaming-ASR/yesno/tdnn.rst", "recipes/Streaming-ASR/index.rst", "recipes/Streaming-ASR/introduction.rst", "recipes/Streaming-ASR/librispeech/index.rst", "recipes/Streaming-ASR/librispeech/lstm_pruned_stateless_transducer.rst", "recipes/Streaming-ASR/librispeech/pruned_transducer_stateless.rst", "recipes/Streaming-ASR/librispeech/zipformer_transducer.rst", "recipes/index.rst"], "titles": ["Follow the code style", "Contributing to Documentation", "How to create a recipe", "Contributing", "LODR for RNN Transducer", "Decoding with language models", "LM rescoring for Transducer", "Shallow fusion for Transducer", "Frequently Asked Questions (FAQs)", "Huggingface", "Pre-trained models", "Huggingface spaces", "Icefall", "Installation", "Export model.state_dict()", "Export to ncnn", "Export ConvEmformer transducer models to ncnn", "Export LSTM transducer models to ncnn", "Export streaming Zipformer transducer models to ncnn", "Export to ONNX", "Export model with torch.jit.script()", "Export model with torch.jit.trace()", "Model export", "Conformer CTC", "aishell", "Stateless Transducer", "TDNN-LSTM CTC", "Non Streaming ASR", "Conformer CTC", "Distillation with HuBERT", "LibriSpeech", "Pruned transducer statelessX", "TDNN-LSTM-CTC", "Zipformer CTC Blank Skip", "Zipformer MMI", "TIMIT", "TDNN-LiGRU-CTC", "TDNN-LSTM-CTC", "YesNo", "TDNN-CTC", "Streaming ASR", "Introduction", "LibriSpeech", "LSTM Transducer", "Pruned transducer statelessX", "Zipformer Transducer", "Recipes"], "terms": {"we": [0, 1, 2, 3, 4, 6, 7, 8, 10, 11, 13, 14, 15, 16, 17, 18, 19, 20, 21, 22, 23, 24, 25, 26, 28, 29, 31, 32, 33, 34, 36, 37, 39, 41, 43, 44, 45, 46], "us": [0, 1, 2, 4, 5, 6, 7, 8, 9, 11, 12, 13, 15, 16, 17, 18, 19, 22, 23, 24, 25, 26, 28, 29, 32, 36, 37, 39, 41], "tool": [0, 8, 16], "make": [0, 1, 3, 16, 17, 18, 23, 25, 28, 41], "consist": [0, 25, 31, 43, 44, 45], "possibl": [0, 2, 3, 13, 23, 28], "black": 0, "format": [0, 16, 17, 18, 23, 25, 26, 28, 31, 32, 33, 34, 36, 37, 39, 43, 44, 45], "flake8": 0, "check": [0, 28], "qualiti": [0, 24], "isort": 0, "sort": [0, 13], "import": [0, 8, 16, 44, 45], "The": [0, 1, 2, 4, 7, 8, 11, 13, 14, 16, 17, 18, 23, 24, 26, 28, 29, 31, 32, 33, 34, 36, 37, 39, 41, 43, 44, 45], "version": [0, 12, 13, 14, 16, 17, 18, 23, 25, 26, 28, 31, 32, 36, 37, 44], "abov": [0, 4, 6, 7, 8, 13, 14, 16, 17, 18, 19, 23, 24, 25, 26, 28, 31, 33, 34, 39, 41, 43, 44, 45], "ar": [0, 1, 3, 4, 5, 6, 7, 8, 13, 14, 16, 17, 18, 23, 24, 25, 26, 28, 29, 31, 32, 33, 34, 36, 37, 39, 43, 44, 45, 46], "22": [0, 13, 16, 17, 28, 36, 37, 39], "3": [0, 4, 6, 7, 8, 12, 14, 15, 19, 22, 26, 29, 31, 32, 33, 34, 39, 43, 44, 45], "0": [0, 1, 4, 6, 7, 12, 14, 16, 17, 18, 19, 23, 25, 26, 28, 29, 31, 32, 33, 34, 36, 37, 39, 43, 44, 45], "5": [0, 7, 15, 22, 23, 25, 26, 28, 29, 31, 32, 33, 34, 36, 37, 39, 43, 44, 45], "4": [0, 4, 5, 6, 7, 8, 12, 14, 15, 22, 23, 25, 26, 28, 29, 31, 32, 33, 34, 36, 37, 39, 43, 44, 45], "10": [0, 7, 12, 13, 14, 16, 17, 18, 23, 25, 26, 28, 29, 31, 32, 33, 34, 36, 37, 39, 43, 44, 45], "1": [0, 4, 6, 7, 12, 14, 15, 19, 20, 21, 22, 29, 31, 32, 33, 34, 36, 37, 39, 43, 44, 45], "after": [0, 1, 6, 11, 13, 14, 16, 17, 18, 23, 25, 26, 28, 29, 31, 32, 33, 34, 36, 37, 39, 41, 43, 44, 45], "run": [0, 2, 8, 11, 13, 16, 17, 18, 19, 22, 23, 25, 26, 28, 29, 31, 32, 33, 34, 36, 37, 39, 43, 44, 45], "command": [0, 1, 4, 6, 7, 8, 13, 14, 16, 17, 21, 23, 25, 26, 28, 29, 31, 32, 33, 34, 36, 37, 39, 43, 44, 45], "git": [0, 4, 6, 7, 13, 14, 16, 17, 18, 19, 23, 25, 26, 28, 32, 36, 37, 39], "clone": [0, 4, 6, 7, 13, 14, 16, 17, 18, 19, 23, 25, 26, 28, 32, 36, 37, 39], "http": [0, 1, 2, 4, 6, 7, 8, 10, 11, 13, 14, 15, 16, 17, 18, 19, 20, 21, 23, 24, 25, 26, 28, 29, 31, 32, 33, 34, 36, 37, 39, 43, 44, 45], "github": [0, 2, 6, 10, 13, 14, 15, 16, 17, 18, 19, 20, 21, 23, 25, 26, 28, 31, 32, 33, 34, 36, 37, 39, 43, 44, 45], "com": [0, 2, 6, 10, 11, 13, 14, 16, 17, 20, 21, 23, 25, 26, 28, 29, 31, 32, 33, 34, 36, 37, 39, 43, 44, 45], "k2": [0, 2, 8, 10, 11, 12, 14, 15, 16, 17, 18, 19, 20, 21, 23, 25, 26, 28, 31, 32, 33, 34, 36, 37, 43, 44, 45], "fsa": [0, 2, 10, 11, 13, 14, 15, 16, 17, 18, 19, 20, 21, 23, 25, 28, 31, 33, 34, 43, 44, 45], "icefal": [0, 2, 3, 4, 6, 7, 8, 10, 11, 14, 15, 19, 20, 21, 22, 23, 25, 26, 28, 29, 31, 32, 33, 34, 36, 37, 39, 41, 43, 44, 45, 46], "cd": [0, 1, 2, 8, 13, 14, 16, 17, 18, 19, 20, 21, 23, 25, 26, 28, 29, 31, 32, 33, 34, 36, 37, 39, 43, 44, 45], "pip": [0, 1, 6, 8, 13, 16, 19, 25], "instal": [0, 1, 4, 6, 8, 9, 11, 12, 14, 15, 19, 22, 29, 31, 33, 34, 39, 43, 44, 45], "pre": [0, 3, 4, 6, 7, 9, 11, 12, 13, 15, 22, 29], "commit": 0, "whenev": 0, "you": [0, 1, 2, 4, 6, 7, 8, 10, 11, 13, 14, 16, 17, 18, 19, 20, 21, 23, 25, 26, 28, 29, 31, 32, 33, 34, 36, 37, 39, 41, 43, 44, 45], "automat": [0, 11, 29], "hook": 0, "invok": 0, "fail": [0, 13], "If": [0, 2, 4, 6, 7, 8, 11, 16, 17, 18, 20, 21, 23, 25, 26, 28, 29, 31, 32, 33, 34, 36, 37, 39, 41, 43, 44, 45], "ani": [0, 4, 6, 7, 13, 23, 25, 26, 28, 29, 31, 33, 34, 39, 43, 44], "your": [0, 1, 2, 4, 6, 7, 9, 11, 12, 16, 17, 18, 19, 23, 25, 26, 28, 29, 31, 32, 33, 34, 36, 37, 39, 43, 44, 45], "wa": [0, 13, 14, 28, 32], "success": [0, 13, 16, 17], "pleas": [0, 1, 2, 4, 6, 7, 8, 11, 13, 15, 16, 17, 18, 19, 20, 21, 23, 25, 26, 28, 29, 31, 32, 33, 34, 36, 37, 39, 41, 43, 44, 45], "fix": [0, 8, 13, 16, 17, 18, 28], "issu": [0, 4, 6, 7, 8, 13, 16, 17, 28, 29, 44, 45], "report": [0, 8, 13, 29], "some": [0, 1, 4, 6, 14, 16, 17, 23, 25, 26, 28, 31, 32, 33, 34, 36, 37, 39, 43, 44, 45], "i": [0, 1, 2, 4, 7, 8, 11, 13, 14, 15, 16, 17, 18, 19, 23, 24, 25, 26, 28, 29, 31, 32, 33, 34, 36, 37, 39, 41, 43, 44, 45], "e": [0, 2, 4, 5, 6, 7, 16, 17, 18, 23, 25, 26, 28, 29, 31, 32, 33, 34, 36, 37, 39, 43, 44, 45], "modifi": [0, 15, 22, 23, 26, 28, 29, 31, 32, 33, 34, 36, 37, 39, 41, 43, 44, 45], "file": [0, 2, 11, 12, 14, 16, 17, 18, 20, 21, 22, 23, 25, 26, 28, 29, 31, 32, 33, 34, 36, 37, 39, 43, 44, 45], "place": [0, 13, 14, 25, 28, 32], "so": [0, 4, 6, 7, 11, 12, 13, 14, 16, 17, 18, 23, 25, 26, 28, 29, 31, 32, 33, 34, 36, 37, 39, 43, 44, 45], "statu": 0, "failur": 0, "see": [0, 1, 6, 7, 11, 13, 16, 17, 18, 19, 20, 21, 23, 25, 26, 28, 29, 31, 32, 33, 34, 36, 37, 39, 41, 43, 44, 45], "which": [0, 2, 4, 6, 7, 11, 14, 16, 17, 18, 19, 23, 24, 25, 26, 28, 29, 31, 32, 33, 34, 36, 37, 39, 44, 45], "ha": [0, 2, 12, 15, 16, 17, 18, 19, 23, 25, 26, 28, 31, 32, 33, 34, 36, 37, 41, 43, 44, 45], "been": [0, 15, 16, 17, 18, 25], "befor": [0, 1, 14, 16, 17, 18, 19, 20, 23, 25, 26, 28, 29, 31, 33, 34, 43, 44, 45], "further": [0, 4, 6, 7], "chang": [0, 4, 6, 7, 8, 13, 16, 17, 18, 23, 25, 26, 28, 29, 31, 32, 33, 34, 36, 37, 39, 43, 44, 45], "all": [0, 10, 11, 14, 16, 17, 18, 20, 23, 25, 26, 28, 29, 31, 32, 33, 34, 36, 37, 39, 41, 43, 44, 45], "again": [0, 16, 17, 39], "should": [0, 2, 4, 6, 16, 17, 18, 23, 25, 26, 28, 29, 31, 32, 33, 34, 36, 37, 39, 43, 44, 45], "succe": 0, "thi": [0, 2, 3, 4, 5, 6, 7, 8, 9, 13, 14, 16, 17, 18, 19, 20, 21, 22, 23, 25, 26, 28, 29, 31, 32, 33, 34, 36, 37, 39, 41, 43, 44, 45, 46], "time": [0, 13, 16, 17, 18, 23, 25, 26, 28, 29, 31, 32, 33, 34, 36, 37, 39, 41, 43, 44, 45], "succeed": 0, "want": [0, 4, 6, 7, 13, 14, 20, 21, 23, 25, 26, 28, 29, 31, 32, 33, 34, 36, 37, 39, 41, 43, 44, 45], "can": [0, 1, 2, 4, 6, 7, 8, 10, 11, 13, 14, 15, 16, 17, 18, 19, 20, 21, 23, 24, 25, 26, 28, 29, 31, 32, 33, 34, 36, 37, 39, 41, 43, 44, 45], "do": [0, 2, 4, 6, 23, 25, 26, 28, 29, 31, 32, 33, 34, 36, 37, 39, 41, 43, 44, 45], "Or": 0, "without": [0, 4, 6, 7, 9, 11, 23, 28], "your_changed_fil": 0, "py": [0, 2, 4, 6, 7, 8, 13, 16, 17, 18, 19, 20, 21, 22, 23, 25, 26, 28, 29, 31, 32, 33, 34, 36, 37, 39, 43, 44, 45], "sphinx": 1, "write": [1, 2, 3], "have": [1, 2, 4, 6, 7, 10, 11, 13, 14, 16, 17, 18, 19, 23, 25, 26, 28, 29, 31, 32, 33, 34, 36, 37, 39, 41, 43, 44, 45], "prepar": [1, 3, 4, 14], "environ": [1, 8, 16, 17, 18, 23, 24, 25, 26, 28, 29, 31, 32, 36, 37, 39, 44, 45], "doc": [1, 14, 41], "r": [1, 13, 16, 17, 18, 36, 37], "requir": [1, 4, 6, 13, 18, 29, 44, 45], "txt": [1, 4, 13, 16, 17, 18, 19, 23, 25, 26, 28, 32, 36, 37, 39], "set": [1, 4, 6, 7, 8, 13, 16, 17, 18, 23, 25, 26, 28, 29, 31, 33, 34, 39, 43, 44, 45], "up": [1, 13, 14, 16, 17, 18, 23, 26, 28, 29, 31, 32, 33, 34, 44, 45], "readi": [1, 23, 28, 29], "refer": [1, 2, 6, 7, 13, 14, 15, 16, 17, 18, 20, 21, 23, 25, 26, 28, 31, 32, 33, 36, 37, 39, 41, 44, 45], "restructuredtext": 1, "primer": 1, "familiar": 1, "build": [1, 13, 14, 16, 17, 18, 23, 25, 28], "local": [1, 13, 31, 33, 34, 43, 44, 45], "preview": 1, "what": [1, 2, 13, 16, 17, 18, 25, 41], "look": [1, 2, 4, 6, 7, 10, 13, 16, 17, 18, 23, 25, 26, 28, 29], "like": [1, 2, 11, 16, 17, 18, 23, 25, 26, 28, 31, 33, 34, 39, 41, 43, 44], "publish": [1, 14, 24], "html": [1, 2, 8, 13, 15, 16, 17, 18, 19, 20, 21, 31, 43, 44, 45], "gener": [1, 6, 14, 16, 17, 18, 19, 20, 21, 23, 25, 26, 28, 29, 31, 33, 34, 43, 44, 45], "view": [1, 16, 17, 18, 23, 25, 26, 28, 31, 33, 34, 39, 43, 44, 45], "follow": [1, 2, 3, 4, 5, 6, 7, 8, 10, 11, 12, 13, 14, 15, 16, 17, 18, 19, 20, 21, 23, 25, 26, 28, 29, 31, 32, 33, 34, 36, 37, 39, 43, 44, 45], "python3": [1, 8, 13, 17, 18], "m": [1, 13, 16, 17, 18, 25, 31, 33, 34, 36, 37, 43, 44, 45], "server": [1, 11, 13, 43], "It": [1, 2, 6, 7, 9, 13, 15, 16, 17, 18, 19, 20, 21, 23, 24, 25, 26, 28, 31, 32, 33, 34, 36, 37, 39, 41, 43, 44, 45], "print": [1, 23, 25, 26, 28, 29, 31, 32, 33, 34, 36, 37, 39, 43, 44, 45], "serv": [1, 31, 33, 34, 43, 44, 45], "port": [1, 29, 31, 33, 34, 43, 44, 45], "8000": [1, 39], "open": [1, 4, 6, 7, 12, 14, 16, 17, 18, 24, 25, 28, 29], "browser": [1, 9, 11, 31, 33, 34, 43, 44, 45], "go": [1, 7, 13, 23, 25, 28, 31, 33, 34, 43, 44, 45], "read": [2, 13, 14, 16, 17, 18, 23, 25, 26, 28, 29, 31, 32, 33, 34, 36, 37, 39, 43, 44, 45], "code": [2, 3, 8, 12, 16, 17, 18, 23, 28, 29, 31, 32, 36, 37, 39, 41, 44, 45], "style": [2, 3, 12], "adjust": 2, "sytl": 2, "design": 2, "python": [2, 13, 14, 16, 17, 18, 19, 20, 21, 23, 25, 28, 31, 33, 34, 43, 44, 45], "recommend": [2, 6, 7, 13, 23, 25, 26, 28, 29, 31, 44, 45], "test": [2, 4, 12, 14, 15, 22, 23, 25, 26, 28, 29, 32, 33, 36, 37], "valid": [2, 13, 18, 23, 25, 26, 28, 31, 32, 33, 34, 36, 37, 39, 43, 44, 45], "dataset": [2, 8, 13, 14, 23, 25, 26, 28, 29, 31, 32, 33, 34, 36, 37, 39, 41, 43, 44, 45], "lhots": [2, 12, 14, 16, 17, 18, 23, 25, 28], "readthedoc": [2, 13], "io": [2, 13, 15, 16, 17, 18, 19, 20, 21, 31, 43, 44, 45], "en": [2, 13, 16], "latest": [2, 11, 13, 28, 29, 31, 32, 33, 34, 43, 44, 45], "index": [2, 13, 15, 16, 17, 18, 19, 20, 21, 43, 44, 45], "yesno": [2, 8, 12, 13, 27, 39, 46], "veri": [2, 3, 7, 16, 17, 18, 25, 36, 37, 39, 44, 45], "good": [2, 7], "exampl": [2, 11, 12, 14, 16, 17, 18, 20, 21, 22, 29, 32, 36, 37, 39], "speech": [2, 11, 12, 13, 15, 24, 25, 39, 46], "pull": [2, 4, 6, 7, 16, 17, 18, 19, 23, 25, 28, 41], "380": [2, 16, 37], "show": [2, 4, 6, 7, 11, 13, 14, 16, 17, 18, 23, 25, 26, 28, 29, 31, 32, 33, 34, 36, 37, 39, 41, 43, 44, 45], "add": [2, 16, 17, 18, 23, 25, 26, 44, 46], "new": [2, 3, 11, 13, 16, 17, 18, 23, 24, 25, 26, 28, 29, 31, 32, 33, 34, 39, 43, 44, 45], "suppos": [2, 44, 45], "would": [2, 13, 14, 16, 17, 18, 28, 32, 44, 45], "name": [2, 8, 14, 16, 17, 18, 19, 23, 25, 31, 33, 34, 44, 45], "foo": [2, 21, 23, 28, 31, 33, 34, 43, 44, 45], "eg": [2, 8, 10, 13, 14, 16, 17, 18, 19, 20, 21, 23, 25, 26, 28, 29, 31, 32, 33, 34, 36, 37, 39, 43, 44, 45], "mkdir": [2, 16, 17, 23, 25, 26, 28, 32, 36, 37, 39], "p": [2, 4, 13, 16, 17, 25, 36, 37], "asr": [2, 4, 6, 7, 8, 10, 12, 13, 14, 16, 17, 18, 19, 20, 21, 23, 25, 26, 28, 29, 31, 32, 33, 34, 36, 37, 39, 41, 43, 44, 45, 46], "touch": 2, "sh": [2, 13, 14, 23, 25, 26, 28, 29, 31, 32, 33, 34, 36, 37, 39, 43, 44, 45], "chmod": 2, "x": [2, 4, 18, 41], "simpl": [2, 25], "own": [2, 29, 31, 44, 45], "otherwis": [2, 16, 17, 18, 23, 25, 28, 29, 31, 33, 34, 43, 44, 45], "librispeech": [2, 4, 6, 7, 8, 10, 12, 14, 16, 17, 18, 19, 20, 21, 27, 28, 29, 31, 32, 33, 34, 40, 41, 43, 44, 45, 46], "assum": [2, 4, 13, 14, 16, 17, 18, 19, 23, 25, 26, 28, 29, 31, 32, 36, 37, 39, 43, 44, 45], "fanci": 2, "call": [2, 8, 19, 29], "bar": [2, 21, 23, 28, 31, 33, 34, 43, 44, 45], "organ": 2, "wai": [2, 3, 22, 31, 33, 34, 41, 43, 44, 45], "readm": [2, 23, 25, 26, 28, 32, 36, 37, 39], "md": [2, 10, 14, 23, 25, 26, 28, 31, 32, 33, 34, 36, 37, 39, 43, 44, 45], "asr_datamodul": [2, 8, 13], "pretrain": [2, 4, 6, 7, 14, 16, 17, 18, 19, 21, 23, 25, 26, 28, 32, 36, 37, 39], "For": [2, 4, 6, 7, 8, 10, 14, 16, 17, 18, 23, 25, 26, 28, 29, 31, 32, 33, 34, 36, 37, 39, 43, 44, 45], "instanc": [2, 8, 10, 16, 17, 18, 23, 25, 26, 28, 31, 32, 33, 34, 36, 37, 39, 43, 44, 45], "tdnn": [2, 8, 13, 24, 27, 30, 35, 38], "its": [2, 4, 14, 15, 16, 17, 18, 21, 25, 33], "directori": [2, 12, 13, 16, 17, 18, 23, 25, 26, 28, 29, 31, 32, 33, 34, 36, 37, 39, 43, 44, 45], "structur": [2, 18], "descript": [2, 23, 25, 26, 28, 31, 32, 33, 34, 36, 37, 39, 43, 44, 45], "contain": [2, 12, 14, 15, 16, 17, 18, 23, 25, 26, 28, 29, 31, 32, 33, 34, 36, 37, 39, 43, 44, 45, 46], "inform": [2, 4, 6, 14, 23, 25, 26, 28, 31, 32, 33, 36, 37, 39, 41, 43, 44, 45], "g": [2, 4, 5, 6, 7, 13, 18, 23, 25, 26, 28, 29, 31, 32, 33, 34, 36, 37, 39, 43, 44, 45], "wer": [2, 13, 14, 31, 32, 33, 34, 36, 37, 39, 43, 44, 45], "etc": [2, 23, 25, 26, 28, 29, 31, 32, 33, 34, 36, 37, 39, 41, 43, 44, 45], "provid": [2, 11, 13, 14, 15, 16, 17, 18, 23, 24, 25, 26, 28, 29, 31, 32, 33, 34, 36, 37, 39, 43, 44, 45, 46], "pytorch": [2, 8, 12, 16, 17, 18, 25], "dataload": [2, 13], "take": [2, 7, 14, 29, 31, 39, 44, 45], "input": [2, 14, 16, 17, 18, 23, 25, 26, 28, 32, 36, 37, 39, 41], "checkpoint": [2, 4, 6, 7, 13, 14, 16, 17, 18, 23, 25, 26, 28, 31, 32, 33, 34, 36, 37, 39, 43, 44, 45], "save": [2, 13, 14, 17, 18, 20, 23, 25, 26, 28, 29, 31, 32, 33, 34, 36, 37, 39, 43, 44, 45], "dure": [2, 4, 5, 7, 8, 11, 14, 23, 25, 26, 28, 29, 31, 32, 33, 34, 36, 37, 39, 43, 44, 45], "stage": [2, 13, 23, 25, 26, 28, 29, 31, 32, 33, 34, 36, 37, 39, 43, 44, 45], "": [2, 4, 6, 7, 13, 14, 16, 17, 18, 19, 20, 23, 25, 26, 28, 29, 31, 32, 33, 34, 36, 37, 39, 43, 44, 45], "definit": [2, 16, 17], "neural": [2, 4, 6, 7, 23, 28], "network": [2, 23, 25, 28, 31, 33, 34, 43, 44, 45], "script": [2, 6, 7, 12, 13, 21, 22, 23, 25, 26, 28, 29, 32, 36, 37, 39, 43], "infer": [2, 14, 16, 17], "tdnn_lstm_ctc": [2, 26, 32, 37], "conformer_ctc": [2, 23, 28], "get": [2, 11, 13, 16, 17, 18, 23, 25, 26, 28, 29, 31, 32, 33, 34, 39, 41, 43, 44, 45], "feel": [2, 29, 43], "result": [2, 4, 7, 10, 11, 14, 16, 17, 18, 23, 25, 26, 28, 29, 31, 32, 33, 34, 36, 37, 39, 43, 44, 45], "everi": [2, 14, 31, 33, 34, 43, 44, 45], "kept": [2, 31, 44, 45], "self": [2, 15, 18, 41], "toler": 2, "duplic": 2, "among": [2, 13], "differ": [2, 13, 16, 17, 18, 19, 23, 24, 28, 29, 31, 41, 43, 44, 45], "invoc": [2, 16, 17], "help": [2, 23, 25, 26, 28, 31, 32, 33, 34, 36, 37, 39, 43, 44, 45], "blob": [2, 10, 14, 21, 31, 33, 34, 43, 44, 45], "master": [2, 6, 10, 13, 14, 17, 18, 20, 21, 25, 29, 31, 33, 34, 43, 44, 45], "transform": [2, 6, 7, 23, 28, 43], "conform": [2, 20, 24, 25, 27, 30, 31, 33, 43, 44, 45], "base": [2, 4, 7, 18, 23, 25, 26, 28, 29, 31, 33, 34, 43, 44, 45], "lstm": [2, 15, 21, 22, 24, 27, 30, 35, 40, 42], "attent": [2, 18, 25, 26, 29, 41, 44, 45], "lm": [2, 4, 7, 12, 13, 25, 31, 32, 36, 37, 39, 44, 45], "rescor": [2, 12, 26, 32, 34, 36, 37, 39], "demonstr": [2, 9, 11, 14, 19], "consid": [2, 4, 18], "colab": 2, "notebook": 2, "welcom": 3, "There": [3, 4, 16, 17, 18, 19, 23, 25, 26, 28, 29, 31, 33, 34, 43, 44, 45], "mani": [3, 44, 45], "two": [3, 4, 16, 17, 18, 23, 25, 26, 28, 29, 31, 32, 33, 34, 36, 37, 39, 41, 43, 44, 45], "them": [3, 5, 6, 9, 10, 11, 13, 16, 18, 23, 25, 26, 28, 29, 31, 32, 33, 34, 36, 37, 39, 43, 44, 45], "To": [3, 4, 6, 7, 11, 13, 23, 25, 26, 28, 29, 31, 32, 33, 34, 36, 37, 39, 43, 44, 45], "document": [3, 12, 14, 15, 16, 17, 18, 19, 34], "repositori": [3, 16, 17, 18, 19], "recip": [3, 4, 6, 7, 10, 12, 13, 14, 19, 23, 25, 26, 28, 29, 31, 32, 36, 37, 39, 41, 43, 44, 45], "In": [3, 4, 6, 8, 11, 13, 14, 16, 17, 18, 19, 20, 21, 22, 23, 25, 26, 28, 29, 32, 36, 37, 39, 41], "page": [3, 11, 20, 23, 25, 26, 28, 29, 31, 32, 33, 34, 36, 37, 39, 41, 43, 44, 45, 46], "describ": [3, 5, 9, 14, 16, 17, 19, 20, 21, 22, 23, 25, 26, 28, 31, 32, 36, 37, 44, 45], "how": [3, 4, 5, 6, 7, 9, 11, 12, 13, 16, 17, 18, 19, 22, 23, 25, 26, 28, 29, 31, 32, 33, 34, 36, 37, 39, 41, 43, 44, 45], "creat": [3, 4, 6, 7, 12, 14, 16, 17, 18, 23, 25, 26, 28, 31, 32, 33, 34, 36, 37, 39, 43, 44], "data": [3, 4, 6, 7, 14, 16, 17, 18, 19, 20, 21, 24], "train": [3, 4, 6, 7, 8, 9, 11, 12, 14, 15, 20, 21, 22, 41], "decod": [3, 4, 8, 11, 12, 16, 17, 18, 21, 22], "model": [3, 4, 6, 7, 9, 11, 12, 13, 15, 29, 41], "As": [4, 5, 6, 7, 16, 25, 28, 29], "type": [4, 6, 7, 13, 14, 16, 17, 18, 23, 25, 28, 31, 33, 34, 39, 41, 43, 44, 45], "e2": [4, 7], "usual": [4, 6, 7, 23, 25, 26, 28, 29, 31, 33, 34, 43, 44, 45], "an": [4, 5, 6, 7, 11, 13, 14, 16, 17, 18, 19, 20, 21, 23, 24, 25, 28, 29, 31, 34, 39, 43, 44, 45], "intern": [4, 5], "languag": [4, 7, 11, 12, 23, 25, 26], "learn": [4, 23, 25, 26, 28, 31, 32, 33, 34, 36, 37, 39, 43, 44, 45], "level": [4, 5], "corpu": [4, 6, 7, 24], "real": 4, "life": 4, "scenario": 4, "often": [4, 23, 25, 26, 28, 31, 33, 34, 43, 44, 45], "mismatch": [4, 44], "between": [4, 7, 31, 44, 45], "target": [4, 11], "space": [4, 9, 12], "problem": [4, 6, 7, 13, 29], "when": [4, 6, 8, 11, 16, 17, 18, 22, 25, 28, 29, 31, 33, 34, 44, 45], "act": 4, "against": 4, "extern": [4, 5, 6, 7], "tutori": [4, 6, 7, 23, 25, 26, 28, 29, 31, 32, 33, 34, 36, 37, 43, 44, 45], "low": [4, 16, 17], "order": [4, 13, 16, 17, 18, 23, 26, 28, 32, 36, 37], "densiti": 4, "ratio": 4, "allevi": 4, "effect": [4, 7, 18], "improv": [4, 5, 6, 7, 25], "perform": [4, 6, 7, 15, 25, 29, 44], "languga": 4, "integr": [4, 11], "pruned_transducer_stateless7_stream": [4, 6, 7, 18, 19, 45], "stream": [4, 6, 7, 12, 15, 16, 17, 19, 22, 23, 28, 36, 37, 43, 46], "howev": [4, 6, 7, 14, 17, 29], "easili": [4, 6, 7, 23, 26, 28], "appli": [4, 6, 7, 25, 41], "other": [4, 7, 14, 17, 18, 19, 25, 28, 29, 31, 32, 36, 37, 39, 41, 44, 45, 46], "encount": [4, 6, 7, 8, 13, 18, 23, 25, 26, 28, 29, 31, 33, 34, 43, 44, 45], "here": [4, 6, 7, 14, 16, 17, 18, 23, 25, 26, 28, 29, 32, 41, 44], "simplic": [4, 6, 7], "same": [4, 6, 7, 13, 14, 16, 17, 18, 23, 25, 26, 28, 29, 31, 32, 33, 34, 36, 37, 39, 41, 43, 44, 45], "domain": [4, 6, 7], "gigaspeech": [4, 6, 7, 10, 20, 43], "first": [4, 6, 8, 13, 16, 17, 18, 23, 25, 26, 28, 29, 31, 32, 33, 34, 36, 37, 39, 43, 44, 45], "let": [4, 6, 7, 13, 16, 17, 18, 23, 28], "background": 4, "predecessor": 4, "dr": 4, "propos": [4, 25, 41, 45], "address": [4, 11, 13, 14, 16, 17, 18, 25, 31, 34, 43, 44, 45], "sourc": [4, 13, 14, 16, 17, 18, 23, 24, 25, 28], "acoust": [4, 44, 45], "similar": [4, 5, 29, 33, 44, 45], "deriv": 4, "formular": 4, "bay": 4, "theorem": 4, "text": [4, 6, 7, 16, 17, 18, 23, 25, 26, 28, 31, 32, 33, 34, 36, 37, 39, 43, 44, 45], "score": [4, 5, 7, 23, 28, 31, 44, 45], "left": [4, 16, 18, 25, 44, 45], "y_u": 4, "mathit": 4, "y": 4, "right": [4, 16, 25, 41, 44], "log": [4, 8, 13, 16, 17, 18, 32, 36, 37, 39], "y_": 4, "u": [4, 13, 16, 17, 18, 23, 25, 26, 28, 29, 39], "lambda_1": 4, "p_": 4, "lambda_2": 4, "where": [4, 8, 44], "weight": [4, 23, 26, 28, 33, 34, 43], "respect": 4, "onli": [4, 6, 14, 16, 17, 18, 23, 25, 26, 28, 29, 31, 32, 33, 34, 36, 37, 39, 41, 43, 44, 45, 46], "compar": [4, 16, 17, 18, 44], "shallow": [4, 12], "fusion": [4, 12], "subtract": [4, 5], "work": [4, 16, 17, 18, 28], "treat": [4, 17, 18], "predictor": 4, "joiner": [4, 16, 17, 18, 19, 21, 25, 31, 43, 44, 45], "weak": 4, "captur": 4, "therefor": [4, 8], "n": [4, 5, 6, 23, 29, 31, 33, 34, 36, 37, 43, 44, 45], "gram": [4, 6, 13, 23, 25, 26, 31, 32, 34, 36, 37, 44, 45], "approxim": [4, 5], "ilm": 4, "lead": [4, 7], "formula": 4, "rnnt": [4, 31, 44, 45], "bi": [4, 6], "addit": 4, "estim": 4, "comar": 4, "li": 4, "choic": 4, "accord": 4, "origin": [4, 5], "paper": [4, 5, 29, 31, 43, 44, 45], "achiev": [4, 6, 7, 41], "both": [4, 31, 33, 34, 41, 43, 44, 45], "intra": 4, "cross": 4, "much": [4, 16, 17], "faster": [4, 6], "evalu": 4, "now": [4, 6, 13, 16, 17, 18, 23, 28, 29, 31, 32, 33, 34, 36, 37, 43, 44, 45], "illustr": [4, 6, 7], "purpos": [4, 6, 7, 16, 17], "from": [4, 6, 7, 8, 9, 11, 13, 14, 16, 17, 18, 19, 23, 24, 25, 26, 28, 29, 31, 32, 33, 34, 36, 37, 39, 41, 43, 44, 45], "link": [4, 6, 7, 10, 13, 14, 15, 31, 33, 34, 43, 44, 45], "scratch": [4, 6, 7, 31, 33, 34, 43, 44, 45], "prune": [4, 6, 7, 14, 18, 19, 25, 27, 29, 30, 40, 41, 42, 43, 45], "statelessx": [4, 6, 7, 27, 29, 30, 40, 41, 42], "initi": [4, 6, 7, 23, 26], "step": [4, 6, 7, 13, 14, 16, 17, 18, 23, 25, 26, 28, 29, 31, 33, 34, 39, 43, 44, 45], "download": [4, 6, 7, 8, 11, 12, 15, 22, 24, 29], "git_lfs_skip_smudg": [4, 6, 7, 16, 17, 18, 19], "huggingfac": [4, 6, 7, 10, 12, 14, 16, 17, 18, 19, 23, 25, 26, 28, 32, 33, 34, 36, 37, 39, 43], "co": [4, 6, 7, 10, 11, 14, 16, 17, 18, 19, 23, 24, 25, 26, 28, 32, 33, 34, 36, 37, 39, 43], "zengwei": [4, 6, 7, 16, 18, 19, 34, 43], "stateless7": [4, 6, 7, 18, 19], "2022": [4, 6, 7, 14, 16, 17, 18, 19, 25, 31, 33, 34, 43, 44], "12": [4, 6, 7, 13, 14, 16, 17, 18, 19, 23, 25, 26, 28, 31, 33, 34, 36, 39, 43, 44, 45], "29": [4, 6, 7, 13, 18, 19, 23, 25, 26, 28, 32, 33, 36, 37], "pushd": [4, 6, 7, 19], "exp": [4, 6, 7, 13, 14, 16, 17, 18, 19, 20, 21, 23, 25, 26, 28, 29, 31, 32, 33, 34, 36, 37, 39, 43, 44, 45], "lf": [4, 6, 7, 14, 16, 17, 18, 19, 23, 25, 26, 28, 32, 34, 36, 37, 39], "includ": [4, 6, 7, 16, 17, 18, 19, 31, 33, 34, 43, 44, 45], "pt": [4, 6, 7, 13, 14, 16, 17, 18, 19, 20, 21, 23, 25, 26, 28, 31, 32, 33, 34, 36, 37, 39, 43, 44, 45], "ln": [4, 6, 7, 14, 16, 17, 18, 19, 23, 28, 31, 33, 34, 43, 44, 45], "epoch": [4, 6, 7, 13, 14, 16, 17, 18, 19, 20, 23, 25, 26, 28, 29, 31, 32, 33, 34, 36, 37, 39, 43, 44, 45], "99": [4, 6, 7, 13, 16, 17, 18, 19], "symbol": [4, 5, 6, 7, 13, 25, 31, 44, 45], "load": [4, 6, 7, 13, 16, 17, 18, 23, 25, 26, 28, 31, 32, 33, 34, 36, 37, 39, 43, 44, 45], "done": [4, 6, 7, 13, 14, 23, 25, 26, 28, 31, 32, 33, 34, 36, 37, 39, 43, 44, 45], "via": [4, 6, 7, 13, 15, 20, 21, 22], "exp_dir": [4, 6, 7, 13, 16, 17, 18, 25, 28, 29, 31, 33, 34, 44, 45], "avg": [4, 6, 7, 13, 14, 16, 17, 18, 19, 20, 21, 25, 28, 29, 31, 32, 33, 34, 36, 37, 39, 43, 44, 45], "averag": [4, 6, 7, 13, 14, 16, 17, 18, 19, 23, 25, 26, 28, 31, 32, 33, 34, 36, 37, 39, 43, 44, 45], "fals": [4, 6, 7, 13, 14, 16, 17, 18, 23, 25, 28, 29], "dir": [4, 6, 7, 14, 16, 17, 18, 19, 20, 21, 23, 25, 26, 28, 29, 31, 32, 33, 34, 36, 37, 39, 43, 44, 45], "bpe": [4, 5, 6, 7, 14, 16, 17, 18, 19, 20, 21, 28, 31, 33, 34, 43, 44, 45], "lang_bpe_500": [4, 6, 7, 14, 16, 17, 18, 19, 20, 21, 28, 31, 33, 34, 43, 44, 45], "max": [4, 6, 7, 14, 16, 17, 23, 25, 26, 28, 29, 31, 33, 34, 43, 44, 45], "durat": [4, 6, 7, 14, 23, 25, 26, 28, 29, 31, 32, 33, 34, 36, 37, 39, 43, 44, 45], "600": [4, 6, 7, 14, 28, 31, 33, 43, 44, 45], "chunk": [4, 6, 7, 16, 18, 19, 44, 45], "len": [4, 6, 7, 18, 19, 45], "32": [4, 6, 7, 16, 17, 18, 19, 23, 25, 26, 45], "method": [4, 7, 11, 13, 14, 23, 25, 26, 28, 29, 31, 32, 33, 34, 36, 37, 43, 44, 45], "modified_beam_search": [4, 5, 6, 7, 11, 25, 29, 31, 33, 43, 44, 45], "clean": [4, 13, 18, 23, 25, 28, 29, 31, 32, 33, 34, 43, 44, 45], "beam_size_4": [4, 6, 7], "11": [4, 6, 7, 8, 13, 16, 17, 19, 23, 25, 26, 28, 31, 32, 33, 34, 36, 37, 39, 43, 44, 45], "best": [4, 5, 6, 7, 16, 17, 18, 23, 26, 28], "7": [4, 6, 7, 13, 14, 15, 18, 22, 23, 26, 28, 31, 32, 36, 37, 43, 44], "93": [4, 6, 7], "Then": [4, 6], "necessari": [4, 29], "note": [4, 5, 6, 7, 8, 14, 16, 23, 25, 26, 28, 31, 32, 33, 34, 36, 37, 39, 43, 44, 45], "960": [4, 28, 31, 33, 34, 43, 44, 45], "hour": [4, 23, 25, 26, 28, 31, 33, 34, 43, 44, 45], "ezerhouni": [4, 6, 7], "popd": [4, 6, 7, 19], "marcoyang": [4, 6], "librispeech_bigram": [4, 6], "2gram": [4, 6], "fst": [4, 13, 25, 39], "modified_beam_search_lm_lodr": 4, "lm_dir": [4, 6, 7, 13, 28], "lm_scale": [4, 6, 7], "42": [4, 13, 17, 23, 28, 39], "lodr_scal": 4, "24": [4, 8, 13, 16, 17, 26, 32, 36, 37, 39], "scale": [4, 6, 7, 16, 17, 23, 28, 29, 32, 34, 36, 37], "embed": [4, 6, 7, 25, 31, 43, 44, 45], "dim": [4, 6, 7, 16, 17, 18, 25, 31, 44], "2048": [4, 6, 7, 14, 16, 17, 18, 25], "hidden": [4, 6, 7, 17, 43], "num": [4, 6, 7, 16, 17, 18, 23, 25, 26, 28, 29, 31, 33, 34, 43, 44, 45], "layer": [4, 6, 7, 16, 17, 18, 25, 29, 31, 41, 43, 44, 45], "vocab": [4, 6, 7, 28], "500": [4, 6, 7, 13, 14, 16, 17, 18, 25, 28, 34, 43], "token": [4, 16, 17, 18, 19, 23, 25, 26, 28, 32, 36, 37, 39], "ngram": [4, 28, 32, 36, 37], "2": [4, 6, 7, 12, 14, 15, 22, 29, 31, 32, 33, 34, 36, 37, 39, 43, 44, 45], "extra": [4, 16, 17, 18, 25, 41, 44], "argument": [4, 7, 29, 41], "need": [4, 6, 11, 13, 14, 15, 16, 17, 18, 23, 25, 26, 28, 29, 31, 32, 33, 34, 36, 37, 39, 41, 43, 44, 45], "given": [4, 13, 14, 16, 17, 18, 23, 25, 26, 28, 31, 32, 33, 34, 44, 45], "specifi": [4, 7, 8, 16, 17, 18, 23, 25, 26, 28, 29, 31, 32, 33, 34, 36, 37, 39, 43, 44, 45], "neg": [4, 25], "number": [4, 7, 11, 14, 16, 17, 18, 23, 25, 26, 28, 31, 32, 33, 34, 36, 37, 39, 43, 44, 45], "obtain": [4, 7, 23, 25, 26, 28, 32, 36, 37], "shown": [4, 7], "below": [4, 7, 13, 16, 17, 18, 23, 25, 26, 28, 31, 32, 33, 34, 36, 37, 39, 43, 44], "61": [4, 6], "6": [4, 6, 7, 8, 13, 15, 22, 23, 25, 28, 31, 32, 36, 37, 43], "74": [4, 6, 13, 14], "recal": 4, "lowest": [4, 31, 33, 34, 43, 44, 45], "77": [4, 6, 7, 13, 28], "08": [4, 6, 7, 13, 18, 28, 32, 34, 36, 37, 39, 43], "inde": 4, "even": [4, 11, 13, 17], "better": [4, 6], "increas": [4, 6, 23, 25, 26, 28, 31, 33, 34, 43, 44, 45], "8": [4, 6, 7, 8, 13, 14, 16, 17, 18, 23, 25, 28, 29, 31, 32, 33, 34, 39, 43, 44, 45], "45": [4, 6, 13, 16, 18, 23, 25, 28], "38": [4, 6, 16, 23, 25, 28, 36], "23": [4, 6, 8, 13, 16, 17, 18, 23, 25, 26, 28, 36, 37, 39], "section": [5, 8, 9, 13, 14, 19, 20, 21, 22, 23, 28], "langugag": 5, "transduc": [5, 12, 14, 15, 19, 22, 24, 27, 29, 30, 40, 41, 42], "avail": [5, 6, 13, 14, 16, 17, 18, 23, 25, 28, 32, 36, 37, 39, 43], "beam": [5, 14, 43], "search": [5, 6, 7, 10, 11], "realli": [5, 23, 26, 28, 31, 33, 34, 43, 44, 45], "valu": [5, 7, 16, 17, 18, 23, 25, 26, 28, 31, 33, 34, 43, 44, 45], "rnn": [5, 6, 7, 12, 17, 25, 31, 33, 43, 44, 45], "t": [5, 13, 16, 17, 18, 19, 20, 23, 25, 26, 28, 29, 31, 32, 33, 34, 36, 37, 39, 43, 44, 45], "doe": [5, 16, 17, 18, 23, 25, 28, 39], "modified_beam_search_lm_shallow_fus": [5, 6, 7], "interpol": 5, "also": [5, 6, 7, 9, 10, 13, 14, 15, 16, 17, 18, 19, 21, 23, 25, 26, 28, 31, 33, 34, 39, 41, 43, 44, 45], "known": 5, "modified_beam_search_lodr": [5, 6], "bigram": 5, "backoff": 5, "modified_beam_search_lm_rescor": [5, 6], "hypothes": [5, 6], "rnnlm": [5, 6], "re": [5, 6, 8, 23, 26, 28, 29, 31, 33, 34, 41, 43, 44, 45], "rank": [5, 6], "modified_beam_search_lm_rescore_lodr": [5, 6], "lodr": [5, 12], "commonli": [6, 7, 23, 25, 26, 28, 32, 36, 37, 39], "approach": 6, "incorpor": 6, "unlik": 6, "head": [6, 25, 41], "80d922c1583b9b7fb7e9b47008302cdc74ef58b7": 6, "more": [6, 13, 16, 17, 18, 23, 28, 29, 39, 41, 43, 44], "effici": [6, 7, 31, 44, 45], "than": [6, 14, 17, 23, 25, 26, 28, 31, 32, 33, 34, 39, 43, 44, 45], "sinc": [6, 16, 17, 18, 29, 39, 43], "less": [6, 14, 28, 32, 39, 44, 45], "comput": [6, 13, 14, 16, 17, 18, 23, 25, 26, 29, 31, 32, 34, 36, 37, 39, 43, 44, 45], "gpu": [6, 7, 13, 16, 17, 23, 25, 26, 28, 29, 31, 33, 34, 36, 37, 39, 43, 44, 45], "try": [6, 8, 9, 11, 29, 31, 33, 34, 43, 44, 45], "might": [6, 7, 17, 18, 44, 45], "ideal": [6, 7], "mai": [6, 7, 13, 16, 17, 18, 23, 25, 26, 28, 31, 33, 34, 43, 44, 45, 46], "With": 6, "43": [6, 17, 18, 28], "great": 6, "made": [6, 16], "boost": [6, 7], "tabl": [6, 11, 16, 17, 18], "67": [6, 13], "59": [6, 13, 16, 26, 28], "86": 6, "fact": 6, "arpa": [6, 39], "performn": 6, "depend": [6, 13, 23, 28], "kenlm": 6, "kpu": 6, "archiv": 6, "zip": 6, "execut": [6, 7, 16, 23, 26, 28, 29, 31, 32, 33, 34, 36, 37, 39, 43, 44, 45], "9": [6, 13, 16, 17, 18, 23, 25, 26, 28, 31, 32, 33, 34, 36, 39, 43, 44, 45], "57": [6, 13, 17, 28, 32], "slightli": 6, "63": [6, 13, 25], "04": [6, 13, 16, 17, 18, 23, 25, 26, 28, 32, 36, 37], "52": [6, 13, 23, 28], "73": [6, 13], "mention": 6, "earlier": 6, "benchmark": [6, 25], "speed": [6, 16, 23, 25, 26, 28, 31, 33, 34, 43, 44, 45], "132": [6, 13], "95": [6, 24], "177": [6, 14, 17, 18, 25, 26, 28], "96": [6, 13], "210": [6, 36, 37], "262": [6, 7], "62": [6, 7, 13, 28, 32], "65": [6, 7, 13, 16], "352": [6, 7, 13, 28], "58": [6, 7, 8, 13, 28], "488": [6, 7, 16, 17, 18], "400": [6, 24], "610": [6, 13], "870": 6, "156": 6, "203": [6, 14, 28], "255": [6, 17, 18], "160": 6, "263": [6, 13, 17], "singl": [6, 23, 25, 26, 28, 31, 32, 33, 34, 36, 37, 39, 43, 44, 45], "32g": 6, "v100": [6, 23, 25, 26, 28], "vari": 6, "word": [7, 23, 25, 26, 28, 32, 36, 37, 39], "error": [7, 8, 13, 16, 17, 18, 28], "rate": [7, 23, 25, 26, 28, 31, 32, 33, 34, 36, 37, 39, 43, 44, 45], "These": [7, 23, 25, 26, 28, 31, 32, 33, 34, 36, 37, 39, 43, 44, 45], "alreadi": [7, 13, 14], "But": [7, 16, 31, 33, 34, 43, 44, 45], "long": [7, 16], "true": [7, 13, 14, 16, 17, 18, 23, 25, 28, 29, 31, 32, 33, 34, 36, 37, 39, 43, 44, 45], "either": [7, 11, 23, 25, 26, 28, 44, 45], "choos": [7, 11, 13, 29, 31, 33, 34, 43, 44, 45], "three": [7, 16, 17, 18, 21, 23, 25, 41], "associ": 7, "dimens": [7, 31, 44, 45], "obviou": 7, "rel": 7, "reduct": [7, 13, 16, 17, 33], "around": 7, "A": [7, 13, 14, 16, 17, 18, 23, 25, 26, 28, 31, 32, 33, 34, 43, 44, 45], "few": [7, 16, 17, 18, 29], "paramet": [7, 14, 16, 17, 18, 20, 23, 25, 26, 28, 31, 32, 33, 34, 36, 37, 43, 44, 45], "tune": [7, 16, 17, 18, 23, 25, 26, 28, 29, 31, 33, 34, 43, 44, 45], "control": [7, 23, 25, 26, 28, 29, 31, 32, 33, 34, 36, 37, 39, 43, 44, 45], "too": 7, "small": [7, 25, 36, 37, 39], "fulli": 7, "util": [7, 8, 13, 28], "larg": 7, "domin": 7, "bad": 7, "typic": [7, 23, 25, 26, 28], "activ": [7, 11], "path": [7, 11, 13, 14, 16, 17, 18, 21, 23, 25, 26, 28, 29, 31, 33, 34, 43, 44, 45], "trade": 7, "off": [7, 16], "accuraci": [7, 16, 17, 24], "larger": [7, 17, 23, 25, 26, 28, 31, 33, 34, 43, 44, 45], "slower": 7, "collect": [8, 13], "user": [8, 13], "post": 8, "correspond": [8, 10, 11], "solut": 8, "One": 8, "torch": [8, 12, 13, 14, 15, 22, 23, 25, 28], "torchaudio": [8, 12, 41], "cu111": 8, "torchvis": 8, "f": [8, 13, 36, 37], "org": [8, 13, 24, 25, 31, 43, 44, 45], "whl": [8, 13], "torch_stabl": [8, 13], "throw": [8, 16, 17, 18], "cuda": [8, 12, 14, 16, 17, 18, 23, 25, 26, 28, 31, 32, 33, 34, 36, 37, 43, 44, 45], "while": [8, 13, 16, 17, 18, 23, 25, 26, 28, 29, 31, 33, 34, 43, 44, 45], "That": [8, 16, 17, 29, 31, 43, 44, 45], "cu11": 8, "correct": 8, "traceback": 8, "most": [8, 44, 45], "recent": [8, 16, 17, 18], "last": 8, "line": [8, 13, 16, 17, 18, 31, 44, 45], "14": [8, 13, 14, 16, 17, 20, 23, 28, 31, 32, 33, 36, 43, 44, 45], "yesnoasrdatamodul": 8, "home": [8, 16, 17, 23, 28], "xxx": [8, 14, 16, 17, 18], "next": [8, 11, 13, 16, 17, 18, 28, 29, 31, 32, 33, 34, 43, 44, 45], "gen": [8, 11, 13, 28, 29, 31, 32, 33, 34, 43, 44, 45], "kaldi": [8, 11, 13, 28, 29, 31, 32, 33, 34, 43, 44, 45], "34": [8, 13, 16, 17], "datamodul": 8, "__init__": [8, 13, 14, 16, 17, 18, 23, 25, 28], "add_eo": 8, "add_so": 8, "get_text": 8, "39": [8, 13, 16, 18, 25, 28, 32, 36], "tensorboard": [8, 13, 23, 25, 26, 28, 31, 32, 33, 34, 36, 37, 39, 43, 44, 45], "summarywrit": 8, "miniconda3": 8, "env": 8, "yyi": 8, "lib": [8, 13, 18], "site": [8, 13, 18], "packag": [8, 13, 18], "loosevers": 8, "uninstal": 8, "setuptool": [8, 13], "conda": [8, 13], "dev": [8, 13, 14, 16, 17, 18, 23, 25, 26, 28, 31, 32, 33, 34, 36, 37, 39, 43, 44, 45], "yangyifan": 8, "anaconda3": 8, "dev20230112": 8, "cuda11": [8, 13], "torch1": [8, 13], "13": [8, 13, 14, 16, 17, 18, 25, 26, 28, 32, 33, 36], "py3": [8, 13], "linux": [8, 11, 13, 15, 16, 17, 18, 19], "x86_64": [8, 13, 16], "egg": [8, 13], "_k2": [8, 13], "determinizeweightpushingtyp": 8, "handl": [8, 23, 26, 28, 29, 31, 32, 33, 34, 36, 37, 39, 43, 44, 45], "except": [8, 14], "anoth": 8, "occur": 8, "pruned_transducer_stateless7_ctc_b": [8, 33], "104": 8, "30": [8, 13, 16, 17, 18, 23, 25, 26, 28, 29, 31, 33, 34, 39, 43, 44, 45], "rais": 8, "anaconda": 8, "maco": [8, 11, 15, 16, 17, 18, 19], "probabl": [8, 13, 25, 31, 33, 43, 44, 45], "variabl": [8, 13, 16, 17, 18, 23, 26, 28, 29, 31, 33, 34, 43, 44, 45], "export": [8, 12, 13, 23, 25, 26, 28, 29, 32, 36, 37, 39], "dyld_library_path": 8, "conda_prefix": 8, "find": [8, 9, 10, 11, 13, 14, 16, 17, 18, 21, 23, 25, 26, 28, 31, 32, 33, 34, 36, 37, 39, 43, 44, 45], "locat": [8, 16], "libpython": 8, "abl": 8, "insid": [8, 21], "codna_prefix": 8, "ld_library_path": 8, "within": [9, 11, 16, 17], "anyth": [9, 11], "youtub": [9, 12, 28, 29, 31, 32, 33, 34, 43, 44, 45], "video": [9, 12, 28, 29, 31, 32, 33, 34, 43, 44, 45], "upload": [10, 11, 23, 25, 26, 28, 29, 31, 32, 33, 34, 36, 37, 39, 43, 44, 45], "visit": [10, 11, 31, 33, 34, 43, 44, 45], "specif": [10, 19, 25], "aishel": [10, 12, 23, 25, 26, 27, 46], "wenetspeech": [10, 20], "framework": [11, 31, 44], "sherpa": [11, 15, 20, 21, 22, 43], "window": [11, 15, 16, 17, 18, 19], "ipad": 11, "phone": 11, "start": [11, 13, 14, 18, 23, 25, 26, 28, 31, 32, 33, 34, 36, 37, 39, 43, 44, 45], "recognit": [11, 12, 15, 16, 17, 24, 25, 39, 46], "screenshot": [11, 23, 25, 26, 28, 29, 31, 39, 43, 44], "select": [11, 16, 17, 18, 31, 32, 36, 37, 39, 43, 44, 45], "current": [11, 16, 17, 25, 29, 41, 43, 44, 45, 46], "chines": [11, 24, 25], "english": [11, 39, 43], "greedi": 11, "record": [11, 17, 18, 23, 24, 25, 26, 28, 31, 32, 33, 34, 36, 37, 39, 43, 44, 45], "click": [11, 13, 23, 25, 26, 28, 31, 33, 34, 39, 43, 44], "button": 11, "submit": 11, "wait": 11, "moment": 11, "bottom": [11, 31, 33, 34, 43, 44, 45], "part": [11, 13, 23, 25, 26, 28, 31, 32, 33, 34, 36, 37, 39, 41, 43, 44, 45], "one": [11, 14, 16, 17, 18, 23, 25, 26, 28, 31, 32, 33, 34, 36, 37, 39, 41, 43, 44, 45], "subscrib": [11, 13, 28, 29, 31, 32, 33, 34, 43, 44, 45], "channel": [11, 13, 23, 25, 26, 28, 29, 31, 32, 33, 34, 36, 37, 39, 43, 44, 45], "nadira": [11, 13, 28, 29, 31, 32, 33, 34, 43, 44, 45], "povei": [11, 13, 28, 29, 31, 32, 33, 34, 43, 44, 45], "www": [11, 13, 24, 28, 29, 31, 32, 33, 34, 43, 44, 45], "uc_vaumpkminz1pnkfxan9mw": [11, 13, 28, 29, 31, 32, 33, 34, 43, 44, 45], "toolkit": 12, "cudnn": 12, "frequent": 12, "ask": 12, "question": 12, "faq": 12, "oserror": 12, "libtorch_hip": 12, "cannot": [12, 16, 17, 18], "share": [12, 13], "object": [12, 13, 23, 25, 26, 31, 39, 43, 44], "attributeerror": 12, "modul": [12, 13, 16, 18, 33, 44], "distutil": 12, "attribut": [12, 18, 28], "importerror": 12, "libpython3": 12, "No": [12, 16, 17, 18, 39], "state_dict": [12, 22, 23, 25, 26, 28, 32, 36, 37, 39], "jit": [12, 15, 22, 28], "trace": [12, 15, 20, 22], "onnx": [12, 14, 22], "ncnn": [12, 22], "non": [12, 28, 41, 44, 46], "timit": [12, 27, 36, 37, 46], "introduct": [12, 40, 46], "contribut": 12, "who": 13, "about": [13, 16, 17, 18, 25, 29, 31, 34, 43, 44, 45], "suggest": [13, 31, 33, 34, 43, 44, 45], "virut": 13, "venv": 13, "my_env": 13, "bin": [13, 16, 17, 18, 23, 28], "matter": [13, 16], "compil": [13, 16, 17, 23, 25, 28], "wheel": [13, 16], "don": [13, 16, 17, 18, 20, 23, 26, 28, 31, 32, 33, 34, 36, 37, 39, 43, 44, 45], "from_sourc": 13, "for_develop": 13, "alwai": [13, 14], "strongli": 13, "pythonpath": [13, 16, 17, 18], "point": [13, 14, 23, 26, 28, 29, 31, 33, 34, 43, 44, 45], "folder": [13, 14, 23, 25, 26, 28, 31, 32, 33, 34, 36, 37, 39, 43, 44, 45], "tmp": [13, 23, 25, 26, 28, 29, 31, 32, 33, 34, 36, 37, 39, 43, 44, 45], "setup": [13, 16, 23, 25, 26, 28, 29, 31, 32, 36, 37, 39, 44, 45], "put": [13, 16, 17, 33, 44], "sever": [13, 14, 23, 25, 26, 28, 29, 31, 32, 33, 34, 36, 37, 39, 41, 43, 44, 45], "switch": [13, 23, 28, 34], "just": [13, 16, 17, 18, 41], "virtualenv": 13, "cpython3": 13, "final": [13, 14, 16, 17, 28, 32], "64": [13, 14, 16, 25, 44], "1540m": 13, "creator": 13, "cpython3posix": 13, "dest": 13, "ceph": [13, 14, 23, 25, 28], "fj": [13, 14, 16, 17, 18, 25, 28], "fangjun": [13, 14, 16, 17, 18, 25, 28], "clear": 13, "no_vcs_ignor": 13, "global": 13, "seeder": 13, "fromappdata": 13, "bundl": 13, "copi": [13, 41], "app_data_dir": 13, "root": [13, 16, 17, 18], "v": [13, 16, 17, 18, 28, 36, 37], "irtualenv": 13, "ad": [13, 16, 17, 18, 23, 25, 26, 28, 31, 33, 34, 39, 41, 43, 44, 45], "seed": 13, "21": [13, 14, 16, 23, 25, 28, 36, 37], "36": [13, 16, 25, 28, 29], "bashactiv": 13, "cshellactiv": 13, "fishactiv": 13, "powershellactiv": 13, "pythonactiv": 13, "xonshactiv": 13, "dev20210822": 13, "cpu": [13, 14, 16, 17, 18, 20, 23, 31, 33, 34, 39, 44, 45], "nightli": 13, "2bcpu": 13, "cp38": 13, "linux_x86_64": 13, "mb": [13, 16, 17, 18], "________________________________": 13, "185": [13, 23, 28, 39], "kb": [13, 16, 17, 18, 36, 37], "graphviz": 13, "17": [13, 14, 16, 17, 18, 23, 28, 36, 37, 43], "none": [13, 23, 28], "18": [13, 16, 17, 18, 23, 25, 26, 28, 31, 32, 36, 37, 43, 44, 45], "cach": [13, 18], "manylinux1_x86_64": 13, "831": [13, 25, 37], "extens": 13, "typing_extens": 13, "26": [13, 16, 17, 18, 25, 28, 37], "successfulli": [13, 16, 17, 18], "req": 13, "7b1b76ge": 13, "q": 13, "audioread": 13, "soundfil": 13, "post1": 13, "py2": 13, "97": [13, 16, 23], "cytoolz": 13, "manylinux_2_17_x86_64": 13, "manylinux2014_x86_64": 13, "dataclass": 13, "h5py": 13, "manylinux_2_12_x86_64": 13, "manylinux2010_x86_64": 13, "684": [13, 23, 39], "intervaltre": 13, "lilcom": 13, "numpi": 13, "15": [13, 14, 16, 17, 18, 25, 26, 28, 36, 39], "40": [13, 16, 17, 18, 26, 28, 32, 36, 37], "pyyaml": 13, "662": 13, "tqdm": 13, "76": [13, 39], "satisfi": 13, "2a1410b": 13, "toolz": 13, "55": [13, 16, 26, 28, 36], "sortedcontain": 13, "cffi": 13, "411": [13, 18, 28], "pycpars": 13, "20": [13, 14, 16, 18, 23, 25, 26, 28, 31, 32, 36, 37, 39, 44], "112": [13, 16, 17, 18], "pypars": 13, "filenam": [13, 16, 17, 18, 19, 20, 21, 33, 34, 43, 45], "dev_2a1410b_clean": 13, "size": [13, 14, 16, 17, 18, 23, 25, 26, 28, 29, 31, 32, 33, 34, 36, 37, 39, 43, 44, 45], "342242": 13, "sha256": 13, "f683444afa4dc0881133206b4646a": 13, "9d0f774224cc84000f55d0a67f6e4a37997": 13, "store": [13, 28], "ephem": 13, "ftu0qysz": 13, "7f": 13, "7a": 13, "8e": 13, "a0bf241336e2e3cb573e1e21e5600952d49f5162454f2e612f": 13, "warn": 13, "built": 13, "invalid": [13, 28], "metadata": [13, 36, 37], "mandat": 13, "pep": 13, "440": 13, "packa": 13, "ging": 13, "deprec": [13, 25], "legaci": 13, "becaus": 13, "could": [13, 16, 17, 18, 23, 26], "replac": [13, 16, 17], "discuss": 13, "regard": 13, "pypa": 13, "sue": 13, "8368": 13, "inter": 13, "valtre": 13, "sor": 13, "tedcontain": 13, "remot": 13, "enumer": 13, "count": 13, "100": [13, 23, 25, 26, 28, 29, 31, 33, 34, 43, 44, 45], "compress": 13, "308": [13, 23, 25, 26], "total": [13, 17, 18, 23, 25, 26, 28, 29, 31, 32, 39, 43, 44], "delta": 13, "reus": 13, "307": 13, "102": [13, 18, 23], "pack": [13, 44, 45], "receiv": 13, "172": 13, "49": [13, 16, 17, 28, 37, 39], "kib": 13, "385": 13, "00": [13, 16, 23, 25, 26, 28, 32, 36, 37, 39], "resolv": 13, "kaldilm": 13, "tar": 13, "gz": 13, "48": [13, 16, 17, 23, 25], "574": 13, "kaldialign": 13, "sentencepiec": [13, 28], "41": [13, 16, 18, 23, 25, 36, 39], "absl": 13, "absl_pi": 13, "googl": [13, 31, 33, 34, 43, 44, 45], "auth": 13, "oauthlib": 13, "google_auth_oauthlib": 13, "grpcio": 13, "ment": 13, "requi": 13, "rement": 13, "protobuf": 13, "manylinux_2_5_x86_64": 13, "werkzeug": 13, "288": 13, "tensorboard_data_serv": 13, "google_auth": 13, "35": [13, 14, 16, 17, 18, 25, 28, 43], "152": 13, "request": [13, 41], "plugin": 13, "wit": 13, "tensorboard_plugin_wit": 13, "781": 13, "markdown": 13, "six": 13, "16": [13, 14, 16, 17, 18, 21, 23, 25, 26, 28, 31, 32, 36, 37, 39, 43, 44, 45], "cachetool": 13, "rsa": 13, "pyasn1": 13, "pyasn1_modul": 13, "155": 13, "requests_oauthlib": 13, "urllib3": 13, "27": [13, 16, 17, 18, 23, 25, 32, 37], "138": [13, 23, 25], "certifi": 13, "2017": 13, "2021": [13, 23, 26, 28, 32, 36, 37, 39], "145": 13, "charset": 13, "normal": [13, 32, 36, 37, 39, 44], "charset_norm": 13, "idna": 13, "146": 13, "897233": 13, "eccb906cafcd45bf9a7e1a1718e4534254bfb": 13, "f4c0d0cbc66eee6c88d68a63862": 13, "85": 13, "7d": 13, "f2dd586369b8797cb36d213bf3a84a789eeb92db93d2e723c9": 13, "etool": 13, "oaut": 13, "hlib": 13, "2023": [13, 16, 17, 18, 33], "05": [13, 14, 16, 17, 23, 25, 26, 28, 37], "main": [13, 23, 28, 41], "dl_dir": [13, 23, 26, 28, 29, 31, 33, 34, 43, 44, 45], "waves_yesno": 13, "_______________________________________________________________": 13, "70m": 13, "06": [13, 14, 16, 26, 28, 32, 39], "54": [13, 17, 18, 28, 32, 36, 37], "4kb": 13, "02": [13, 14, 16, 17, 18, 25, 28, 31, 37, 43, 44], "19": [13, 14, 16, 17, 18, 23, 28, 32, 36, 37], "manifest": [13, 29], "fbank": [13, 14, 16, 17, 18, 23, 25, 26, 28, 32, 36, 37, 39], "199": [13, 28, 32], "info": [13, 14, 16, 17, 18, 23, 25, 26, 28, 32, 36, 37, 39], "compute_fbank_yesno": 13, "process": [13, 14, 16, 17, 23, 25, 26, 28, 31, 33, 34, 43, 44, 45], "extract": [13, 23, 25, 26, 28, 29, 31, 32, 33, 34, 36, 37, 39, 43, 44, 45], "featur": [13, 23, 25, 26, 28, 29, 31, 32, 33, 34, 36, 37, 39, 41, 43, 44, 45], "90": [13, 16], "212": 13, "60it": 13, "640": [13, 18], "304": [13, 17], "53it": 13, "51": [13, 16, 23, 28, 39], "lang": [13, 14, 25, 28, 34], "66": [13, 17], "project": 13, "csrc": [13, 28], "arpa_file_pars": 13, "cc": 13, "void": 13, "arpafilepars": 13, "std": 13, "istream": 13, "79": 13, "140": [13, 26], "92": [13, 28], "hlg": [13, 32, 36, 37, 39], "28": [13, 16, 17, 25, 28, 32], "581": [13, 16, 32], "compile_hlg": 13, "124": [13, 23, 28], "lang_phon": [13, 26, 32, 36, 37, 39], "582": 13, "lexicon": [13, 23, 25, 26, 28, 29, 31, 33, 34, 39, 43, 44, 45], "171": [13, 26, 28, 36, 37], "convert": [13, 16, 17, 18, 28], "l": [13, 16, 17, 18, 25, 36, 37, 39], "linv": [13, 25, 28, 39], "609": 13, "ctc_topo": 13, "max_token_id": 13, "611": 13, "intersect": [13, 31, 44, 45], "613": 13, "lg": [13, 31, 34, 44, 45], "shape": [13, 18], "connect": [13, 14, 28, 31, 32, 43, 44, 45], "614": 13, "68": [13, 28], "70": 13, "class": [13, 28], "tensor": [13, 17, 18, 23, 25, 26, 28, 31, 39, 43, 44], "71": [13, 28, 32], "determin": 13, "615": 13, "rag": 13, "raggedtensor": 13, "remov": [13, 23, 25, 26, 28, 32, 36, 37], "disambigu": 13, "616": 13, "91": 13, "remove_epsilon": 13, "617": 13, "arc": 13, "compos": 13, "h": 13, "619": 13, "106": [13, 17, 28], "109": [13, 23, 28], "111": [13, 28], "127": [13, 16, 17, 39], "cuda_visible_devic": [13, 23, 25, 26, 28, 29, 31, 32, 33, 34, 36, 37, 39, 43, 44, 45], "case": [13, 14, 16, 17, 18, 31, 33, 34, 43, 44, 45], "segment": 13, "fault": 13, "core": 13, "dump": 13, "protocol_buffers_python_implement": 13, "674": 13, "interest": [13, 29, 31, 33, 34, 43, 44, 45], "759": [13, 25], "481": 13, "482": 13, "posixpath": [13, 16, 17, 18, 25, 28], "lang_dir": [13, 25, 28], "lr": [13, 25, 43], "01": [13, 16, 25, 26, 28, 29, 33], "feature_dim": [13, 14, 16, 17, 18, 23, 25, 28, 39], "weight_decai": 13, "1e": 13, "start_epoch": 13, "best_train_loss": [13, 14, 16, 17, 18], "inf": [13, 14, 16, 17, 18], "best_valid_loss": [13, 14, 16, 17, 18], "best_train_epoch": [13, 14, 16, 17, 18], "best_valid_epoch": [13, 14, 17, 18], "batch_idx_train": [13, 14, 16, 17, 18], "log_interv": [13, 14, 16, 17, 18], "reset_interv": [13, 14, 16, 17, 18], "valid_interv": [13, 14, 16, 17, 18], "beam_siz": [13, 14, 25], "sum": 13, "use_double_scor": [13, 23, 28, 39], "world_siz": [13, 29], "master_port": 13, "12354": 13, "num_epoch": 13, "feature_dir": [13, 28], "max_dur": [13, 28], "bucketing_sampl": [13, 28], "num_bucket": [13, 28], "concatenate_cut": [13, 28], "duration_factor": [13, 28], "gap": [13, 28], "on_the_fly_feat": [13, 28], "shuffl": [13, 28], "return_cut": [13, 28], "num_work": [13, 28], "env_info": [13, 14, 16, 17, 18, 23, 25, 28], "releas": [13, 14, 16, 17, 18, 23, 25, 28], "sha1": [13, 14, 16, 17, 18, 23, 25, 28], "3b7f09fa35e72589914f67089c0da9f196a92ca4": 13, "date": [13, 14, 16, 17, 18, 23, 25, 28], "mon": [13, 17, 18], "6fcfced": 13, "cu118": 13, "branch": [13, 14, 16, 17, 18, 23, 25, 28, 33], "30bde4b": 13, "thu": [13, 14, 16, 17, 18, 25, 28, 32], "37": [13, 17, 23, 25, 28, 36], "47": [13, 16, 17, 18, 23, 28], "dev20230512": 13, "torch2": 13, "hostnam": [13, 14, 16, 17, 18, 25], "host": [13, 14], "ip": [13, 14, 16, 17, 18, 25], "761": 13, "168": [13, 32], "764": 13, "495": 13, "devic": [13, 14, 16, 17, 18, 23, 25, 26, 28, 31, 32, 33, 34, 36, 37, 39, 44, 45], "791": [13, 32], "cut": [13, 28], "244": 13, "852": 13, "149": [13, 16, 28], "singlecutsampl": 13, "205": [13, 28], "853": 13, "218": [13, 17], "252": 13, "986": 13, "422": 13, "batch": [13, 16, 17, 18, 23, 25, 26, 28, 31, 33, 34, 43, 44, 45], "loss": [13, 16, 17, 23, 26, 28, 29, 31, 32, 33, 34, 36, 37, 39, 43, 44, 45], "065": 13, "over": [13, 23, 25, 26, 28, 31, 33, 34, 43, 44, 45], "2436": 13, "frame": [13, 25, 31, 33, 44, 45], "tot_loss": 13, "4561": 13, "2828": 13, "7076": 13, "22192": 13, "691": 13, "444": 13, "9002": 13, "18067": 13, "996": 13, "2555": 13, "2695": 13, "484": 13, "34971": 13, "217": [13, 23, 28], "4688": 13, "251": [13, 36, 37], "75": [13, 16], "389": [13, 26, 28], "2532": 13, "637": 13, "1139": 13, "1592": 13, "859": 13, "1629": 13, "094": 13, "0767": 13, "118": [13, 28], "350": 13, "06778": 13, "395": 13, "789": 13, "01056": 13, "016": 13, "009022": 13, "009985": 13, "271": [13, 14, 17], "01088": 13, "497": 13, "01174": 13, "01077": 13, "747": 13, "01087": 13, "783": 13, "921": 13, "01045": 13, "008957": 13, "009903": 13, "374": 13, "01092": 13, "598": [13, 28], "01169": 13, "01065": 13, "824": 13, "862": [13, 17], "865": [13, 17], "555": 13, "483": 13, "264": [13, 18], "search_beam": [13, 23, 28, 39], "output_beam": [13, 23, 28, 39], "min_active_st": [13, 23, 28, 39], "max_active_st": [13, 23, 28, 39], "10000": [13, 23, 28, 39], "487": 13, "273": [13, 14, 25], "513": 13, "291": 13, "521": 13, "675": 13, "204": [13, 18, 28], "until": [13, 28, 33], "923": 13, "241": [13, 23], "transcript": [13, 23, 24, 25, 26, 28, 31, 32, 36, 37, 43, 44, 45], "recog": [13, 25, 28], "test_set": [13, 39], "924": 13, "558": 13, "240": [13, 23, 39], "ins": [13, 28, 39], "del": [13, 28, 39], "sub": [13, 28, 39], "925": 13, "249": [13, 17], "wrote": [13, 28], "detail": [13, 15, 19, 23, 25, 26, 28, 29, 31, 32, 33, 34, 36, 37, 39, 41, 43, 44, 45], "stat": [13, 28], "err": [13, 25, 28], "316": [13, 28], "congratul": [13, 16, 17, 18, 23, 26, 28, 32, 36, 37, 39], "fun": [13, 16, 17], "debug": 13, "variou": [13, 19, 22, 46], "period": [14, 16], "disk": 14, "optim": [14, 23, 25, 26, 28, 31, 32, 33, 34, 36, 37, 39, 43, 44, 45], "relat": [14, 23, 25, 28, 32, 36, 37, 39], "resum": [14, 23, 25, 26, 28, 31, 32, 33, 34, 36, 37, 39, 43, 44, 45], "strip": 14, "reduc": [14, 23, 25, 26, 28, 31, 32, 33, 34, 36, 37, 39, 43, 44, 45], "each": [14, 16, 17, 19, 23, 25, 26, 28, 31, 33, 34, 41, 43, 44, 45], "well": [14, 39, 46], "usag": [14, 16, 17, 18, 20, 21, 32, 36, 37, 39], "pruned_transducer_stateless3": [14, 20, 41], "almost": [14, 31, 41, 44, 45], "dict": [14, 18], "csukuangfj": [14, 16, 17, 19, 23, 25, 26, 28, 32, 36, 37, 39, 43], "stateless3": [14, 16], "repo": [14, 19], "prefix": 14, "those": 14, "wave": [14, 16, 17, 18, 23, 28], "iter": [14, 16, 17, 18, 21, 31, 33, 34, 43, 44, 45], "1224000": 14, "greedy_search": [14, 25, 31, 33, 43, 44, 45], "test_wav": [14, 16, 17, 18, 19, 23, 25, 26, 28, 32, 36, 37, 39], "1089": [14, 16, 17, 18, 19, 28, 32], "134686": [14, 16, 17, 18, 19, 28, 32], "0001": [14, 16, 17, 18, 19, 28, 32], "wav": [14, 16, 17, 18, 19, 21, 23, 25, 26, 28, 31, 33, 34, 36, 37, 39, 43, 44, 45], "1221": [14, 16, 17, 28, 32], "135766": [14, 16, 17, 28, 32], "0002": [14, 16, 17, 28, 32], "multipl": [14, 23, 25, 26, 28, 32, 36, 37, 39], "sound": [14, 16, 17, 18, 21, 22, 23, 25, 26, 28, 32, 36, 37, 39], "Its": [14, 16, 17, 18, 28], "output": [14, 16, 17, 18, 23, 25, 26, 28, 29, 31, 32, 33, 34, 36, 37, 39, 41, 43, 44, 45], "09": [14, 17, 23, 25, 26, 28, 43], "233": [14, 16, 17], "265": 14, "50": [14, 16, 17, 18, 28, 31, 36, 43, 44, 45], "200": [14, 16, 17, 18, 23, 28, 29, 36, 37, 39], "3000": [14, 16, 17, 18], "80": [14, 16, 17, 18, 23, 25, 28], "subsampling_factor": [14, 17, 18, 23, 25, 28], "encoder_dim": [14, 16, 17, 18], "512": [14, 16, 17, 18, 23, 25, 28], "nhead": [14, 16, 18, 23, 25, 28, 31, 44], "dim_feedforward": [14, 16, 17, 25], "num_encoder_lay": [14, 16, 17, 18, 25], "decoder_dim": [14, 16, 17, 18], "joiner_dim": [14, 16, 17, 18], "model_warm_step": [14, 16, 17], "4810e00d8738f1a21278b0156a42ff396a2d40ac": 14, "fri": 14, "oct": [14, 28], "03": [14, 17, 25, 28, 36, 37, 43], "miss": [14, 16, 17, 18, 25, 28], "cu102": [14, 16, 17, 18], "1013": 14, "c39cba5": 14, "dirti": [14, 16, 17, 23, 28], "jsonl": 14, "de": [14, 16, 17, 18, 25], "74279": [14, 16, 17, 18, 25], "0324160024": 14, "65bfd8b584": 14, "jjlbn": 14, "bpe_model": [14, 16, 17, 18, 28], "sound_fil": [14, 23, 25, 28, 39], "sample_r": [14, 23, 25, 28, 39], "16000": [14, 23, 25, 26, 28, 32, 33, 36, 37], "max_context": 14, "max_stat": 14, "context_s": [14, 16, 17, 18, 25], "max_sym_per_fram": [14, 25], "simulate_stream": 14, "decode_chunk_s": 14, "left_context": 14, "dynamic_chunk_train": 14, "causal_convolut": 14, "short_chunk_s": [14, 18, 44, 45], "25": [14, 16, 17, 23, 28, 31, 36, 37, 39, 44], "num_left_chunk": [14, 18], "blank_id": [14, 16, 17, 18, 25], "unk_id": 14, "vocab_s": [14, 16, 17, 18, 25], "612": 14, "458": 14, "disabl": [14, 16, 17], "giga": [14, 17, 43], "623": 14, "277": 14, "78648040": 14, "951": [14, 28], "285": [14, 25, 28], "construct": [14, 16, 17, 18, 23, 25, 26, 28, 32, 36, 37, 39], "952": 14, "295": [14, 23, 25, 26, 28], "957": 14, "301": [14, 28], "700": 14, "329": [14, 17, 28], "912": 14, "388": 14, "earli": [14, 16, 17, 18, 28, 32], "nightfal": [14, 16, 17, 18, 28, 32], "THE": [14, 16, 17, 18, 28, 32], "yellow": [14, 16, 17, 18, 28, 32], "lamp": [14, 16, 17, 18, 28, 32], "light": [14, 16, 17, 18, 28, 32], "AND": [14, 16, 17, 18, 28, 32], "THERE": [14, 16, 17, 18, 28, 32], "squalid": [14, 16, 17, 18, 28, 32], "quarter": [14, 16, 17, 18, 28, 32], "OF": [14, 16, 17, 18, 28, 32], "brothel": [14, 16, 17, 18, 28, 32], "god": [14, 28, 32], "AS": [14, 28, 32], "direct": [14, 28, 32], "consequ": [14, 28, 32], "sin": [14, 28, 32], "man": [14, 28, 32], "punish": [14, 28, 32], "had": [14, 28, 32], "her": [14, 28, 32], "love": [14, 28, 32], "child": [14, 28, 32], "whose": [14, 25, 28, 32], "ON": [14, 16, 28, 32], "THAT": [14, 28, 32], "dishonor": [14, 28, 32], "bosom": [14, 28, 32], "TO": [14, 28, 32], "parent": [14, 28, 32], "forev": [14, 28, 32], "WITH": [14, 28, 32], "race": [14, 28, 32], "descent": [14, 28, 32], "mortal": [14, 28, 32], "BE": [14, 28, 32], "bless": [14, 28, 32], "soul": [14, 28, 32], "IN": [14, 28, 32], "heaven": [14, 28, 32], "yet": [14, 16, 17, 28, 32], "THESE": [14, 28, 32], "thought": [14, 28, 32], "affect": [14, 28, 32], "hester": [14, 28, 32], "prynn": [14, 28, 32], "hope": [14, 24, 28, 32], "apprehens": [14, 28, 32], "390": 14, "down": [14, 23, 28, 31, 33, 34, 43, 44, 45], "reproduc": [14, 28], "9999": [14, 33, 34, 43], "symlink": 14, "pass": [14, 18, 23, 25, 26, 28, 31, 33, 34, 41, 43, 44, 45], "reason": [14, 16, 17, 18, 44], "support": [15, 16, 17, 18, 23, 25, 28, 31, 33, 34, 41, 43, 44, 45], "zipform": [15, 19, 22, 27, 30, 40, 42], "convemform": [15, 22, 41], "platform": [15, 19], "android": [15, 16, 17, 18, 19], "raspberri": [15, 19], "pi": [15, 19], "\u7231\u82af\u6d3e": 15, "maix": 15, "iii": 15, "axera": 15, "rv1126": 15, "static": 15, "produc": [15, 31, 33, 34, 43, 44, 45], "binari": [15, 16, 17, 18, 23, 25, 26, 28, 31, 39, 43, 44], "everyth": 15, "pnnx": [15, 22], "torchscript": [15, 20, 21, 22], "encod": [15, 19, 21, 22, 23, 25, 26, 28, 31, 32, 33, 39, 41, 43, 44, 45], "option": [15, 19, 22, 25, 29, 32, 36, 37, 39], "int8": [15, 22], "quantiz": [15, 22, 29], "conv": [16, 17], "emform": [16, 17, 20], "stateless2": [16, 17, 43], "07": [16, 17, 18, 23, 25, 26, 28], "ubuntu": [16, 17, 18], "cpp": [16, 20], "pretrained_model": [16, 17, 18], "online_transduc": 16, "continu": [16, 17, 18, 19, 23, 25, 26, 28, 31, 33, 34, 39, 43, 44], "jit_xxx": [16, 17, 18], "anywher": [16, 17], "submodul": 16, "updat": [16, 17, 18], "recurs": 16, "init": 16, "cmake": [16, 17, 23, 28], "dcmake_build_typ": [16, 23, 28], "dncnn_python": 16, "dncnn_build_benchmark": 16, "dncnn_build_exampl": 16, "dncnn_build_tool": 16, "j4": 16, "pwd": 16, "src": [16, 18], "compon": [16, 41], "ncnn2int8": [16, 17], "our": [16, 17, 18, 20, 21, 28, 29, 31, 41, 44, 45], "cpython": 16, "gnu": 16, "am": 16, "sai": [16, 17, 18, 23, 25, 26, 28, 29, 31, 32, 33, 34, 36, 37, 39, 43, 44, 45], "later": [16, 17, 18, 23, 26, 28, 31, 32, 33, 34, 36, 37, 43, 44, 45], "termin": 16, "tencent": [16, 17], "modif": [16, 25], "offic": 16, "synchron": 16, "offici": 16, "renam": [16, 17, 18], "conv_emformer_transducer_stateless2": [16, 41], "length": [16, 18, 25, 44, 45], "cnn": [16, 18], "kernel": [16, 18, 25], "31": [16, 17, 18, 28], "context": [16, 25, 31, 41, 43, 44, 45], "memori": [16, 23, 25, 28, 41], "configur": [16, 18, 25, 29, 32, 36, 37, 39], "accordingli": [16, 17, 18], "yourself": [16, 17, 18, 29, 44, 45], "combin": [16, 17, 18], "677": 16, "220": [16, 25, 26, 28], "681": 16, "229": [16, 23], "best_v": 16, "alid_epoch": 16, "subsampl": [16, 44, 45], "ing_factor": 16, "a34171ed85605b0926eebbd0463d059431f4f74a": 16, "wed": [16, 23, 25, 28], "dec": 16, "ver": 16, "ion": 16, "530e8a1": 16, "tue": [16, 28], "star": [16, 17, 18], "op": 16, "1220120619": [16, 17, 18], "7695ff496b": [16, 17, 18], "s9n4w": [16, 17, 18], "icefa": 16, "ll": 16, "transdu": 16, "cer": 16, "use_averaged_model": [16, 17, 18], "cnn_module_kernel": [16, 18], "left_context_length": 16, "chunk_length": 16, "right_context_length": 16, "memory_s": 16, "231": [16, 17, 18], "053": 16, "022": 16, "708": [16, 23, 25, 28, 39], "315": [16, 23, 25, 26, 28, 32], "75490012": 16, "318": [16, 17], "320": [16, 25], "682": 16, "lh": [16, 17, 18], "rw": [16, 17, 18], "kuangfangjun": [16, 17, 18], "289m": 16, "jan": [16, 17, 18], "289": 16, "roughli": [16, 17, 18], "equal": [16, 17, 18, 44, 45], "1024": [16, 17, 18, 43], "287": [16, 39], "1010k": [16, 17], "decoder_jit_trac": [16, 17, 18, 21, 43, 45], "283m": 16, "encoder_jit_trac": [16, 17, 18, 21, 43, 45], "0m": [16, 17], "joiner_jit_trac": [16, 17, 18, 21, 43, 45], "sure": [16, 17, 18], "found": [16, 17, 18, 23, 25, 26, 28, 31, 33, 34, 39, 43, 44], "param": [16, 17, 18], "503k": [16, 17], "437": [16, 17, 18], "142m": 16, "79k": 16, "5m": [16, 17], "architectur": [16, 17, 18, 43], "editor": [16, 17, 18], "content": [16, 17, 18], "283": [16, 18], "1010": [16, 17], "142": [16, 23, 26, 28], "503": [16, 17], "convers": [16, 17, 18], "half": [16, 17, 18, 31, 44, 45], "default": [16, 17, 18, 23, 25, 26, 28, 29, 31, 32, 33, 34, 36, 37, 39, 43, 44, 45], "float32": [16, 17, 18], "float16": [16, 17, 18], "occupi": [16, 17, 18], "byte": [16, 17, 18], "twice": [16, 17, 18], "smaller": [16, 17, 18, 23, 25, 26, 28, 31, 33, 34, 43, 44, 45], "fp16": [16, 17, 18, 31, 33, 34, 43, 44, 45], "won": [16, 17, 18, 19, 23, 26, 28, 29, 31, 33, 34, 43, 44, 45], "accept": [16, 17, 18], "216": [16, 23, 28, 36, 37], "encoder_param_filenam": [16, 17, 18], "encoder_bin_filenam": [16, 17, 18], "decoder_param_filenam": [16, 17, 18], "decoder_bin_filenam": [16, 17, 18], "joiner_param_filenam": [16, 17, 18], "joiner_bin_filenam": [16, 17, 18], "sound_filenam": [16, 17, 18], "141": 16, "328": 16, "151": 16, "331": [16, 17, 28, 32], "176": [16, 25, 28], "336": 16, "106000": [16, 17, 18, 28, 32], "381": 16, "7767517": [16, 17, 18], "1060": 16, "1342": 16, "in0": [16, 17, 18], "explan": [16, 17, 18], "magic": [16, 17, 18], "intermedi": [16, 17, 18], "mean": [16, 17, 18, 23, 25, 26, 28, 31, 32, 33, 34, 36, 37, 39, 41, 43, 44, 45], "increment": [16, 17, 18], "1061": 16, "sherpametadata": [16, 17, 18], "sherpa_meta_data1": [16, 17, 18], "still": [16, 17, 18], "newli": [16, 17, 18], "must": [16, 17, 18, 44], "kei": [16, 17, 18, 28], "eas": [16, 17, 18], "list": [16, 17, 18, 23, 25, 26, 28, 32, 36, 37], "pair": [16, 17, 18], "sad": [16, 17, 18], "rememb": [16, 17, 18], "anymor": [16, 17, 18], "flexibl": [16, 17, 18], "edit": [16, 17, 18], "arm": [16, 17, 18], "aarch64": [16, 17, 18], "onc": [16, 17], "mayb": [16, 17], "year": [16, 17], "_jit_trac": [16, 17], "56": [16, 17, 28, 36], "fp32": [16, 17], "doubl": [16, 17], "j": [16, 17, 23, 28], "py38": [16, 17, 18], "arg": [16, 17], "wave_filenam": [16, 17], "16k": [16, 17], "hz": [16, 17, 36, 37], "mono": [16, 17], "calibr": [16, 17], "cat": [16, 17], "eof": [16, 17], "calcul": [16, 17, 33, 44, 45], "has_gpu": [16, 17], "config": [16, 17], "use_vulkan_comput": [16, 17], "88": [16, 25], "conv_87": 16, "942385": [16, 17], "threshold": [16, 17, 33], "938493": 16, "968131": 16, "conv_88": 16, "442448": 16, "549335": 16, "167552": 16, "conv_89": 16, "228289": 16, "001738": 16, "871552": 16, "linear_90": 16, "976146": 16, "101789": 16, "115": [16, 17, 23, 28], "267128": 16, "linear_91": 16, "962030": 16, "162033": 16, "602713": 16, "linear_92": 16, "323041": 16, "853959": 16, "953129": 16, "linear_94": 16, "905416": 16, "648006": 16, "323545": 16, "linear_93": 16, "474093": 16, "200188": 16, "linear_95": 16, "888012": 16, "403563": 16, "483986": 16, "linear_96": 16, "856741": 16, "398679": 16, "524273": 16, "linear_97": 16, "635942": 16, "613655": 16, "590950": 16, "linear_98": 16, "460340": 16, "670146": 16, "398010": 16, "linear_99": 16, "532276": 16, "585537": 16, "119396": 16, "linear_101": 16, "585871": 16, "719224": 16, "205809": 16, "linear_100": 16, "751382": 16, "081648": 16, "linear_102": 16, "593344": 16, "450581": 16, "87": 16, "551147": 16, "linear_103": 16, "592681": 16, "705824": 16, "257959": 16, "linear_104": 16, "752957": 16, "980955": 16, "110489": 16, "linear_105": 16, "696240": 16, "877193": 16, "608953": 16, "linear_106": 16, "059659": 16, "643138": 16, "048950": 16, "linear_108": 16, "975461": 16, "589567": 16, "671457": 16, "linear_107": 16, "190381": 16, "515701": 16, "linear_109": 16, "710759": 16, "305635": 16, "082436": 16, "linear_110": 16, "531228": 16, "731162": 16, "159557": 16, "linear_111": 16, "528083": 16, "259322": 16, "211544": 16, "linear_112": 16, "148807": 16, "500842": 16, "087374": 16, "linear_113": 16, "592566": 16, "948851": 16, "166611": 16, "linear_115": 16, "437109": 16, "608947": 16, "642395": 16, "linear_114": 16, "193942": 16, "503904": 16, "linear_116": 16, "966980": 16, "200896": 16, "676392": 16, "linear_117": 16, "451303": 16, "061664": 16, "951344": 16, "linear_118": 16, "077262": 16, "965800": 16, "023804": 16, "linear_119": 16, "671615": 16, "847613": 16, "198460": 16, "linear_120": 16, "625638": 16, "131427": 16, "556595": 16, "linear_122": 16, "274080": 16, "888716": 16, "978189": 16, "linear_121": 16, "420480": 16, "429659": 16, "linear_123": 16, "826197": 16, "599617": 16, "281532": 16, "linear_124": 16, "396383": 16, "325849": 16, "335875": 16, "linear_125": 16, "337198": 16, "941410": 16, "221970": 16, "linear_126": 16, "699965": 16, "842878": 16, "224073": 16, "linear_127": 16, "775370": 16, "884215": 16, "696438": 16, "linear_129": 16, "872276": 16, "837319": 16, "254213": 16, "linear_128": 16, "180057": 16, "687883": 16, "linear_130": 16, "150427": 16, "454298": 16, "765789": 16, "linear_131": 16, "112692": 16, "924847": 16, "025545": 16, "linear_132": 16, "852893": 16, "116593": 16, "749626": 16, "linear_133": 16, "517084": 16, "024665": 16, "275314": 16, "linear_134": 16, "683807": 16, "878618": 16, "743618": 16, "linear_136": 16, "421055": 16, "322729": 16, "086264": 16, "linear_135": 16, "309880": 16, "917679": 16, "linear_137": 16, "827781": 16, "744595": 16, "33": [16, 17, 23, 24, 25, 28, 36], "915554": 16, "linear_138": 16, "422395": 16, "742882": 16, "402161": 16, "linear_139": 16, "527538": 16, "866123": 16, "849449": 16, "linear_140": 16, "128619": 16, "657793": 16, "266134": 16, "linear_141": 16, "839593": 16, "845993": 16, "021378": 16, "linear_143": 16, "442304": 16, "099039": 16, "889746": 16, "linear_142": 16, "325038": 16, "849592": 16, "linear_144": 16, "929444": 16, "618206": 16, "605080": 16, "linear_145": 16, "382126": 16, "321095": 16, "625010": 16, "linear_146": 16, "894987": 16, "867645": 16, "836517": 16, "linear_147": 16, "915313": 16, "906028": 16, "886522": 16, "linear_148": 16, "614287": 16, "908151": 16, "496181": 16, "linear_150": 16, "724932": 16, "485588": 16, "312899": 16, "linear_149": 16, "161146": 16, "606939": 16, "linear_151": 16, "164453": 16, "847355": 16, "719223": 16, "linear_152": 16, "086471": 16, "984121": 16, "222834": 16, "linear_153": 16, "099524": 16, "991601": 16, "816805": 16, "linear_154": 16, "054585": 16, "489706": 16, "286930": 16, "linear_155": 16, "389185": 16, "100321": 16, "963501": 16, "linear_157": 16, "982999": 16, "154796": 16, "637253": 16, "linear_156": 16, "537706": 16, "875190": 16, "linear_158": 16, "420287": 16, "502287": 16, "531588": 16, "linear_159": 16, "014746": 16, "423280": 16, "477261": 16, "linear_160": 16, "633553": 16, "715335": 16, "220921": 16, "linear_161": 16, "371849": 16, "117830": 16, "815203": 16, "linear_162": 16, "492933": 16, "126283": 16, "623318": 16, "linear_164": 16, "697504": 16, "825712": 16, "317358": 16, "linear_163": 16, "078367": 16, "008038": 16, "linear_165": 16, "023975": 16, "836278": 16, "577358": 16, "linear_166": 16, "860619": 16, "259792": 16, "493614": 16, "linear_167": 16, "380934": 16, "496160": 16, "107042": 16, "linear_168": 16, "691216": 16, "733317": 16, "831076": 16, "linear_169": 16, "723948": 16, "952728": 16, "129707": 16, "linear_171": 16, "034811": 16, "366547": 16, "665123": 16, "linear_170": 16, "356277": 16, "710501": 16, "linear_172": 16, "556884": 16, "729481": 16, "166058": 16, "linear_173": 16, "033039": 16, "207264": 16, "442120": 16, "linear_174": 16, "597379": 16, "658676": 16, "768131": 16, "linear_2": [16, 17], "293503": 16, "305265": 16, "877850": 16, "linear_1": [16, 17], "812222": 16, "766452": 16, "487047": 16, "linear_3": [16, 17], "999999": 16, "999755": 16, "031174": 16, "wish": [16, 17], "955k": 16, "18k": 16, "inparam": [16, 17], "inbin": [16, 17], "outparam": [16, 17], "outbin": [16, 17], "99m": 16, "78k": 16, "774k": [16, 17], "496": [16, 17, 28, 32], "774": [16, 17], "linear": [16, 17, 25], "convolut": [16, 17, 33, 41, 44], "exact": [16, 17], "4x": [16, 17], "comparison": 16, "44": [16, 17, 28, 36, 37], "468000": [17, 21, 43], "lstm_transducer_stateless2": [17, 21, 43], "222": [17, 26, 28], "is_pnnx": 17, "62e404dd3f3a811d73e424199b3408e309c06e1a": [17, 18], "6d7a559": [17, 18], "feb": [17, 18, 25], "147": [17, 18], "rnn_hidden_s": 17, "aux_layer_period": 17, "235": 17, "239": [17, 25], "472": 17, "595": 17, "324": 17, "83137520": 17, "596": 17, "325": 17, "257024": 17, "326": 17, "781812": 17, "327": 17, "84176356": 17, "182": [17, 18, 23, 32], "158": 17, "183": [17, 36, 37], "335": 17, "101": 17, "tracerwarn": [17, 18], "boolean": [17, 18], "caus": [17, 18, 23, 25, 26, 28, 31, 33, 34, 43, 44, 45], "incorrect": [17, 18, 25], "flow": [17, 18], "constant": [17, 18], "futur": [17, 18, 25, 46], "need_pad": 17, "bool": 17, "259": [17, 23], "180": [17, 23, 28], "339": 17, "207": [17, 26, 28], "84": [17, 23], "324m": 17, "321": [17, 23], "107": [17, 32], "318m": 17, "159m": 17, "21k": 17, "159": [17, 28, 39], "861": 17, "425": [17, 28], "427": [17, 28], "266": [17, 18, 28, 32], "431": 17, "342": 17, "343": 17, "267": [17, 25, 36, 37], "379": 17, "268": [17, 28, 32], "317m": 17, "317": 17, "conv_15": 17, "930708": 17, "972025": 17, "conv_16": 17, "978855": 17, "031788": 17, "456645": 17, "conv_17": 17, "868437": 17, "830528": 17, "218575": 17, "linear_18": 17, "107259": 17, "194808": 17, "293236": 17, "linear_19": 17, "193777": 17, "634748": 17, "401705": 17, "linear_20": 17, "259933": 17, "606617": 17, "722160": 17, "linear_21": 17, "186600": 17, "790260": 17, "512129": 17, "linear_22": 17, "759041": 17, "265832": 17, "050053": 17, "linear_23": 17, "931209": 17, "099090": 17, "979767": 17, "linear_24": 17, "324160": 17, "215561": 17, "321835": 17, "linear_25": 17, "800708": 17, "599352": 17, "284134": 17, "linear_26": 17, "492444": 17, "153369": 17, "274391": 17, "linear_27": 17, "660161": 17, "720994": 17, "46": [17, 23, 28], "674126": 17, "linear_28": 17, "415265": 17, "174434": 17, "007133": 17, "linear_29": 17, "038418": 17, "118534": 17, "724262": 17, "linear_30": 17, "072084": 17, "936867": 17, "259155": 17, "linear_31": 17, "342712": 17, "599489": 17, "282787": 17, "linear_32": 17, "340535": 17, "120308": 17, "701103": 17, "linear_33": 17, "846987": 17, "630030": 17, "985939": 17, "linear_34": 17, "686298": 17, "204571": 17, "607586": 17, "linear_35": 17, "904821": 17, "575518": 17, "756420": 17, "linear_36": 17, "806659": 17, "585589": 17, "118401": 17, "linear_37": 17, "402340": 17, "047157": 17, "162680": 17, "linear_38": 17, "174589": 17, "923361": 17, "030258": 17, "linear_39": 17, "178576": 17, "556058": 17, "807705": 17, "linear_40": 17, "901954": 17, "301267": 17, "956539": 17, "linear_41": 17, "839805": 17, "597429": 17, "716181": 17, "linear_42": 17, "178945": 17, "651595": 17, "895699": 17, "829245": 17, "627592": 17, "637907": 17, "746186": 17, "255032": 17, "167313": 17, "000000": 17, "999756": 17, "031013": 17, "345k": 17, "17k": 17, "218m": 17, "counterpart": 17, "bit": [17, 23, 25, 26, 28, 32, 39], "4532": 17, "feedforward": [18, 25, 31, 44], "384": [18, 28], "192": [18, 28], "unmask": 18, "256": [18, 36, 37], "downsampl": [18, 24], "factor": [18, 23, 25, 26, 28, 29, 31, 33, 34, 43, 44, 45], "473": [18, 28], "246": [18, 25, 28, 36, 37], "477": 18, "warm_step": 18, "2000": [18, 26], "feedforward_dim": 18, "attention_dim": [18, 23, 25, 28], "encoder_unmasked_dim": 18, "zipformer_downsampling_factor": 18, "decode_chunk_len": 18, "257": [18, 25, 36, 37], "023": 18, "zipformer2": 18, "419": 18, "At": [18, 23, 28], "stack": 18, "downsampling_factor": 18, "037": 18, "655": 18, "346": 18, "68944004": 18, "347": 18, "260096": 18, "348": [18, 36], "716276": 18, "656": [18, 28], "349": 18, "69920376": 18, "351": 18, "353": 18, "174": [18, 28], "175": 18, "1344": 18, "assert": 18, "cached_len": 18, "num_lay": 18, "1348": 18, "cached_avg": 18, "1352": 18, "cached_kei": 18, "1356": 18, "cached_v": 18, "1360": 18, "cached_val2": 18, "1364": 18, "cached_conv1": 18, "1368": 18, "cached_conv2": 18, "1373": 18, "left_context_len": 18, "1884": 18, "x_size": 18, "2442": 18, "2449": 18, "2469": 18, "2473": 18, "2483": 18, "kv_len": 18, "k": [18, 31, 36, 37, 43, 44, 45], "2570": 18, "attn_output": 18, "bsz": 18, "num_head": 18, "seq_len": 18, "head_dim": 18, "2926": 18, "lorder": 18, "2652": 18, "2653": 18, "embed_dim": 18, "2666": 18, "1543": 18, "in_x_siz": 18, "1637": 18, "1643": 18, "in_channel": 18, "1571": 18, "1763": 18, "src1": 18, "src2": 18, "1779": 18, "dim1": 18, "1780": 18, "dim2": 18, "_trace": 18, "958": 18, "tracer": 18, "instead": [18, 25, 44], "tupl": 18, "namedtupl": 18, "absolut": 18, "know": [18, 29], "side": 18, "strict": [18, 24], "allow": [18, 31, 44], "behavior": [18, 25], "_c": 18, "_create_method_from_trac": 18, "646": 18, "357": 18, "embedding_out": 18, "686": 18, "361": [18, 28, 32], "735": 18, "69": 18, "269m": 18, "53": [18, 23, 31, 32, 37, 43, 44], "269": [18, 23, 36, 37], "725": [18, 32], "1022k": 18, "266m": 18, "8m": 18, "509k": 18, "133m": 18, "152k": 18, "4m": 18, "1022": 18, "133": 18, "509": 18, "260": [18, 28], "360": 18, "365": 18, "280": [18, 28], "372": [18, 23], "state": [18, 23, 25, 26, 28, 31, 33, 34, 43, 44, 45], "026": 18, "410": 18, "2028": 18, "2547": 18, "2029": 18, "23316": 18, "23317": 18, "23318": 18, "23319": 18, "23320": 18, "amount": [18, 24], "pad": [18, 23, 25, 26, 28, 31, 33, 34, 43, 44, 45], "conv2dsubsampl": 18, "v2": [18, 23, 28], "arrai": 18, "23300": 18, "element": 18, "onnx_pretrain": 19, "onnxruntim": 19, "separ": 19, "deploi": [19, 23, 28], "repo_url": 19, "basenam": 19, "tree": [20, 21, 23, 25, 26, 28, 32, 36, 37, 39, 43], "cpu_jit": [20, 23, 28, 31, 33, 34, 44, 45], "confus": 20, "move": [20, 31, 33, 34, 44, 45], "why": 20, "streaming_asr": [20, 21, 43, 44, 45], "conv_emform": 20, "offline_asr": [20, 31], "jit_pretrain": [21, 33, 34, 43], "baz": 21, "1best": [23, 26, 28, 32, 33, 34, 36, 37], "automag": [23, 26, 28, 29, 31, 32, 33, 34, 36, 37, 39, 43, 44, 45], "stop": [23, 25, 26, 28, 29, 31, 32, 33, 34, 36, 37, 39, 43, 44, 45], "By": [23, 26, 28, 29, 31, 32, 33, 34, 36, 37, 39, 43, 44, 45], "musan": [23, 26, 28, 29, 31, 33, 34, 43, 44, 45], "thei": [23, 25, 26, 28, 29, 31, 33, 34, 43, 44, 45], "intal": [23, 26], "sudo": [23, 26], "apt": [23, 26], "permiss": [23, 26], "commandlin": [23, 25, 26, 28, 31, 33, 34, 43, 44, 45], "quit": [23, 25, 26, 28, 31, 33, 34, 43, 44, 45], "experi": [23, 25, 26, 28, 29, 31, 33, 34, 39, 43, 44, 45], "world": [23, 25, 26, 28, 29, 31, 32, 33, 34, 43, 44, 45], "multi": [23, 25, 26, 28, 29, 31, 33, 34, 41, 43, 44, 45], "machin": [23, 25, 26, 28, 31, 33, 34, 43, 44, 45], "ddp": [23, 25, 26, 28, 31, 33, 34, 43, 44, 45], "implement": [23, 25, 26, 28, 29, 31, 33, 34, 41, 43, 44, 45], "present": [23, 25, 26, 28, 31, 33, 34, 43, 44, 45], "second": [23, 25, 26, 28, 29, 31, 33, 34, 39, 43, 44, 45], "utter": [23, 25, 26, 28, 31, 33, 34, 43, 44, 45], "oom": [23, 25, 26, 28, 31, 33, 34, 43, 44, 45], "nvidia": [23, 25, 26, 28], "due": [23, 25, 26, 28, 31, 33, 34, 43, 44, 45], "decai": [23, 26, 28, 33, 34, 43], "warmup": [23, 25, 26, 28, 31, 33, 34, 43, 44, 45], "function": [23, 25, 26, 28, 31, 32, 33, 34, 36, 37, 39, 43, 44, 45], "get_param": [23, 25, 26, 28, 31, 32, 33, 34, 36, 37, 39, 43, 44, 45], "directli": [23, 25, 26, 28, 29, 31, 33, 34, 43, 44, 45], "perturb": [23, 25, 26, 28, 31, 33, 34, 43, 44, 45], "actual": [23, 25, 26, 28, 31, 33, 34, 43, 44, 45], "3x150": [23, 25, 26], "450": [23, 25, 26], "visual": [23, 25, 26, 28, 31, 32, 33, 34, 36, 37, 39, 43, 44, 45], "logdir": [23, 25, 26, 28, 31, 32, 33, 34, 36, 37, 39, 43, 44, 45], "labelsmooth": 23, "someth": [23, 25, 26, 28, 31, 33, 34, 39, 43, 44], "tensorflow": [23, 25, 26, 28, 31, 33, 34, 39, 43, 44], "press": [23, 25, 26, 28, 31, 33, 34, 39, 43, 44, 45], "ctrl": [23, 25, 26, 28, 31, 33, 34, 39, 43, 44, 45], "engw8ksktzqs24zbv5dgcg": 23, "22t11": 23, "scan": [23, 25, 26, 28, 31, 39, 43, 44], "116068": 23, "scalar": [23, 25, 26, 28, 31, 39, 43, 44], "listen": [23, 25, 26, 31, 39, 43, 44], "url": [23, 25, 26, 28, 31, 33, 34, 39, 43, 44], "xxxx": [23, 25, 26, 28, 31, 32, 33, 34, 36, 37, 39, 43, 44, 45], "saw": [23, 25, 26, 28, 31, 32, 33, 34, 36, 37, 39, 43, 44, 45], "consol": [23, 25, 26, 28, 31, 32, 33, 34, 36, 37, 39, 43, 44, 45], "avoid": [23, 25, 28], "nbest": [23, 28, 34], "lattic": [23, 26, 28, 31, 32, 36, 37, 44, 45], "uniqu": [23, 28, 31, 44, 45], "pkufool": [23, 26, 32], "icefall_asr_aishell_conformer_ctc": 23, "transcrib": [23, 25, 26, 28], "v1": [23, 26, 28, 32, 36, 37], "lang_char": [23, 25], "bac009s0764w0121": [23, 25, 26], "bac009s0764w0122": [23, 25, 26], "bac009s0764w0123": [23, 25, 26], "tran": [23, 26, 28, 32, 36, 37], "graph": [23, 26, 28, 31, 32, 36, 37, 44, 45], "id": [23, 26, 28, 32, 36, 37], "conveni": [23, 26, 28, 29], "eo": [23, 26, 28], "soxi": [23, 25, 26, 28, 32, 39], "sampl": [23, 25, 26, 28, 32, 33, 39, 44, 45], "precis": [23, 25, 26, 28, 31, 32, 39, 44, 45], "67263": [23, 25, 26], "cdda": [23, 25, 26, 28, 32, 39], "sector": [23, 25, 26, 28, 32, 39], "135k": [23, 25, 26], "256k": [23, 25, 26, 28], "sign": [23, 25, 26, 28, 39], "integ": [23, 25, 26, 28, 39], "pcm": [23, 25, 26, 28, 39], "65840": [23, 25, 26], "625": [23, 25, 26], "132k": [23, 25, 26], "64000": [23, 25, 26], "300": [23, 25, 26, 28, 29, 31, 44], "128k": [23, 25, 26, 39], "displai": [23, 25, 26, 28], "topologi": [23, 28], "707": [23, 28], "num_decoder_lay": [23, 28], "vgg_frontend": [23, 25, 28], "use_feat_batchnorm": [23, 28], "f2fd997f752ed11bbef4c306652c433e83f9cf12": 23, "sun": 23, "sep": 23, "33cfe45": 23, "d57a873": 23, "nov": [23, 28], "hw": 23, "kangwei": 23, "icefall_aishell3": 23, "k2_releas": 23, "tokens_fil": 23, "words_fil": [23, 28, 39], "num_path": [23, 28, 31, 44, 45], "ngram_lm_scal": [23, 28], "attention_decoder_scal": [23, 28], "nbest_scal": [23, 28], "sos_id": [23, 28], "eos_id": [23, 28], "num_class": [23, 28, 39], "4336": [23, 25], "242": [23, 28], "131": [23, 28], "134": 23, "275": 23, "293": [23, 28], "704": [23, 36], "369": [23, 28], "\u751a": [23, 25], "\u81f3": [23, 25], "\u51fa": [23, 25], "\u73b0": [23, 25], "\u4ea4": [23, 25], "\u6613": [23, 25], "\u51e0": [23, 25], "\u4e4e": [23, 25], "\u505c": [23, 25], "\u6b62": 23, "\u7684": [23, 25, 26], "\u60c5": [23, 25], "\u51b5": [23, 25], "\u4e00": [23, 25], "\u4e8c": [23, 25], "\u7ebf": [23, 25, 26], "\u57ce": [23, 25], "\u5e02": [23, 25], "\u867d": [23, 25], "\u7136": [23, 25], "\u4e5f": [23, 25, 26], "\u5904": [23, 25], "\u4e8e": [23, 25], "\u8c03": [23, 25], "\u6574": [23, 25], "\u4e2d": [23, 25, 26], "\u4f46": [23, 25, 26], "\u56e0": [23, 25], "\u4e3a": [23, 25], "\u805a": [23, 25], "\u96c6": [23, 25], "\u4e86": [23, 25, 26], "\u8fc7": [23, 25], "\u591a": [23, 25], "\u516c": [23, 25], "\u5171": [23, 25], "\u8d44": [23, 25], "\u6e90": [23, 25], "371": 23, "683": 23, "651": [23, 39], "654": 23, "659": 23, "752": 23, "887": 23, "340": 23, "370": 23, "\u751a\u81f3": [23, 26], "\u51fa\u73b0": [23, 26], "\u4ea4\u6613": [23, 26], "\u51e0\u4e4e": [23, 26], "\u505c\u6b62": 23, "\u60c5\u51b5": [23, 26], "\u4e00\u4e8c": [23, 26], "\u57ce\u5e02": [23, 26], "\u867d\u7136": [23, 26], "\u5904\u4e8e": [23, 26], "\u8c03\u6574": [23, 26], "\u56e0\u4e3a": [23, 26], "\u805a\u96c6": [23, 26], "\u8fc7\u591a": [23, 26], "\u516c\u5171": [23, 26], "\u8d44\u6e90": [23, 26], "recor": [23, 28], "highest": [23, 28], "965": 23, "966": 23, "821": 23, "822": 23, "826": 23, "916": 23, "345": 23, "888": 23, "889": 23, "limit": [23, 25, 28, 41, 44], "upgrad": [23, 28], "pro": [23, 28], "finish": [23, 25, 26, 28, 29, 31, 32, 36, 37, 39, 44, 45], "NOT": [23, 25, 28, 39], "checkout": [23, 28], "hlg_decod": [23, 28], "four": [23, 28], "messag": [23, 28, 31, 33, 34, 43, 44, 45], "nn_model": [23, 28], "use_gpu": [23, 28], "word_tabl": [23, 28], "caution": [23, 28], "forward": [23, 28, 33], "89": 23, "cu": [23, 28], "int": [23, 28], "char": [23, 28], "98": 23, "150": [23, 28], "693": [23, 36], "165": [23, 28], "nnet_output": [23, 28], "489": 23, "mandarin": 24, "beij": 24, "shell": 24, "technologi": 24, "ltd": 24, "peopl": 24, "accent": 24, "area": 24, "china": 24, "invit": 24, "particip": 24, "conduct": 24, "quiet": 24, "indoor": 24, "high": 24, "fidel": 24, "microphon": 24, "16khz": 24, "manual": 24, "through": 24, "profession": 24, "annot": 24, "inspect": 24, "free": [24, 29, 43], "academ": 24, "moder": 24, "research": 24, "field": 24, "openslr": 24, "ctc": [24, 27, 30, 34, 35, 38], "stateless": [24, 27, 31, 43, 44, 45], "conv1d": [25, 31, 43, 44, 45], "nn": [25, 31, 33, 34, 43, 44, 45], "tanh": 25, "borrow": 25, "ieeexplor": 25, "ieee": 25, "stamp": 25, "jsp": 25, "arnumb": 25, "9054419": 25, "predict": [25, 29, 31, 43, 44, 45], "charact": 25, "unit": 25, "vocabulari": 25, "87939824": 25, "optimized_transduc": 25, "technqiu": 25, "end": [25, 31, 33, 34, 39, 43, 44, 45], "furthermor": 25, "maximum": 25, "emit": 25, "per": [25, 31, 44, 45], "simplifi": [25, 41], "significantli": 25, "degrad": 25, "exactli": 25, "unprun": 25, "advantag": 25, "minim": 25, "pruned_transducer_stateless": [25, 31, 41, 44], "altern": 25, "though": 25, "transducer_stateless_modifi": 25, "pr": 25, "gb": 25, "ram": 25, "tri": 25, "prob": [25, 43], "219": [25, 28], "c": [25, 26, 31, 33, 34, 39, 43, 44, 45], "lagz6hrcqxoigbfd5e0y3q": 25, "03t14": 25, "8477": 25, "250": [25, 32], "sym": [25, 31, 44, 45], "beam_search": [25, 31, 44, 45], "decoding_method": 25, "beam_4": 25, "ensur": 25, "give": 25, "poor": 25, "531": [25, 26], "994": [25, 28], "027": 25, "encoder_out_dim": 25, "f4fefe4882bc0ae59af951da3f47335d5495ef71": 25, "50d2281": 25, "mar": 25, "0815224919": 25, "75d558775b": 25, "mmnv8": 25, "72": [25, 28], "248": 25, "878": [25, 37], "880": 25, "891": 25, "113": [25, 28], "userwarn": 25, "__floordiv__": 25, "round": 25, "toward": 25, "trunc": 25, "floor": 25, "keep": [25, 31, 44, 45], "div": 25, "b": [25, 28, 36, 37], "rounding_mod": 25, "divis": 25, "x_len": 25, "163": [25, 28], "\u6ede": 25, "322": 25, "760": 25, "919": 25, "922": 25, "929": 25, "046": 25, "047": 25, "319": [25, 28], "798": 25, "214": [25, 28], "215": [25, 28, 32], "402": 25, "topk_hyp_index": 25, "topk_index": 25, "logit": 25, "583": [25, 37], "lji9mwuorlow3jkdhxwk8a": 26, "13t11": 26, "4454": 26, "icefall_asr_aishell_tdnn_lstm_ctc": 26, "858": [26, 28], "154": 26, "161": [26, 28], "536": 26, "539": 26, "917": 26, "129": 26, "\u505c\u6ede": 26, "mmi": [27, 30], "blank": [27, 30], "skip": [27, 29, 30, 31, 43, 44, 45], "distil": [27, 30], "hubert": [27, 30], "ligru": [27, 35], "full": [28, 29, 31, 33, 34, 43, 44, 45], "libri": [28, 29, 31, 33, 34, 43, 44, 45], "subset": [28, 31, 33, 34, 43, 44, 45], "3x960": [28, 31, 33, 34, 43, 44, 45], "2880": [28, 31, 33, 34, 43, 44, 45], "lzgnetjwrxc3yghnmd4kpw": 28, "24t16": 28, "4540": 28, "sentenc": 28, "piec": 28, "And": [28, 31, 33, 34, 43, 44, 45], "neither": 28, "nor": 28, "5000": 28, "033": 28, "537": 28, "538": 28, "full_libri": [28, 29], "406": 28, "464": 28, "548": 28, "776": 28, "652": [28, 39], "109226120": 28, "714": [28, 36], "206": 28, "944": 28, "1328": 28, "443": [28, 32], "2563": 28, "494": 28, "592": 28, "1715": 28, "52576": 28, "128": 28, "1424": 28, "807": 28, "506": 28, "808": [28, 36], "522": 28, "362": 28, "565": 28, "1477": 28, "2922": 28, "208": 28, "4295": 28, "52343": 28, "396": 28, "3584": 28, "432": 28, "433": 28, "680": [28, 36], "_pickl": 28, "unpicklingerror": 28, "hlg_modifi": 28, "g_4_gram": [28, 32, 36, 37], "875": [28, 32], "212k": 28, "267440": [28, 32], "1253": [28, 32], "535k": 28, "83": [28, 32], "77200": [28, 32], "154k": 28, "554": 28, "7178d67e594bc7fa89c2b331ad7bd1c62a6a9eb4": 28, "8d93169": 28, "601": 28, "758": 28, "025": 28, "broffel": 28, "osom": 28, "723": 28, "775": 28, "881": 28, "234": 28, "571": 28, "whole": [28, 32, 36, 37, 44, 45], "857": 28, "979": 28, "980": 28, "055": 28, "117": 28, "051": 28, "363": 28, "959": [28, 37], "546": 28, "599": [28, 32], "833": 28, "834": 28, "915": 28, "076": 28, "110": 28, "397": 28, "999": [28, 31, 44, 45], "concaten": 28, "bucket": 28, "sampler": 28, "1000": 28, "ctc_decod": 28, "ngram_lm_rescor": 28, "attention_rescor": 28, "kind": [28, 31, 33, 34, 43, 44, 45], "105": 28, "221": 28, "125": [28, 39], "136": 28, "228": 28, "144": 28, "543": 28, "topo": 28, "547": 28, "729": 28, "702": 28, "703": 28, "545": 28, "279": 28, "122": 28, "126": 28, "135": [28, 39], "153": [28, 39], "945": 28, "475": 28, "191": [28, 36, 37], "398": 28, "515": 28, "w": [28, 36, 37], "deseri": 28, "441": 28, "fsaclass": 28, "loadfsa": 28, "const": 28, "string": 28, "c10": 28, "ignor": 28, "dummi": 28, "589": 28, "attention_scal": 28, "162": 28, "169": [28, 36, 37], "188": 28, "984": 28, "624": 28, "519": [28, 37], "632": 28, "645": [28, 39], "243": 28, "970": 28, "303": 28, "179": 28, "knowledg": 29, "vector": 29, "mvq": 29, "kd": 29, "pruned_transducer_stateless4": [29, 31, 41, 44], "theoret": 29, "applic": 29, "minor": 29, "out": 29, "thing": 29, "distillation_with_hubert": 29, "Of": 29, "cours": 29, "xl": 29, "proce": 29, "960h": [29, 33], "use_extracted_codebook": 29, "augment": 29, "th": [29, 36, 37], "fine": 29, "embedding_lay": 29, "num_codebook": 29, "under": 29, "vq_fbank_layer36_cb8": 29, "whola": 29, "snippet": 29, "echo": 29, "awk": 29, "split": 29, "_": 29, "pruned_transducer_stateless6": 29, "12359": 29, "spec": 29, "aug": 29, "warp": 29, "enabl": 29, "paid": 29, "suitabl": [31, 43, 44, 45], "pruned_transducer_stateless2": [31, 41, 44], "pruned_transducer_stateless5": [31, 41, 44], "scroll": [31, 33, 34, 43, 44, 45], "arxiv": [31, 43, 44, 45], "ab": [31, 43, 44, 45], "2206": [31, 43, 44, 45], "13236": [31, 43, 44, 45], "rework": [31, 41, 44], "daniel": [31, 44, 45], "joint": [31, 43, 44, 45], "contrari": [31, 43, 44, 45], "convent": [31, 43, 44, 45], "recurr": [31, 43, 44, 45], "2x": [31, 44, 45], "littl": [31, 44], "436000": [31, 33, 34, 43, 44, 45], "438000": [31, 33, 34, 43, 44, 45], "qogspbgsr8kzcrmmie9jgw": 31, "20t15": [31, 43, 44], "4468": [31, 43, 44], "210171": [31, 43, 44], "access": [31, 33, 34, 43, 44, 45], "6008": [31, 33, 34, 43, 44, 45], "localhost": [31, 33, 34, 43, 44, 45], "expos": [31, 33, 34, 43, 44, 45], "proxi": [31, 33, 34, 43, 44, 45], "bind_al": [31, 33, 34, 43, 44, 45], "fast_beam_search": [31, 33, 43, 44, 45], "474000": [31, 43, 44, 45], "largest": [31, 44, 45], "posterior": [31, 33, 44, 45], "algorithm": [31, 44, 45], "pdf": [31, 34, 44, 45], "1211": [31, 44, 45], "3711": [31, 44, 45], "espnet": [31, 44, 45], "net": [31, 44, 45], "beam_search_transduc": [31, 44, 45], "basicli": [31, 44, 45], "topk": [31, 44, 45], "expand": [31, 44, 45], "mode": [31, 44, 45], "being": [31, 44, 45], "hardcod": [31, 44, 45], "composit": [31, 44, 45], "log_prob": [31, 44, 45], "hard": [31, 41, 44, 45], "2211": [31, 44, 45], "00484": [31, 44, 45], "fast_beam_search_lg": [31, 44, 45], "trivial": [31, 44, 45], "fast_beam_search_nbest": [31, 44, 45], "random_path": [31, 44, 45], "shortest": [31, 44, 45], "fast_beam_search_nbest_lg": [31, 44, 45], "logic": [31, 44, 45], "smallest": [31, 43, 44, 45], "icefall_asr_librispeech_tdnn": 32, "lstm_ctc": 32, "flac": 32, "116k": 32, "140k": 32, "343k": 32, "164k": 32, "105k": 32, "174k": 32, "pretraind": 32, "170": 32, "584": [32, 37], "209": 32, "245": 32, "098": 32, "099": 32, "methond": [32, 36, 37], "403": 32, "631": 32, "190": 32, "121": 32, "010": 32, "guidanc": 33, "bigger": 33, "simpli": 33, "discard": 33, "prevent": 33, "lconv": 33, "encourag": [33, 34, 43], "stabil": [33, 34], "doesn": 33, "warm": [33, 34], "xyozukpeqm62hbilud4upa": [33, 34], "ctc_guide_decode_b": 33, "pretrained_ctc": 33, "jit_pretrained_ctc": 33, "100h": 33, "yfyeung": 33, "wechat": 34, "zipformer_mmi": 34, "worker": [34, 43], "hp": 34, "tdnn_ligru_ctc": 36, "enough": [36, 37, 39], "luomingshuang": [36, 37], "icefall_asr_timit_tdnn_ligru_ctc": 36, "pretrained_average_9_25": 36, "fdhc0_si1559": [36, 37], "felc0_si756": [36, 37], "fmgd0_si1564": [36, 37], "ffprobe": [36, 37], "show_format": [36, 37], "nistspher": [36, 37], "database_id": [36, 37], "database_vers": [36, 37], "utterance_id": [36, 37], "dhc0_si1559": [36, 37], "sample_min": [36, 37], "4176": [36, 37], "sample_max": [36, 37], "5984": [36, 37], "bitrat": [36, 37], "258": [36, 37], "audio": [36, 37], "pcm_s16le": [36, 37], "s16": [36, 37], "elc0_si756": [36, 37], "1546": [36, 37], "1989": [36, 37], "mgd0_si1564": [36, 37], "7626": [36, 37], "10573": [36, 37], "660": 36, "695": 36, "697": 36, "819": 36, "829": 36, "sil": [36, 37], "dh": [36, 37], "ih": [36, 37], "uw": [36, 37], "ah": [36, 37], "ii": [36, 37], "z": [36, 37], "aa": [36, 37], "ei": [36, 37], "dx": [36, 37], "d": [36, 37], "uh": [36, 37], "ng": [36, 37], "eh": [36, 37], "jh": [36, 37], "er": [36, 37], "ai": [36, 37], "hh": [36, 37], "aw": 36, "ae": [36, 37], "705": 36, "715": 36, "720": 36, "ch": 36, "icefall_asr_timit_tdnn_lstm_ctc": 37, "pretrained_average_16_25": 37, "816": 37, "827": 37, "387": 37, "unk": 37, "739": 37, "971": 37, "977": 37, "978": 37, "981": 37, "ow": 37, "ykubhb5wrmosxykid1z9eg": 39, "23t23": 39, "icefall_asr_yesno_tdnn": 39, "l_disambig": 39, "lexicon_disambig": 39, "0_0_0_1_0_0_0_1": 39, "0_0_1_0_0_0_1_0": 39, "0_0_1_0_0_1_1_1": 39, "0_0_1_0_1_0_0_1": 39, "0_0_1_1_0_0_0_1": 39, "0_0_1_1_0_1_1_0": 39, "0_0_1_1_1_0_0_0": 39, "0_0_1_1_1_1_0_0": 39, "0_1_0_0_0_1_0_0": 39, "0_1_0_0_1_0_1_0": 39, "0_1_0_1_0_0_0_0": 39, "0_1_0_1_1_1_0_0": 39, "0_1_1_0_0_1_1_1": 39, "0_1_1_1_0_0_1_0": 39, "0_1_1_1_1_0_1_0": 39, "1_0_0_0_0_0_0_0": 39, "1_0_0_0_0_0_1_1": 39, "1_0_0_1_0_1_1_1": 39, "1_0_1_1_0_1_1_1": 39, "1_0_1_1_1_1_0_1": 39, "1_1_0_0_0_1_1_1": 39, "1_1_0_0_1_0_1_1": 39, "1_1_0_1_0_1_0_0": 39, "1_1_0_1_1_0_0_1": 39, "1_1_0_1_1_1_1_0": 39, "1_1_1_0_0_1_0_1": 39, "1_1_1_0_1_0_1_0": 39, "1_1_1_1_0_0_1_0": 39, "1_1_1_1_1_0_0_0": 39, "1_1_1_1_1_1_1_1": 39, "54080": 39, "507": 39, "108k": 39, "ye": 39, "hebrew": 39, "NO": 39, "621": 39, "119": 39, "650": 39, "139": 39, "143": 39, "198": 39, "181": 39, "186": 39, "187": 39, "213": 39, "correctli": 39, "simplest": 39, "former": 41, "idea": 41, "mask": [41, 44, 45], "wenet": 41, "did": 41, "metion": 41, "complic": 41, "techniqu": 41, "bank": 41, "memor": 41, "histori": 41, "introduc": 41, "variant": 41, "pruned_stateless_emformer_rnnt2": 41, "conv_emformer_transducer_stateless": 41, "ourself": 41, "mechan": 41, "onlin": 43, "lstm_transducer_stateless": 43, "lower": 43, "prepare_giga_speech": 43, "cj2vtpiwqhkn9q1tx6ptpg": 43, "dynam": [44, 45], "causal": 44, "short": [44, 45], "2012": 44, "05481": 44, "flag": 44, "indic": [44, 45], "whether": 44, "sequenc": [44, 45], "uniformli": [44, 45], "seen": [44, 45], "97vkxf80ru61cnp2alwzzg": 44, "streaming_decod": [44, 45], "wise": [44, 45], "parallel": [44, 45], "bath": [44, 45], "parallelli": [44, 45], "seem": 44, "benefit": 44, "mdoel": 44, "320m": 45, "550": 45, "scriptmodul": 45, "jit_trace_export": 45, "jit_trace_pretrain": 45, "task": 46}, "objects": {}, "objtypes": {}, "objnames": {}, "titleterms": {"follow": 0, "code": 0, "style": 0, "contribut": [1, 3], "document": 1, "how": [2, 14, 20, 21], "creat": [2, 13], "recip": [2, 46], "data": [2, 13, 23, 25, 26, 28, 29, 31, 32, 33, 34, 36, 37, 39, 43, 44, 45], "prepar": [2, 13, 23, 25, 26, 28, 29, 31, 32, 33, 34, 36, 37, 39, 43, 44, 45], "train": [2, 10, 13, 16, 17, 18, 19, 23, 25, 26, 28, 29, 31, 32, 33, 34, 36, 37, 39, 43, 44, 45], "decod": [2, 5, 6, 7, 13, 14, 19, 23, 25, 26, 28, 29, 31, 32, 33, 34, 36, 37, 39, 43, 44, 45], "pre": [2, 10, 16, 17, 18, 19, 23, 25, 26, 28, 31, 32, 33, 34, 36, 37, 39, 43, 44, 45], "model": [2, 5, 10, 14, 16, 17, 18, 19, 20, 21, 22, 23, 25, 26, 28, 31, 32, 33, 34, 36, 37, 39, 43, 44, 45], "lodr": [4, 6], "rnn": 4, "transduc": [4, 6, 7, 16, 17, 18, 25, 31, 43, 44, 45], "wer": [4, 5, 6, 7, 28], "differ": [4, 6, 7], "beam": [4, 6, 7, 25], "size": [4, 6, 7], "languag": 5, "lm": [5, 6, 28], "rescor": [5, 6, 23, 28], "base": [5, 6], "method": [5, 6], "v": [5, 6], "shallow": [5, 6, 7], "fusion": [5, 6, 7], "The": [5, 6, 25], "number": [5, 6], "each": [5, 6], "field": [5, 6], "i": [5, 6], "test": [5, 6, 7, 13, 16, 17, 18], "clean": [5, 6, 7], "other": [5, 6], "time": [5, 6, 7], "frequent": 8, "ask": 8, "question": 8, "faq": 8, "oserror": 8, "libtorch_hip": 8, "so": 8, "cannot": 8, "open": 8, "share": 8, "object": 8, "file": [8, 19], "directori": 8, "attributeerror": 8, "modul": 8, "distutil": 8, "ha": 8, "attribut": 8, "version": 8, "importerror": 8, "libpython3": 8, "10": 8, "1": [8, 13, 16, 17, 18, 23, 25, 26, 28], "0": [8, 13], "No": 8, "huggingfac": [9, 11], "space": 11, "youtub": [11, 13], "video": [11, 13], "icefal": [12, 13, 16, 17, 18], "content": [12, 46], "instal": [13, 16, 17, 18, 23, 25, 26, 28, 32, 36, 37], "cuda": 13, "toolkit": 13, "cudnn": 13, "pytorch": 13, "torchaudio": 13, "2": [13, 16, 17, 18, 23, 25, 26, 28], "k2": 13, "3": [13, 16, 17, 18, 23, 25, 28], "lhots": 13, "4": [13, 16, 17, 18], "download": [13, 16, 17, 18, 19, 23, 25, 26, 28, 31, 32, 33, 34, 36, 37, 39, 43, 44, 45], "exampl": [13, 19, 23, 25, 26, 28, 31, 33, 34, 43, 44, 45], "virtual": 13, "environ": 13, "activ": 13, "your": 13, "5": [13, 16, 17, 18], "export": [14, 15, 16, 17, 18, 19, 20, 21, 22, 31, 33, 34, 43, 44, 45], "state_dict": [14, 31, 33, 34, 43, 44, 45], "when": [14, 20, 21], "us": [14, 20, 21, 31, 33, 34, 43, 44, 45], "run": 14, "py": 14, "ncnn": [15, 16, 17, 18], "convemform": 16, "pnnx": [16, 17, 18], "via": [16, 17, 18], "torch": [16, 17, 18, 20, 21, 31, 33, 34, 43, 44, 45], "jit": [16, 17, 18, 20, 21, 31, 33, 34, 43, 44, 45], "trace": [16, 17, 18, 21, 43, 45], "torchscript": [16, 17, 18], "6": [16, 17, 18], "modifi": [16, 17, 18, 25], "encod": [16, 17, 18], "sherpa": [16, 17, 18, 19, 31, 44, 45], "7": [16, 17], "option": [16, 17, 23, 26, 28, 31, 33, 34, 43, 44, 45], "int8": [16, 17], "quantiz": [16, 17], "lstm": [17, 26, 32, 37, 43], "stream": [18, 27, 40, 41, 44, 45], "zipform": [18, 33, 34, 45], "onnx": 19, "sound": 19, "script": [20, 31, 33, 34, 44, 45], "conform": [23, 28, 41], "ctc": [23, 26, 28, 32, 33, 36, 37, 39], "configur": [23, 26, 28, 31, 33, 34, 43, 44, 45], "log": [23, 25, 26, 28, 31, 33, 34, 43, 44, 45], "usag": [23, 25, 26, 28, 31, 33, 34, 43, 44, 45], "case": [23, 25, 26, 28], "kaldifeat": [23, 25, 26, 28, 32, 36, 37, 39], "hlg": [23, 26, 28], "attent": [23, 28], "colab": [23, 25, 26, 28, 32, 36, 37, 39], "notebook": [23, 25, 26, 28, 32, 36, 37, 39], "deploy": [23, 28], "c": [23, 28], "aishel": 24, "stateless": 25, "loss": 25, "todo": 25, "greedi": 25, "search": 25, "tdnn": [26, 32, 36, 37, 39], "non": 27, "asr": [27, 40], "comput": 28, "n": 28, "gram": 28, "distil": 29, "hubert": 29, "codebook": 29, "index": 29, "librispeech": [30, 42], "prune": [31, 44], "statelessx": [31, 44], "pretrain": [31, 33, 34, 43, 44, 45], "deploi": [31, 44, 45], "infer": [32, 36, 37, 39], "blank": 33, "skip": 33, "mmi": 34, "timit": 35, "ligru": 36, "yesno": 38, "introduct": 41, "emform": 41, "which": 43, "simul": [44, 45], "real": [44, 45], "tabl": 46}, "envversion": {"sphinx.domains.c": 2, "sphinx.domains.changeset": 1, "sphinx.domains.citation": 1, "sphinx.domains.cpp": 8, "sphinx.domains.index": 1, "sphinx.domains.javascript": 2, "sphinx.domains.math": 2, "sphinx.domains.python": 3, "sphinx.domains.rst": 2, "sphinx.domains.std": 2, "sphinx.ext.todo": 2, "sphinx": 57}, "alltitles": {"Follow the code style": [[0, "follow-the-code-style"]], "Contributing to Documentation": [[1, "contributing-to-documentation"]], "How to create a recipe": [[2, "how-to-create-a-recipe"]], "Data Preparation": [[2, "data-preparation"], [25, "data-preparation"]], "Training": [[2, "training"], [13, "training"], [23, "training"], [25, "training"], [26, "training"], [28, "training"], [29, "training"], [31, "training"], [32, "training"], [33, "training"], [34, "training"], [36, "training"], [37, "training"], [39, "training"], [43, "training"], [44, "training"], [45, "training"]], "Decoding": [[2, "decoding"], [13, "decoding"], [23, "decoding"], [25, "decoding"], [26, "decoding"], [28, "decoding"], [29, "decoding"], [31, "decoding"], [32, "decoding"], [33, "decoding"], [34, "decoding"], [36, "decoding"], [37, "decoding"], [39, "decoding"], [43, "decoding"], [44, "decoding"], [45, "decoding"]], "Pre-trained model": [[2, "pre-trained-model"]], "Contributing": [[3, "contributing"]], "LODR for RNN Transducer": [[4, "lodr-for-rnn-transducer"]], "WER of LODR with different beam sizes": [[4, "id1"]], "Decoding with language models": [[5, "decoding-with-language-models"]], "LM-rescoring-based methods vs shallow-fusion-based methods (The numbers in each field is WER on test-clean, WER on test-other and decoding time on test-clean)": [[5, "id1"], [6, "id3"]], "LM rescoring for Transducer": [[6, "lm-rescoring-for-transducer"]], "WERs of LM rescoring with different beam sizes": [[6, "id1"]], "WERs of LM rescoring + LODR with different beam sizes": [[6, "id2"]], "Shallow fusion for Transducer": [[7, "shallow-fusion-for-transducer"]], "WERs and decoding time (on test-clean) of shallow fusion with different beam sizes": [[7, "id2"]], "Frequently Asked Questions (FAQs)": [[8, "frequently-asked-questions-faqs"]], "OSError: libtorch_hip.so: cannot open shared object file: no such file or directory": [[8, "oserror-libtorch-hip-so-cannot-open-shared-object-file-no-such-file-or-directory"]], "AttributeError: module \u2018distutils\u2019 has no attribute \u2018version\u2019": [[8, "attributeerror-module-distutils-has-no-attribute-version"]], "ImportError: libpython3.10.so.1.0: cannot open shared object file: No such file or directory": [[8, "importerror-libpython3-10-so-1-0-cannot-open-shared-object-file-no-such-file-or-directory"]], "Huggingface": [[9, "huggingface"]], "Pre-trained models": [[10, "pre-trained-models"]], "Huggingface spaces": [[11, "huggingface-spaces"]], "YouTube Video": [[11, "youtube-video"], [13, "youtube-video"]], "Icefall": [[12, "icefall"]], "Contents:": [[12, null]], "Installation": [[13, "installation"]], "(0) Install CUDA toolkit and cuDNN": [[13, "install-cuda-toolkit-and-cudnn"]], "(1) Install PyTorch and torchaudio": [[13, "install-pytorch-and-torchaudio"]], "(2) Install k2": [[13, "install-k2"]], "(3) Install lhotse": [[13, "install-lhotse"]], "(4) Download icefall": [[13, "download-icefall"]], "Installation example": [[13, "installation-example"]], "(1) Create a virtual environment": [[13, "create-a-virtual-environment"]], "(2) Activate your virtual environment": [[13, "activate-your-virtual-environment"]], "(3) Install k2": [[13, "id1"]], "(4) Install lhotse": [[13, "id2"]], "(5) Download icefall": [[13, "id3"]], "Test Your Installation": [[13, "test-your-installation"]], "Data preparation": [[13, "data-preparation"], [23, "data-preparation"], [26, "data-preparation"], [28, "data-preparation"], [29, "data-preparation"], [31, "data-preparation"], [32, "data-preparation"], [33, "data-preparation"], [34, "data-preparation"], [36, "data-preparation"], [37, "data-preparation"], [39, "data-preparation"], [43, "data-preparation"], [44, "data-preparation"], [45, "data-preparation"]], "Export model.state_dict()": [[14, "export-model-state-dict"], [31, "export-model-state-dict"], [33, "export-model-state-dict"], [34, "export-model-state-dict"], [43, "export-model-state-dict"], [44, "export-model-state-dict"], [45, "export-model-state-dict"]], "When to use it": [[14, "when-to-use-it"], [20, "when-to-use-it"], [21, "when-to-use-it"]], "How to export": [[14, "how-to-export"], [20, "how-to-export"], [21, "how-to-export"]], "How to use the exported model": [[14, "how-to-use-the-exported-model"], [20, "how-to-use-the-exported-model"]], "Use the exported model to run decode.py": [[14, "use-the-exported-model-to-run-decode-py"]], "Export to ncnn": [[15, "export-to-ncnn"]], "Export ConvEmformer transducer models to ncnn": [[16, "export-convemformer-transducer-models-to-ncnn"]], "1. Download the pre-trained model": [[16, "download-the-pre-trained-model"], [17, "download-the-pre-trained-model"], [18, "download-the-pre-trained-model"]], "2. Install ncnn and pnnx": [[16, "install-ncnn-and-pnnx"], [17, "install-ncnn-and-pnnx"], [18, "install-ncnn-and-pnnx"]], "3. Export the model via torch.jit.trace()": [[16, "export-the-model-via-torch-jit-trace"], [17, "export-the-model-via-torch-jit-trace"], [18, "export-the-model-via-torch-jit-trace"]], "4. Export torchscript model via pnnx": [[16, "export-torchscript-model-via-pnnx"], [17, "export-torchscript-model-via-pnnx"], [18, "export-torchscript-model-via-pnnx"]], "5. Test the exported models in icefall": [[16, "test-the-exported-models-in-icefall"], [17, "test-the-exported-models-in-icefall"], [18, "test-the-exported-models-in-icefall"]], "6. Modify the exported encoder for sherpa-ncnn": [[16, "modify-the-exported-encoder-for-sherpa-ncnn"], [17, "modify-the-exported-encoder-for-sherpa-ncnn"], [18, "modify-the-exported-encoder-for-sherpa-ncnn"]], "7. (Optional) int8 quantization with sherpa-ncnn": [[16, "optional-int8-quantization-with-sherpa-ncnn"], [17, "optional-int8-quantization-with-sherpa-ncnn"]], "Export LSTM transducer models to ncnn": [[17, "export-lstm-transducer-models-to-ncnn"]], "Export streaming Zipformer transducer models to ncnn": [[18, "export-streaming-zipformer-transducer-models-to-ncnn"]], "Export to ONNX": [[19, "export-to-onnx"]], "sherpa-onnx": [[19, "sherpa-onnx"]], "Example": [[19, "example"]], "Download the pre-trained model": [[19, "download-the-pre-trained-model"], [23, "download-the-pre-trained-model"], [25, "download-the-pre-trained-model"], [26, "download-the-pre-trained-model"], [28, "download-the-pre-trained-model"], [32, "download-the-pre-trained-model"], [36, "download-the-pre-trained-model"], [37, "download-the-pre-trained-model"], [39, "download-the-pre-trained-model"]], "Export the model to ONNX": [[19, "export-the-model-to-onnx"]], "Decode sound files with exported ONNX models": [[19, "decode-sound-files-with-exported-onnx-models"]], "Export model with torch.jit.script()": [[20, "export-model-with-torch-jit-script"]], "Export model with torch.jit.trace()": [[21, "export-model-with-torch-jit-trace"]], "How to use the exported models": [[21, "how-to-use-the-exported-models"]], "Model export": [[22, "model-export"]], "Conformer CTC": [[23, "conformer-ctc"], [28, "conformer-ctc"]], "Configurable options": [[23, "configurable-options"], [26, "configurable-options"], [28, "configurable-options"], [31, "configurable-options"], [33, "configurable-options"], [34, "configurable-options"], [43, "configurable-options"], [44, "configurable-options"], [45, "configurable-options"]], "Pre-configured options": [[23, "pre-configured-options"], [26, "pre-configured-options"], [28, "pre-configured-options"], [31, "pre-configured-options"], [33, "pre-configured-options"], [34, "pre-configured-options"], [43, "pre-configured-options"], [44, "pre-configured-options"], [45, "pre-configured-options"]], "Training logs": [[23, "training-logs"], [25, "training-logs"], [26, "training-logs"], [28, "training-logs"], [31, "training-logs"], [33, "training-logs"], [34, "training-logs"], [43, "training-logs"], [44, "training-logs"], [45, "training-logs"]], "Usage examples": [[23, "usage-examples"], [25, "usage-examples"], [26, "usage-examples"], [28, "usage-examples"]], "Case 1": [[23, "case-1"], [25, "case-1"], [26, "case-1"], [28, "case-1"]], "Case 2": [[23, "case-2"], [25, "case-2"], [26, "case-2"], [28, "case-2"]], "Case 3": [[23, "case-3"], [25, "case-3"], [28, "case-3"]], "Pre-trained Model": [[23, "pre-trained-model"], [25, "pre-trained-model"], [26, "pre-trained-model"], [28, "pre-trained-model"], [32, "pre-trained-model"], [36, "pre-trained-model"], [37, "pre-trained-model"], [39, "pre-trained-model"]], "Install kaldifeat": [[23, "install-kaldifeat"], [25, "install-kaldifeat"], [26, "install-kaldifeat"], [28, "install-kaldifeat"], [32, "install-kaldifeat"], [36, "install-kaldifeat"], [37, "install-kaldifeat"]], "Usage": [[23, "usage"], [25, "usage"], [26, "usage"], [28, "usage"]], "CTC decoding": [[23, "ctc-decoding"], [28, "ctc-decoding"], [28, "id2"]], "HLG decoding": [[23, "hlg-decoding"], [23, "id2"], [26, "hlg-decoding"], [28, "hlg-decoding"], [28, "id3"]], "HLG decoding + attention decoder rescoring": [[23, "hlg-decoding-attention-decoder-rescoring"]], "Colab notebook": [[23, "colab-notebook"], [25, "colab-notebook"], [26, "colab-notebook"], [28, "colab-notebook"], [32, "colab-notebook"], [36, "colab-notebook"], [37, "colab-notebook"], [39, "colab-notebook"]], "Deployment with C++": [[23, "deployment-with-c"], [28, "deployment-with-c"]], "aishell": [[24, "aishell"]], "Stateless Transducer": [[25, "stateless-transducer"]], "The Model": [[25, "the-model"]], "The Loss": [[25, "the-loss"]], "Todo": [[25, "id1"]], "Greedy search": [[25, "greedy-search"]], "Beam search": [[25, "beam-search"]], "Modified Beam search": [[25, "modified-beam-search"]], "TDNN-LSTM CTC": [[26, "tdnn-lstm-ctc"]], "Non Streaming ASR": [[27, "non-streaming-asr"]], "HLG decoding + LM rescoring": [[28, "hlg-decoding-lm-rescoring"]], "HLG decoding + LM rescoring + attention decoder rescoring": [[28, "hlg-decoding-lm-rescoring-attention-decoder-rescoring"]], "Compute WER with the pre-trained model": [[28, "compute-wer-with-the-pre-trained-model"]], "HLG decoding + n-gram LM rescoring": [[28, "hlg-decoding-n-gram-lm-rescoring"]], "HLG decoding + n-gram LM rescoring + attention decoder rescoring": [[28, "hlg-decoding-n-gram-lm-rescoring-attention-decoder-rescoring"]], "Distillation with HuBERT": [[29, "distillation-with-hubert"]], "Codebook index preparation": [[29, "codebook-index-preparation"]], "LibriSpeech": [[30, "librispeech"], [42, "librispeech"]], "Pruned transducer statelessX": [[31, "pruned-transducer-statelessx"], [44, "pruned-transducer-statelessx"]], "Usage example": [[31, "usage-example"], [33, "usage-example"], [34, "usage-example"], [43, "usage-example"], [44, "usage-example"], [45, "usage-example"]], "Export Model": [[31, "export-model"], [44, "export-model"], [45, "export-model"]], "Export model using torch.jit.script()": [[31, "export-model-using-torch-jit-script"], [33, "export-model-using-torch-jit-script"], [34, "export-model-using-torch-jit-script"], [44, "export-model-using-torch-jit-script"], [45, "export-model-using-torch-jit-script"]], "Download pretrained models": [[31, "download-pretrained-models"], [33, "download-pretrained-models"], [34, "download-pretrained-models"], [43, "download-pretrained-models"], [44, "download-pretrained-models"], [45, "download-pretrained-models"]], "Deploy with Sherpa": [[31, "deploy-with-sherpa"], [44, "deploy-with-sherpa"], [45, "deploy-with-sherpa"]], "TDNN-LSTM-CTC": [[32, "tdnn-lstm-ctc"], [37, "tdnn-lstm-ctc"]], "Inference with a pre-trained model": [[32, "inference-with-a-pre-trained-model"], [36, "inference-with-a-pre-trained-model"], [37, "inference-with-a-pre-trained-model"], [39, "inference-with-a-pre-trained-model"]], "Zipformer CTC Blank Skip": [[33, "zipformer-ctc-blank-skip"]], "Export models": [[33, "export-models"], [34, "export-models"], [43, "export-models"]], "Zipformer MMI": [[34, "zipformer-mmi"]], "TIMIT": [[35, "timit"]], "TDNN-LiGRU-CTC": [[36, "tdnn-ligru-ctc"]], "YesNo": [[38, "yesno"]], "TDNN-CTC": [[39, "tdnn-ctc"]], "Download kaldifeat": [[39, "download-kaldifeat"]], "Streaming ASR": [[40, "streaming-asr"]], "Introduction": [[41, "introduction"]], "Streaming Conformer": [[41, "streaming-conformer"]], "Streaming Emformer": [[41, "streaming-emformer"]], "LSTM Transducer": [[43, "lstm-transducer"]], "Which model to use": [[43, "which-model-to-use"]], "Export model using torch.jit.trace()": [[43, "export-model-using-torch-jit-trace"], [45, "export-model-using-torch-jit-trace"]], "Simulate streaming decoding": [[44, "simulate-streaming-decoding"], [45, "simulate-streaming-decoding"]], "Real streaming decoding": [[44, "real-streaming-decoding"], [45, "real-streaming-decoding"]], "Zipformer Transducer": [[45, "zipformer-transducer"]], "Recipes": [[46, "recipes"]], "Table of Contents": [[46, null]]}, "indexentries": {}}) \ No newline at end of file +Search.setIndex({"docnames": ["contributing/code-style", "contributing/doc", "contributing/how-to-create-a-recipe", "contributing/index", "decoding-with-langugage-models/LODR", "decoding-with-langugage-models/index", "decoding-with-langugage-models/rescoring", "decoding-with-langugage-models/shallow-fusion", "faqs", "huggingface/index", "huggingface/pretrained-models", "huggingface/spaces", "index", "installation/index", "model-export/export-model-state-dict", "model-export/export-ncnn", "model-export/export-ncnn-conv-emformer", "model-export/export-ncnn-lstm", "model-export/export-ncnn-zipformer", "model-export/export-onnx", "model-export/export-with-torch-jit-script", "model-export/export-with-torch-jit-trace", "model-export/index", "recipes/Non-streaming-ASR/aishell/conformer_ctc", "recipes/Non-streaming-ASR/aishell/index", "recipes/Non-streaming-ASR/aishell/stateless_transducer", "recipes/Non-streaming-ASR/aishell/tdnn_lstm_ctc", "recipes/Non-streaming-ASR/index", "recipes/Non-streaming-ASR/librispeech/conformer_ctc", "recipes/Non-streaming-ASR/librispeech/distillation", "recipes/Non-streaming-ASR/librispeech/index", "recipes/Non-streaming-ASR/librispeech/pruned_transducer_stateless", "recipes/Non-streaming-ASR/librispeech/tdnn_lstm_ctc", "recipes/Non-streaming-ASR/librispeech/zipformer_ctc_blankskip", "recipes/Non-streaming-ASR/librispeech/zipformer_mmi", "recipes/Non-streaming-ASR/timit/index", "recipes/Non-streaming-ASR/timit/tdnn_ligru_ctc", "recipes/Non-streaming-ASR/timit/tdnn_lstm_ctc", "recipes/Non-streaming-ASR/yesno/index", "recipes/Non-streaming-ASR/yesno/tdnn", "recipes/Streaming-ASR/index", "recipes/Streaming-ASR/introduction", "recipes/Streaming-ASR/librispeech/index", "recipes/Streaming-ASR/librispeech/lstm_pruned_stateless_transducer", "recipes/Streaming-ASR/librispeech/pruned_transducer_stateless", "recipes/Streaming-ASR/librispeech/zipformer_transducer", "recipes/index"], "filenames": ["contributing/code-style.rst", "contributing/doc.rst", "contributing/how-to-create-a-recipe.rst", "contributing/index.rst", "decoding-with-langugage-models/LODR.rst", "decoding-with-langugage-models/index.rst", "decoding-with-langugage-models/rescoring.rst", "decoding-with-langugage-models/shallow-fusion.rst", "faqs.rst", "huggingface/index.rst", "huggingface/pretrained-models.rst", "huggingface/spaces.rst", "index.rst", "installation/index.rst", "model-export/export-model-state-dict.rst", "model-export/export-ncnn.rst", "model-export/export-ncnn-conv-emformer.rst", "model-export/export-ncnn-lstm.rst", "model-export/export-ncnn-zipformer.rst", "model-export/export-onnx.rst", "model-export/export-with-torch-jit-script.rst", "model-export/export-with-torch-jit-trace.rst", "model-export/index.rst", "recipes/Non-streaming-ASR/aishell/conformer_ctc.rst", "recipes/Non-streaming-ASR/aishell/index.rst", "recipes/Non-streaming-ASR/aishell/stateless_transducer.rst", "recipes/Non-streaming-ASR/aishell/tdnn_lstm_ctc.rst", "recipes/Non-streaming-ASR/index.rst", "recipes/Non-streaming-ASR/librispeech/conformer_ctc.rst", "recipes/Non-streaming-ASR/librispeech/distillation.rst", "recipes/Non-streaming-ASR/librispeech/index.rst", "recipes/Non-streaming-ASR/librispeech/pruned_transducer_stateless.rst", "recipes/Non-streaming-ASR/librispeech/tdnn_lstm_ctc.rst", "recipes/Non-streaming-ASR/librispeech/zipformer_ctc_blankskip.rst", "recipes/Non-streaming-ASR/librispeech/zipformer_mmi.rst", "recipes/Non-streaming-ASR/timit/index.rst", "recipes/Non-streaming-ASR/timit/tdnn_ligru_ctc.rst", "recipes/Non-streaming-ASR/timit/tdnn_lstm_ctc.rst", "recipes/Non-streaming-ASR/yesno/index.rst", "recipes/Non-streaming-ASR/yesno/tdnn.rst", "recipes/Streaming-ASR/index.rst", "recipes/Streaming-ASR/introduction.rst", "recipes/Streaming-ASR/librispeech/index.rst", "recipes/Streaming-ASR/librispeech/lstm_pruned_stateless_transducer.rst", "recipes/Streaming-ASR/librispeech/pruned_transducer_stateless.rst", "recipes/Streaming-ASR/librispeech/zipformer_transducer.rst", "recipes/index.rst"], "titles": ["Follow the code style", "Contributing to Documentation", "How to create a recipe", "Contributing", "LODR for RNN Transducer", "Decoding with language models", "LM rescoring for Transducer", "Shallow fusion for Transducer", "Frequently Asked Questions (FAQs)", "Huggingface", "Pre-trained models", "Huggingface spaces", "Icefall", "Installation", "Export model.state_dict()", "Export to ncnn", "Export ConvEmformer transducer models to ncnn", "Export LSTM transducer models to ncnn", "Export streaming Zipformer transducer models to ncnn", "Export to ONNX", "Export model with torch.jit.script()", "Export model with torch.jit.trace()", "Model export", "Conformer CTC", "aishell", "Stateless Transducer", "TDNN-LSTM CTC", "Non Streaming ASR", "Conformer CTC", "Distillation with HuBERT", "LibriSpeech", "Pruned transducer statelessX", "TDNN-LSTM-CTC", "Zipformer CTC Blank Skip", "Zipformer MMI", "TIMIT", "TDNN-LiGRU-CTC", "TDNN-LSTM-CTC", "YesNo", "TDNN-CTC", "Streaming ASR", "Introduction", "LibriSpeech", "LSTM Transducer", "Pruned transducer statelessX", "Zipformer Transducer", "Recipes"], "terms": {"we": [0, 1, 2, 3, 4, 6, 7, 8, 10, 11, 13, 14, 15, 16, 17, 18, 19, 20, 21, 22, 23, 24, 25, 26, 28, 29, 31, 32, 33, 34, 36, 37, 39, 41, 43, 44, 45, 46], "us": [0, 1, 2, 4, 5, 6, 7, 8, 9, 11, 12, 13, 15, 16, 17, 18, 19, 22, 23, 24, 25, 26, 28, 29, 32, 36, 37, 39, 41], "tool": [0, 8, 13, 16], "make": [0, 1, 3, 16, 17, 18, 23, 25, 28, 41], "consist": [0, 25, 31, 43, 44, 45], "possibl": [0, 2, 3, 23, 28], "black": 0, "format": [0, 16, 17, 18, 23, 25, 26, 28, 31, 32, 33, 34, 36, 37, 39, 43, 44, 45], "flake8": 0, "check": [0, 13, 28], "qualiti": [0, 24], "isort": 0, "sort": [0, 13], "import": [0, 8, 13, 16, 44, 45], "The": [0, 1, 2, 4, 7, 8, 11, 13, 14, 16, 17, 18, 23, 24, 26, 28, 29, 31, 32, 33, 34, 36, 37, 39, 41, 43, 44, 45], "version": [0, 12, 13, 14, 16, 17, 18, 23, 25, 26, 28, 31, 32, 36, 37, 44], "abov": [0, 4, 6, 7, 8, 14, 16, 17, 18, 19, 23, 24, 25, 26, 28, 31, 33, 34, 39, 41, 43, 44, 45], "ar": [0, 1, 3, 4, 5, 6, 7, 8, 13, 14, 16, 17, 18, 23, 24, 25, 26, 28, 29, 31, 32, 33, 34, 36, 37, 39, 43, 44, 45, 46], "22": [0, 13, 16, 17, 28, 36, 37, 39], "3": [0, 4, 6, 7, 8, 12, 14, 15, 19, 22, 26, 29, 31, 32, 33, 34, 39, 43, 44, 45], "0": [0, 1, 4, 6, 7, 12, 14, 16, 17, 18, 19, 23, 25, 26, 28, 29, 31, 32, 33, 34, 36, 37, 39, 43, 44, 45], "5": [0, 7, 15, 22, 23, 25, 26, 28, 29, 31, 32, 33, 34, 36, 37, 39, 43, 44, 45], "4": [0, 4, 5, 6, 7, 8, 12, 14, 15, 22, 23, 25, 26, 28, 29, 31, 32, 33, 34, 36, 37, 39, 43, 44, 45], "10": [0, 7, 12, 13, 14, 16, 17, 18, 23, 25, 26, 28, 29, 31, 32, 33, 34, 36, 37, 39, 43, 44, 45], "1": [0, 4, 6, 7, 12, 14, 15, 19, 20, 21, 22, 29, 31, 32, 33, 34, 36, 37, 39, 43, 44, 45], "after": [0, 1, 6, 11, 13, 14, 16, 17, 18, 23, 25, 26, 28, 29, 31, 32, 33, 34, 36, 37, 39, 41, 43, 44, 45], "run": [0, 2, 8, 11, 13, 16, 17, 18, 19, 22, 23, 25, 26, 28, 29, 31, 32, 33, 34, 36, 37, 39, 43, 44, 45], "command": [0, 1, 4, 6, 7, 8, 13, 14, 16, 17, 21, 23, 25, 26, 28, 29, 31, 32, 33, 34, 36, 37, 39, 43, 44, 45], "git": [0, 4, 6, 7, 13, 14, 16, 17, 18, 19, 23, 25, 26, 28, 32, 36, 37, 39], "clone": [0, 4, 6, 7, 13, 14, 16, 17, 18, 19, 23, 25, 26, 28, 32, 36, 37, 39], "http": [0, 1, 2, 4, 6, 7, 8, 10, 11, 13, 14, 15, 16, 17, 18, 19, 20, 21, 23, 24, 25, 26, 28, 29, 31, 32, 33, 34, 36, 37, 39, 43, 44, 45], "github": [0, 2, 6, 10, 13, 14, 15, 16, 17, 18, 19, 20, 21, 23, 25, 26, 28, 31, 32, 33, 34, 36, 37, 39, 43, 44, 45], "com": [0, 2, 6, 10, 11, 13, 14, 16, 17, 20, 21, 23, 25, 26, 28, 29, 31, 32, 33, 34, 36, 37, 39, 43, 44, 45], "k2": [0, 2, 8, 10, 11, 12, 14, 15, 16, 17, 18, 19, 20, 21, 23, 25, 26, 28, 31, 32, 33, 34, 36, 37, 43, 44, 45], "fsa": [0, 2, 10, 11, 13, 14, 15, 16, 17, 18, 19, 20, 21, 23, 25, 28, 31, 33, 34, 43, 44, 45], "icefal": [0, 2, 3, 4, 6, 7, 8, 10, 11, 14, 15, 19, 20, 21, 22, 23, 25, 26, 28, 29, 31, 32, 33, 34, 36, 37, 39, 41, 43, 44, 45, 46], "cd": [0, 1, 2, 8, 13, 14, 16, 17, 18, 19, 20, 21, 23, 25, 26, 28, 29, 31, 32, 33, 34, 36, 37, 39, 43, 44, 45], "pip": [0, 1, 6, 8, 13, 16, 19, 25], "instal": [0, 1, 4, 6, 8, 9, 11, 12, 14, 15, 19, 22, 29, 31, 33, 34, 39, 43, 44, 45], "pre": [0, 3, 4, 6, 7, 9, 11, 12, 13, 15, 22, 29], "commit": [0, 13], "whenev": 0, "you": [0, 1, 2, 4, 6, 7, 8, 10, 11, 13, 14, 16, 17, 18, 19, 20, 21, 23, 25, 26, 28, 29, 31, 32, 33, 34, 36, 37, 39, 41, 43, 44, 45], "automat": [0, 11, 29], "hook": 0, "invok": 0, "fail": 0, "If": [0, 2, 4, 6, 7, 8, 11, 16, 17, 18, 20, 21, 23, 25, 26, 28, 29, 31, 32, 33, 34, 36, 37, 39, 41, 43, 44, 45], "ani": [0, 4, 6, 7, 13, 23, 25, 26, 28, 29, 31, 33, 34, 39, 43, 44], "your": [0, 1, 2, 4, 6, 7, 9, 11, 12, 16, 17, 18, 19, 23, 25, 26, 28, 29, 31, 32, 33, 34, 36, 37, 39, 43, 44, 45], "wa": [0, 14, 28, 32], "success": [0, 13, 16, 17], "pleas": [0, 1, 2, 4, 6, 7, 8, 11, 13, 15, 16, 17, 18, 19, 20, 21, 23, 25, 26, 28, 29, 31, 32, 33, 34, 36, 37, 39, 41, 43, 44, 45], "fix": [0, 8, 16, 17, 18, 28], "issu": [0, 4, 6, 7, 8, 13, 16, 17, 28, 29, 44, 45], "report": [0, 8, 29], "some": [0, 1, 4, 6, 14, 16, 17, 23, 25, 26, 28, 31, 32, 33, 34, 36, 37, 39, 43, 44, 45], "i": [0, 1, 2, 4, 7, 8, 11, 13, 14, 15, 16, 17, 18, 19, 23, 24, 25, 26, 28, 29, 31, 32, 33, 34, 36, 37, 39, 41, 43, 44, 45], "e": [0, 2, 4, 5, 6, 7, 16, 17, 18, 23, 25, 26, 28, 29, 31, 32, 33, 34, 36, 37, 39, 43, 44, 45], "modifi": [0, 15, 22, 23, 26, 28, 29, 31, 32, 33, 34, 36, 37, 39, 41, 43, 44, 45], "file": [0, 2, 11, 12, 14, 16, 17, 18, 20, 21, 22, 23, 25, 26, 28, 29, 31, 32, 33, 34, 36, 37, 39, 43, 44, 45], "place": [0, 13, 14, 25, 28, 32], "so": [0, 4, 6, 7, 11, 12, 13, 14, 16, 17, 18, 23, 25, 26, 28, 29, 31, 32, 33, 34, 36, 37, 39, 43, 44, 45], "statu": 0, "failur": 0, "see": [0, 1, 6, 7, 11, 13, 16, 17, 18, 19, 20, 21, 23, 25, 26, 28, 29, 31, 32, 33, 34, 36, 37, 39, 41, 43, 44, 45], "which": [0, 2, 4, 6, 7, 11, 13, 14, 16, 17, 18, 19, 23, 24, 25, 26, 28, 29, 31, 32, 33, 34, 36, 37, 39, 44, 45], "ha": [0, 2, 12, 13, 15, 16, 17, 18, 19, 23, 25, 26, 28, 31, 32, 33, 34, 36, 37, 41, 43, 44, 45], "been": [0, 13, 15, 16, 17, 18, 25], "befor": [0, 1, 13, 14, 16, 17, 18, 19, 20, 23, 25, 26, 28, 29, 31, 33, 34, 43, 44, 45], "further": [0, 4, 6, 7], "chang": [0, 4, 6, 7, 8, 13, 16, 17, 18, 23, 25, 26, 28, 29, 31, 32, 33, 34, 36, 37, 39, 43, 44, 45], "all": [0, 10, 11, 14, 16, 17, 18, 20, 23, 25, 26, 28, 29, 31, 32, 33, 34, 36, 37, 39, 41, 43, 44, 45], "again": [0, 16, 17, 39], "should": [0, 2, 4, 6, 16, 17, 18, 23, 25, 26, 28, 29, 31, 32, 33, 34, 36, 37, 39, 43, 44, 45], "succe": 0, "thi": [0, 2, 3, 4, 5, 6, 7, 8, 9, 13, 14, 16, 17, 18, 19, 20, 21, 22, 23, 25, 26, 28, 29, 31, 32, 33, 34, 36, 37, 39, 41, 43, 44, 45, 46], "time": [0, 13, 16, 17, 18, 23, 25, 26, 28, 29, 31, 32, 33, 34, 36, 37, 39, 41, 43, 44, 45], "succeed": 0, "want": [0, 4, 6, 7, 13, 14, 20, 21, 23, 25, 26, 28, 29, 31, 32, 33, 34, 36, 37, 39, 41, 43, 44, 45], "can": [0, 1, 2, 4, 6, 7, 8, 10, 11, 13, 14, 15, 16, 17, 18, 19, 20, 21, 23, 24, 25, 26, 28, 29, 31, 32, 33, 34, 36, 37, 39, 41, 43, 44, 45], "do": [0, 2, 4, 6, 23, 25, 26, 28, 29, 31, 32, 33, 34, 36, 37, 39, 41, 43, 44, 45], "Or": 0, "without": [0, 4, 6, 7, 9, 11, 23, 28], "your_changed_fil": 0, "py": [0, 2, 4, 6, 7, 8, 13, 16, 17, 18, 19, 20, 21, 22, 23, 25, 26, 28, 29, 31, 32, 33, 34, 36, 37, 39, 43, 44, 45], "sphinx": 1, "write": [1, 2, 3], "have": [1, 2, 4, 6, 7, 10, 11, 13, 14, 16, 17, 18, 19, 23, 25, 26, 28, 29, 31, 32, 33, 34, 36, 37, 39, 41, 43, 44, 45], "prepar": [1, 3, 4, 14], "environ": [1, 8, 16, 17, 18, 23, 24, 25, 26, 28, 29, 31, 32, 36, 37, 39, 44, 45], "doc": [1, 14, 41], "r": [1, 13, 16, 17, 18, 36, 37], "requir": [1, 4, 6, 13, 18, 29, 44, 45], "txt": [1, 4, 13, 16, 17, 18, 19, 23, 25, 26, 28, 32, 36, 37, 39], "set": [1, 4, 6, 7, 8, 13, 16, 17, 18, 23, 25, 26, 28, 29, 31, 33, 34, 39, 43, 44, 45], "up": [1, 13, 14, 16, 17, 18, 23, 26, 28, 29, 31, 32, 33, 34, 44, 45], "readi": [1, 23, 28, 29], "refer": [1, 2, 6, 7, 13, 14, 15, 16, 17, 18, 20, 21, 23, 25, 26, 28, 31, 32, 33, 36, 37, 39, 41, 44, 45], "restructuredtext": 1, "primer": 1, "familiar": 1, "build": [1, 13, 14, 16, 17, 18, 23, 25, 28], "local": [1, 13, 31, 33, 34, 43, 44, 45], "preview": 1, "what": [1, 2, 13, 16, 17, 18, 25, 41], "look": [1, 2, 4, 6, 7, 10, 13, 16, 17, 18, 23, 25, 26, 28, 29], "like": [1, 2, 11, 16, 17, 18, 23, 25, 26, 28, 31, 33, 34, 39, 41, 43, 44], "publish": [1, 14, 24], "html": [1, 2, 8, 13, 15, 16, 17, 18, 19, 20, 21, 31, 43, 44, 45], "gener": [1, 6, 14, 16, 17, 18, 19, 20, 21, 23, 25, 26, 28, 29, 31, 33, 34, 43, 44, 45], "view": [1, 16, 17, 18, 23, 25, 26, 28, 31, 33, 34, 39, 43, 44, 45], "follow": [1, 2, 3, 4, 5, 6, 7, 8, 10, 11, 12, 13, 14, 15, 16, 17, 18, 19, 20, 21, 23, 25, 26, 28, 29, 31, 32, 33, 34, 36, 37, 39, 43, 44, 45], "python3": [1, 8, 13, 17, 18], "m": [1, 13, 16, 17, 18, 25, 31, 33, 34, 36, 37, 43, 44, 45], "server": [1, 11, 43], "It": [1, 2, 6, 7, 9, 13, 15, 16, 17, 18, 19, 20, 21, 23, 24, 25, 26, 28, 31, 32, 33, 34, 36, 37, 39, 41, 43, 44, 45], "print": [1, 13, 23, 25, 26, 28, 29, 31, 32, 33, 34, 36, 37, 39, 43, 44, 45], "serv": [1, 31, 33, 34, 43, 44, 45], "port": [1, 29, 31, 33, 34, 43, 44, 45], "8000": [1, 39], "open": [1, 4, 6, 7, 12, 14, 16, 17, 18, 24, 25, 28, 29], "browser": [1, 9, 11, 31, 33, 34, 43, 44, 45], "go": [1, 7, 23, 25, 28, 31, 33, 34, 43, 44, 45], "read": [2, 13, 14, 16, 17, 18, 23, 25, 26, 28, 29, 31, 32, 33, 34, 36, 37, 39, 43, 44, 45], "code": [2, 3, 8, 12, 13, 16, 17, 18, 23, 28, 29, 31, 32, 36, 37, 39, 41, 44, 45], "style": [2, 3, 12], "adjust": 2, "sytl": 2, "design": 2, "python": [2, 13, 14, 16, 17, 18, 19, 20, 21, 23, 25, 28, 31, 33, 34, 43, 44, 45], "recommend": [2, 6, 7, 13, 23, 25, 26, 28, 29, 31, 44, 45], "test": [2, 4, 12, 14, 15, 22, 23, 25, 26, 28, 29, 32, 33, 36, 37], "valid": [2, 13, 18, 23, 25, 26, 28, 31, 32, 33, 34, 36, 37, 39, 43, 44, 45], "dataset": [2, 8, 13, 14, 23, 25, 26, 28, 29, 31, 32, 33, 34, 36, 37, 39, 41, 43, 44, 45], "lhots": [2, 12, 14, 16, 17, 18, 23, 25, 28], "readthedoc": [2, 13], "io": [2, 13, 15, 16, 17, 18, 19, 20, 21, 31, 43, 44, 45], "en": [2, 13, 16], "latest": [2, 11, 13, 28, 29, 31, 32, 33, 34, 43, 44, 45], "index": [2, 13, 15, 16, 17, 18, 19, 20, 21, 43, 44, 45], "yesno": [2, 8, 12, 13, 27, 39, 46], "veri": [2, 3, 7, 16, 17, 18, 25, 36, 37, 39, 44, 45], "good": [2, 7], "exampl": [2, 11, 12, 14, 16, 17, 18, 20, 21, 22, 29, 32, 36, 37, 39], "speech": [2, 11, 12, 13, 15, 24, 25, 39, 46], "pull": [2, 4, 6, 7, 16, 17, 18, 19, 23, 25, 28, 41], "380": [2, 16, 37], "show": [2, 4, 6, 7, 11, 13, 14, 16, 17, 18, 23, 25, 26, 28, 29, 31, 32, 33, 34, 36, 37, 39, 41, 43, 44, 45], "add": [2, 16, 17, 18, 23, 25, 26, 44, 46], "new": [2, 3, 11, 13, 16, 17, 18, 23, 24, 25, 26, 28, 29, 31, 32, 33, 34, 39, 43, 44, 45], "suppos": [2, 44, 45], "would": [2, 14, 16, 17, 18, 28, 32, 44, 45], "name": [2, 8, 14, 16, 17, 18, 19, 23, 25, 31, 33, 34, 44, 45], "foo": [2, 21, 23, 28, 31, 33, 34, 43, 44, 45], "eg": [2, 8, 10, 13, 14, 16, 17, 18, 19, 20, 21, 23, 25, 26, 28, 29, 31, 32, 33, 34, 36, 37, 39, 43, 44, 45], "mkdir": [2, 16, 17, 23, 25, 26, 28, 32, 36, 37, 39], "p": [2, 4, 13, 16, 17, 25, 36, 37], "asr": [2, 4, 6, 7, 8, 10, 12, 13, 14, 16, 17, 18, 19, 20, 21, 23, 25, 26, 28, 29, 31, 32, 33, 34, 36, 37, 39, 41, 43, 44, 45, 46], "touch": 2, "sh": [2, 13, 14, 23, 25, 26, 28, 29, 31, 32, 33, 34, 36, 37, 39, 43, 44, 45], "chmod": 2, "x": [2, 4, 18, 41], "simpl": [2, 13, 25], "own": [2, 29, 31, 44, 45], "otherwis": [2, 16, 17, 18, 23, 25, 28, 29, 31, 33, 34, 43, 44, 45], "librispeech": [2, 4, 6, 7, 8, 10, 12, 14, 16, 17, 18, 19, 20, 21, 27, 28, 29, 31, 32, 33, 34, 40, 41, 43, 44, 45, 46], "assum": [2, 4, 13, 14, 16, 17, 18, 19, 23, 25, 26, 28, 29, 31, 32, 36, 37, 39, 43, 44, 45], "fanci": 2, "call": [2, 8, 19, 29], "bar": [2, 21, 23, 28, 31, 33, 34, 43, 44, 45], "organ": 2, "wai": [2, 3, 22, 31, 33, 34, 41, 43, 44, 45], "readm": [2, 23, 25, 26, 28, 32, 36, 37, 39], "md": [2, 10, 14, 23, 25, 26, 28, 31, 32, 33, 34, 36, 37, 39, 43, 44, 45], "asr_datamodul": [2, 8, 13], "pretrain": [2, 4, 6, 7, 14, 16, 17, 18, 19, 21, 23, 25, 26, 28, 32, 36, 37, 39], "For": [2, 4, 6, 7, 8, 10, 14, 16, 17, 18, 23, 25, 26, 28, 29, 31, 32, 33, 34, 36, 37, 39, 43, 44, 45], "instanc": [2, 8, 10, 16, 17, 18, 23, 25, 26, 28, 31, 32, 33, 34, 36, 37, 39, 43, 44, 45], "tdnn": [2, 8, 13, 24, 27, 30, 35, 38], "its": [2, 4, 14, 15, 16, 17, 18, 21, 25, 33], "directori": [2, 12, 13, 16, 17, 18, 23, 25, 26, 28, 29, 31, 32, 33, 34, 36, 37, 39, 43, 44, 45], "structur": [2, 18], "descript": [2, 23, 25, 26, 28, 31, 32, 33, 34, 36, 37, 39, 43, 44, 45], "contain": [2, 12, 14, 15, 16, 17, 18, 23, 25, 26, 28, 29, 31, 32, 33, 34, 36, 37, 39, 43, 44, 45, 46], "inform": [2, 4, 6, 13, 14, 23, 25, 26, 28, 31, 32, 33, 36, 37, 39, 41, 43, 44, 45], "g": [2, 4, 5, 6, 7, 13, 18, 23, 25, 26, 28, 29, 31, 32, 33, 34, 36, 37, 39, 43, 44, 45], "wer": [2, 13, 14, 31, 32, 33, 34, 36, 37, 39, 43, 44, 45], "etc": [2, 23, 25, 26, 28, 29, 31, 32, 33, 34, 36, 37, 39, 41, 43, 44, 45], "provid": [2, 11, 13, 14, 15, 16, 17, 18, 23, 24, 25, 26, 28, 29, 31, 32, 33, 34, 36, 37, 39, 43, 44, 45, 46], "pytorch": [2, 8, 13, 16, 17, 18, 25], "dataload": [2, 13], "take": [2, 7, 14, 29, 31, 39, 44, 45], "input": [2, 14, 16, 17, 18, 23, 25, 26, 28, 32, 36, 37, 39, 41], "checkpoint": [2, 4, 6, 7, 13, 14, 16, 17, 18, 23, 25, 26, 28, 31, 32, 33, 34, 36, 37, 39, 43, 44, 45], "save": [2, 13, 14, 17, 18, 20, 23, 25, 26, 28, 29, 31, 32, 33, 34, 36, 37, 39, 43, 44, 45], "dure": [2, 4, 5, 7, 8, 11, 14, 23, 25, 26, 28, 29, 31, 32, 33, 34, 36, 37, 39, 43, 44, 45], "stage": [2, 13, 23, 25, 26, 28, 29, 31, 32, 33, 34, 36, 37, 39, 43, 44, 45], "": [2, 4, 6, 7, 13, 14, 16, 17, 18, 19, 20, 23, 25, 26, 28, 29, 31, 32, 33, 34, 36, 37, 39, 43, 44, 45], "definit": [2, 16, 17], "neural": [2, 4, 6, 7, 23, 28], "network": [2, 23, 25, 28, 31, 33, 34, 43, 44, 45], "script": [2, 6, 7, 12, 13, 21, 22, 23, 25, 26, 28, 29, 32, 36, 37, 39, 43], "infer": [2, 14, 16, 17], "tdnn_lstm_ctc": [2, 26, 32, 37], "conformer_ctc": [2, 23, 28], "get": [2, 11, 13, 16, 17, 18, 23, 25, 26, 28, 29, 31, 32, 33, 34, 39, 41, 43, 44, 45], "feel": [2, 29, 43], "result": [2, 4, 7, 10, 11, 14, 16, 17, 18, 23, 25, 26, 28, 29, 31, 32, 33, 34, 36, 37, 39, 43, 44, 45], "everi": [2, 14, 31, 33, 34, 43, 44, 45], "kept": [2, 31, 44, 45], "self": [2, 15, 18, 41], "toler": 2, "duplic": 2, "among": [2, 13], "differ": [2, 13, 16, 17, 18, 19, 23, 24, 28, 29, 31, 41, 43, 44, 45], "invoc": [2, 16, 17], "help": [2, 23, 25, 26, 28, 31, 32, 33, 34, 36, 37, 39, 43, 44, 45], "blob": [2, 10, 13, 14, 21, 31, 33, 34, 43, 44, 45], "master": [2, 6, 10, 13, 14, 17, 18, 20, 21, 25, 29, 31, 33, 34, 43, 44, 45], "transform": [2, 6, 7, 23, 28, 43], "conform": [2, 20, 24, 25, 27, 30, 31, 33, 43, 44, 45], "base": [2, 4, 7, 18, 23, 25, 26, 28, 29, 31, 33, 34, 43, 44, 45], "lstm": [2, 15, 21, 22, 24, 27, 30, 35, 40, 42], "attent": [2, 18, 25, 26, 29, 41, 44, 45], "lm": [2, 4, 7, 12, 13, 25, 31, 32, 36, 37, 39, 44, 45], "rescor": [2, 12, 26, 32, 34, 36, 37, 39], "demonstr": [2, 9, 11, 14, 19], "consid": [2, 4, 18], "colab": [2, 13], "notebook": [2, 13], "welcom": 3, "There": [3, 4, 16, 17, 18, 19, 23, 25, 26, 28, 29, 31, 33, 34, 43, 44, 45], "mani": [3, 13, 44, 45], "two": [3, 4, 16, 17, 18, 23, 25, 26, 28, 29, 31, 32, 33, 34, 36, 37, 39, 41, 43, 44, 45], "them": [3, 5, 6, 9, 10, 11, 16, 18, 23, 25, 26, 28, 29, 31, 32, 33, 34, 36, 37, 39, 43, 44, 45], "To": [3, 4, 6, 7, 11, 13, 23, 25, 26, 28, 29, 31, 32, 33, 34, 36, 37, 39, 43, 44, 45], "document": [3, 12, 14, 15, 16, 17, 18, 19, 34], "repositori": [3, 16, 17, 18, 19], "recip": [3, 4, 6, 7, 10, 12, 13, 14, 19, 23, 25, 26, 28, 29, 31, 32, 36, 37, 39, 41, 43, 44, 45], "In": [3, 4, 6, 8, 11, 13, 14, 16, 17, 18, 19, 20, 21, 22, 23, 25, 26, 28, 29, 32, 36, 37, 39, 41], "page": [3, 11, 20, 23, 25, 26, 28, 29, 31, 32, 33, 34, 36, 37, 39, 41, 43, 44, 45, 46], "describ": [3, 5, 9, 14, 16, 17, 19, 20, 21, 22, 23, 25, 26, 28, 31, 32, 36, 37, 44, 45], "how": [3, 4, 5, 6, 7, 9, 11, 12, 13, 16, 17, 18, 19, 22, 23, 25, 26, 28, 29, 31, 32, 33, 34, 36, 37, 39, 41, 43, 44, 45], "creat": [3, 4, 6, 7, 12, 14, 16, 17, 18, 23, 25, 26, 28, 31, 32, 33, 34, 36, 37, 39, 43, 44], "data": [3, 4, 6, 7, 14, 16, 17, 18, 19, 20, 21, 24], "train": [3, 4, 6, 7, 8, 9, 11, 12, 14, 15, 20, 21, 22, 41], "decod": [3, 4, 8, 11, 12, 16, 17, 18, 21, 22], "model": [3, 4, 6, 7, 9, 11, 12, 13, 15, 29, 41], "As": [4, 5, 6, 7, 16, 25, 28, 29], "type": [4, 6, 7, 13, 14, 16, 17, 18, 23, 25, 28, 31, 33, 34, 39, 41, 43, 44, 45], "e2": [4, 7, 13], "usual": [4, 6, 7, 23, 25, 26, 28, 29, 31, 33, 34, 43, 44, 45], "an": [4, 5, 6, 7, 11, 13, 14, 16, 17, 18, 19, 20, 21, 23, 24, 25, 28, 29, 31, 34, 39, 43, 44, 45], "intern": [4, 5], "languag": [4, 7, 11, 12, 23, 25, 26], "learn": [4, 23, 25, 26, 28, 31, 32, 33, 34, 36, 37, 39, 43, 44, 45], "level": [4, 5], "corpu": [4, 6, 7, 24], "real": 4, "life": 4, "scenario": 4, "often": [4, 23, 25, 26, 28, 31, 33, 34, 43, 44, 45], "mismatch": [4, 44], "between": [4, 7, 31, 44, 45], "target": [4, 11, 13], "space": [4, 9, 12], "problem": [4, 6, 7, 13, 29], "when": [4, 6, 8, 11, 16, 17, 18, 22, 25, 28, 29, 31, 33, 34, 44, 45], "act": 4, "against": [4, 13], "extern": [4, 5, 6, 7], "tutori": [4, 6, 7, 23, 25, 26, 28, 29, 31, 32, 33, 34, 36, 37, 43, 44, 45], "low": [4, 16, 17], "order": [4, 13, 16, 17, 18, 23, 26, 28, 32, 36, 37], "densiti": 4, "ratio": 4, "allevi": 4, "effect": [4, 7, 18], "improv": [4, 5, 6, 7, 25], "perform": [4, 6, 7, 15, 25, 29, 44], "languga": 4, "integr": [4, 11], "pruned_transducer_stateless7_stream": [4, 6, 7, 18, 19, 45], "stream": [4, 6, 7, 12, 15, 16, 17, 19, 22, 23, 28, 36, 37, 43, 46], "howev": [4, 6, 7, 14, 17, 29], "easili": [4, 6, 7, 23, 26, 28], "appli": [4, 6, 7, 25, 41], "other": [4, 7, 14, 17, 18, 19, 25, 28, 29, 31, 32, 36, 37, 39, 41, 44, 45, 46], "encount": [4, 6, 7, 8, 13, 18, 23, 25, 26, 28, 29, 31, 33, 34, 43, 44, 45], "here": [4, 6, 7, 14, 16, 17, 18, 23, 25, 26, 28, 29, 32, 41, 44], "simplic": [4, 6, 7], "same": [4, 6, 7, 13, 14, 16, 17, 18, 23, 25, 26, 28, 29, 31, 32, 33, 34, 36, 37, 39, 41, 43, 44, 45], "domain": [4, 6, 7], "gigaspeech": [4, 6, 7, 10, 20, 43], "first": [4, 6, 8, 13, 16, 17, 18, 23, 25, 26, 28, 29, 31, 32, 33, 34, 36, 37, 39, 43, 44, 45], "let": [4, 6, 7, 13, 16, 17, 18, 23, 28], "background": 4, "predecessor": 4, "dr": 4, "propos": [4, 25, 41, 45], "address": [4, 11, 13, 14, 16, 17, 18, 25, 31, 34, 43, 44, 45], "sourc": [4, 13, 14, 16, 17, 18, 23, 24, 25, 28], "acoust": [4, 44, 45], "similar": [4, 5, 29, 33, 44, 45], "deriv": 4, "formular": 4, "bay": 4, "theorem": 4, "text": [4, 6, 7, 16, 17, 18, 23, 25, 26, 28, 31, 32, 33, 34, 36, 37, 39, 43, 44, 45], "score": [4, 5, 7, 23, 28, 31, 44, 45], "left": [4, 16, 18, 25, 44, 45], "y_u": 4, "mathit": 4, "y": 4, "right": [4, 16, 25, 41, 44], "log": [4, 8, 13, 16, 17, 18, 32, 36, 37, 39], "y_": 4, "u": [4, 13, 16, 17, 18, 23, 25, 26, 28, 29, 39], "lambda_1": 4, "p_": 4, "lambda_2": 4, "where": [4, 8, 44], "weight": [4, 23, 26, 28, 33, 34, 43], "respect": 4, "onli": [4, 6, 14, 16, 17, 18, 23, 25, 26, 28, 29, 31, 32, 33, 34, 36, 37, 39, 41, 43, 44, 45, 46], "compar": [4, 16, 17, 18, 44], "shallow": [4, 12], "fusion": [4, 12], "subtract": [4, 5], "work": [4, 16, 17, 18, 28], "treat": [4, 17, 18], "predictor": 4, "joiner": [4, 16, 17, 18, 19, 21, 25, 31, 43, 44, 45], "weak": 4, "captur": 4, "therefor": [4, 8], "n": [4, 5, 6, 13, 23, 29, 31, 33, 34, 36, 37, 43, 44, 45], "gram": [4, 6, 13, 23, 25, 26, 31, 32, 34, 36, 37, 44, 45], "approxim": [4, 5], "ilm": 4, "lead": [4, 7], "formula": 4, "rnnt": [4, 31, 44, 45], "bi": [4, 6], "addit": 4, "estim": 4, "comar": 4, "li": 4, "choic": 4, "accord": 4, "origin": [4, 5], "paper": [4, 5, 29, 31, 43, 44, 45], "achiev": [4, 6, 7, 41], "both": [4, 31, 33, 34, 41, 43, 44, 45], "intra": 4, "cross": 4, "much": [4, 16, 17], "faster": [4, 6], "evalu": 4, "now": [4, 6, 13, 16, 17, 18, 23, 28, 29, 31, 32, 33, 34, 36, 37, 43, 44, 45], "illustr": [4, 6, 7], "purpos": [4, 6, 7, 16, 17], "from": [4, 6, 7, 8, 9, 11, 13, 14, 16, 17, 18, 19, 23, 24, 25, 26, 28, 29, 31, 32, 33, 34, 36, 37, 39, 41, 43, 44, 45], "link": [4, 6, 7, 10, 13, 14, 15, 31, 33, 34, 43, 44, 45], "scratch": [4, 6, 7, 31, 33, 34, 43, 44, 45], "prune": [4, 6, 7, 14, 18, 19, 25, 27, 29, 30, 40, 41, 42, 43, 45], "statelessx": [4, 6, 7, 27, 29, 30, 40, 41, 42], "initi": [4, 6, 7, 23, 26], "step": [4, 6, 7, 13, 14, 16, 17, 18, 23, 25, 26, 28, 29, 31, 33, 34, 39, 43, 44, 45], "download": [4, 6, 7, 8, 11, 12, 15, 22, 24, 29], "git_lfs_skip_smudg": [4, 6, 7, 16, 17, 18, 19], "huggingfac": [4, 6, 7, 10, 12, 13, 14, 16, 17, 18, 19, 23, 25, 26, 28, 32, 33, 34, 36, 37, 39, 43], "co": [4, 6, 7, 10, 11, 13, 14, 16, 17, 18, 19, 23, 24, 25, 26, 28, 32, 33, 34, 36, 37, 39, 43], "zengwei": [4, 6, 7, 16, 18, 19, 34, 43], "stateless7": [4, 6, 7, 18, 19], "2022": [4, 6, 7, 14, 16, 17, 18, 19, 25, 31, 33, 34, 43, 44], "12": [4, 6, 7, 13, 14, 16, 17, 18, 19, 23, 25, 26, 28, 31, 33, 34, 36, 39, 43, 44, 45], "29": [4, 6, 7, 13, 18, 19, 23, 25, 26, 28, 32, 33, 36, 37], "pushd": [4, 6, 7, 19], "exp": [4, 6, 7, 13, 14, 16, 17, 18, 19, 20, 21, 23, 25, 26, 28, 29, 31, 32, 33, 34, 36, 37, 39, 43, 44, 45], "lf": [4, 6, 7, 14, 16, 17, 18, 19, 23, 25, 26, 28, 32, 34, 36, 37, 39], "includ": [4, 6, 7, 16, 17, 18, 19, 31, 33, 34, 43, 44, 45], "pt": [4, 6, 7, 13, 14, 16, 17, 18, 19, 20, 21, 23, 25, 26, 28, 31, 32, 33, 34, 36, 37, 39, 43, 44, 45], "ln": [4, 6, 7, 14, 16, 17, 18, 19, 23, 28, 31, 33, 34, 43, 44, 45], "epoch": [4, 6, 7, 13, 14, 16, 17, 18, 19, 20, 23, 25, 26, 28, 29, 31, 32, 33, 34, 36, 37, 39, 43, 44, 45], "99": [4, 6, 7, 13, 16, 17, 18, 19], "symbol": [4, 5, 6, 7, 13, 25, 31, 44, 45], "load": [4, 6, 7, 13, 16, 17, 18, 23, 25, 26, 28, 31, 32, 33, 34, 36, 37, 39, 43, 44, 45], "done": [4, 6, 7, 13, 14, 23, 25, 26, 28, 31, 32, 33, 34, 36, 37, 39, 43, 44, 45], "via": [4, 6, 7, 13, 15, 20, 21, 22], "exp_dir": [4, 6, 7, 13, 16, 17, 18, 25, 28, 29, 31, 33, 34, 44, 45], "avg": [4, 6, 7, 13, 14, 16, 17, 18, 19, 20, 21, 25, 28, 29, 31, 32, 33, 34, 36, 37, 39, 43, 44, 45], "averag": [4, 6, 7, 13, 14, 16, 17, 18, 19, 23, 25, 26, 28, 31, 32, 33, 34, 36, 37, 39, 43, 44, 45], "fals": [4, 6, 7, 13, 14, 16, 17, 18, 23, 25, 28, 29], "dir": [4, 6, 7, 14, 16, 17, 18, 19, 20, 21, 23, 25, 26, 28, 29, 31, 32, 33, 34, 36, 37, 39, 43, 44, 45], "bpe": [4, 5, 6, 7, 14, 16, 17, 18, 19, 20, 21, 28, 31, 33, 34, 43, 44, 45], "lang_bpe_500": [4, 6, 7, 14, 16, 17, 18, 19, 20, 21, 28, 31, 33, 34, 43, 44, 45], "max": [4, 6, 7, 13, 14, 16, 17, 23, 25, 26, 28, 29, 31, 33, 34, 43, 44, 45], "durat": [4, 6, 7, 14, 23, 25, 26, 28, 29, 31, 32, 33, 34, 36, 37, 39, 43, 44, 45], "600": [4, 6, 7, 13, 14, 28, 31, 33, 43, 44, 45], "chunk": [4, 6, 7, 16, 18, 19, 44, 45], "len": [4, 6, 7, 18, 19, 45], "32": [4, 6, 7, 13, 16, 17, 18, 19, 23, 25, 26, 45], "method": [4, 7, 11, 14, 23, 25, 26, 28, 29, 31, 32, 33, 34, 36, 37, 43, 44, 45], "modified_beam_search": [4, 5, 6, 7, 11, 25, 29, 31, 33, 43, 44, 45], "clean": [4, 13, 18, 23, 25, 28, 29, 31, 32, 33, 34, 43, 44, 45], "beam_size_4": [4, 6, 7], "11": [4, 6, 7, 8, 13, 16, 17, 19, 23, 25, 26, 28, 31, 32, 33, 34, 36, 37, 39, 43, 44, 45], "best": [4, 5, 6, 7, 16, 17, 18, 23, 26, 28], "7": [4, 6, 7, 13, 14, 15, 18, 22, 23, 26, 28, 31, 32, 36, 37, 43, 44], "93": [4, 6, 7], "Then": [4, 6], "necessari": [4, 29], "note": [4, 5, 6, 7, 8, 14, 16, 23, 25, 26, 28, 31, 32, 33, 34, 36, 37, 39, 43, 44, 45], "960": [4, 28, 31, 33, 34, 43, 44, 45], "hour": [4, 23, 25, 26, 28, 31, 33, 34, 43, 44, 45], "ezerhouni": [4, 6, 7], "popd": [4, 6, 7, 19], "marcoyang": [4, 6], "librispeech_bigram": [4, 6], "2gram": [4, 6], "fst": [4, 13, 25, 39], "modified_beam_search_lm_lodr": 4, "lm_dir": [4, 6, 7, 13, 28], "lm_scale": [4, 6, 7], "42": [4, 13, 17, 23, 28, 39], "lodr_scal": 4, "24": [4, 8, 13, 16, 17, 26, 32, 36, 37, 39], "scale": [4, 6, 7, 16, 17, 23, 28, 29, 32, 34, 36, 37], "embed": [4, 6, 7, 25, 31, 43, 44, 45], "dim": [4, 6, 7, 16, 17, 18, 25, 31, 44], "2048": [4, 6, 7, 14, 16, 17, 18, 25], "hidden": [4, 6, 7, 17, 43], "num": [4, 6, 7, 16, 17, 18, 23, 25, 26, 28, 29, 31, 33, 34, 43, 44, 45], "layer": [4, 6, 7, 16, 17, 18, 25, 29, 31, 41, 43, 44, 45], "vocab": [4, 6, 7, 28], "500": [4, 6, 7, 14, 16, 17, 18, 25, 28, 34, 43], "token": [4, 16, 17, 18, 19, 23, 25, 26, 28, 32, 36, 37, 39], "ngram": [4, 28, 32, 36, 37], "2": [4, 6, 7, 12, 14, 15, 22, 29, 31, 32, 33, 34, 36, 37, 39, 43, 44, 45], "extra": [4, 16, 17, 18, 25, 41, 44], "argument": [4, 7, 29, 41], "need": [4, 6, 11, 13, 14, 15, 16, 17, 18, 23, 25, 26, 28, 29, 31, 32, 33, 34, 36, 37, 39, 41, 43, 44, 45], "given": [4, 13, 14, 16, 17, 18, 23, 25, 26, 28, 31, 32, 33, 34, 44, 45], "specifi": [4, 7, 8, 16, 17, 18, 23, 25, 26, 28, 29, 31, 32, 33, 34, 36, 37, 39, 43, 44, 45], "neg": [4, 25], "number": [4, 7, 11, 14, 16, 17, 18, 23, 25, 26, 28, 31, 32, 33, 34, 36, 37, 39, 43, 44, 45], "obtain": [4, 7, 23, 25, 26, 28, 32, 36, 37], "shown": [4, 7], "below": [4, 7, 13, 16, 17, 18, 23, 25, 26, 28, 31, 32, 33, 34, 36, 37, 39, 43, 44], "61": [4, 6], "6": [4, 6, 7, 8, 15, 22, 23, 25, 28, 31, 32, 36, 37, 43], "74": [4, 6, 13, 14], "recal": 4, "lowest": [4, 31, 33, 34, 43, 44, 45], "77": [4, 6, 7, 13, 28], "08": [4, 6, 7, 18, 28, 32, 34, 36, 37, 39, 43], "inde": 4, "even": [4, 11, 13, 17], "better": [4, 6], "increas": [4, 6, 23, 25, 26, 28, 31, 33, 34, 43, 44, 45], "8": [4, 6, 7, 8, 13, 14, 16, 17, 18, 23, 25, 28, 29, 31, 32, 33, 34, 39, 43, 44, 45], "45": [4, 6, 13, 16, 18, 23, 25, 28], "38": [4, 6, 13, 16, 23, 25, 28, 36], "23": [4, 6, 8, 13, 16, 17, 18, 23, 25, 26, 28, 36, 37, 39], "section": [5, 8, 9, 13, 14, 19, 20, 21, 22, 23, 28], "langugag": 5, "transduc": [5, 12, 14, 15, 19, 22, 24, 27, 29, 30, 40, 41, 42], "avail": [5, 6, 13, 14, 16, 17, 18, 23, 25, 28, 32, 36, 37, 39, 43], "beam": [5, 14, 43], "search": [5, 6, 7, 10, 11], "realli": [5, 23, 26, 28, 31, 33, 34, 43, 44, 45], "valu": [5, 7, 16, 17, 18, 23, 25, 26, 28, 31, 33, 34, 43, 44, 45], "rnn": [5, 6, 7, 12, 17, 25, 31, 33, 43, 44, 45], "t": [5, 13, 16, 17, 18, 19, 20, 23, 25, 26, 28, 29, 31, 32, 33, 34, 36, 37, 39, 43, 44, 45], "doe": [5, 16, 17, 18, 23, 25, 28, 39], "modified_beam_search_lm_shallow_fus": [5, 6, 7], "interpol": 5, "also": [5, 6, 7, 9, 10, 13, 14, 15, 16, 17, 18, 19, 21, 23, 25, 26, 28, 31, 33, 34, 39, 41, 43, 44, 45], "known": 5, "modified_beam_search_lodr": [5, 6], "bigram": 5, "backoff": 5, "modified_beam_search_lm_rescor": [5, 6], "hypothes": [5, 6], "rnnlm": [5, 6], "re": [5, 6, 8, 23, 26, 28, 29, 31, 33, 34, 41, 43, 44, 45], "rank": [5, 6], "modified_beam_search_lm_rescore_lodr": [5, 6], "lodr": [5, 12], "commonli": [6, 7, 23, 25, 26, 28, 32, 36, 37, 39], "approach": 6, "incorpor": 6, "unlik": 6, "more": [6, 13, 16, 17, 18, 23, 28, 29, 39, 41, 43, 44], "effici": [6, 7, 31, 44, 45], "than": [6, 13, 14, 17, 23, 25, 26, 28, 31, 32, 33, 34, 39, 43, 44, 45], "sinc": [6, 13, 16, 17, 18, 29, 39, 43], "less": [6, 14, 28, 32, 39, 44, 45], "comput": [6, 13, 14, 16, 17, 18, 23, 25, 26, 29, 31, 32, 34, 36, 37, 39, 43, 44, 45], "gpu": [6, 7, 13, 16, 17, 23, 25, 26, 28, 29, 31, 33, 34, 36, 37, 39, 43, 44, 45], "try": [6, 8, 9, 11, 29, 31, 33, 34, 43, 44, 45], "might": [6, 7, 17, 18, 44, 45], "ideal": [6, 7], "mai": [6, 7, 13, 16, 17, 18, 23, 25, 26, 28, 31, 33, 34, 43, 44, 45, 46], "With": [6, 13], "43": [6, 17, 18, 28], "great": 6, "made": [6, 16], "boost": [6, 7], "tabl": [6, 11, 16, 17, 18], "67": [6, 13], "59": [6, 13, 16, 26, 28], "86": 6, "fact": 6, "arpa": [6, 39], "performn": 6, "depend": [6, 13, 23, 28], "kenlm": 6, "kpu": 6, "archiv": 6, "zip": 6, "execut": [6, 7, 16, 23, 26, 28, 29, 31, 32, 33, 34, 36, 37, 39, 43, 44, 45], "9": [6, 13, 16, 17, 18, 23, 25, 26, 28, 31, 32, 33, 34, 36, 39, 43, 44, 45], "57": [6, 13, 17, 28, 32], "slightli": 6, "63": [6, 25], "04": [6, 16, 17, 18, 23, 25, 26, 28, 32, 36, 37], "52": [6, 13, 23, 28], "73": 6, "mention": 6, "earlier": 6, "benchmark": [6, 25], "speed": [6, 16, 23, 25, 26, 28, 31, 33, 34, 43, 44, 45], "132": 6, "95": [6, 24], "177": [6, 13, 14, 17, 18, 25, 26, 28], "96": [6, 13], "210": [6, 36, 37], "262": [6, 7], "62": [6, 7, 13, 28, 32], "65": [6, 7, 13, 16], "352": [6, 7, 28], "58": [6, 7, 8, 13, 28], "488": [6, 7, 16, 17, 18], "400": [6, 24], "610": 6, "870": 6, "156": 6, "203": [6, 14, 28], "255": [6, 17, 18], "160": 6, "263": [6, 13, 17], "singl": [6, 23, 25, 26, 28, 31, 32, 33, 34, 36, 37, 39, 43, 44, 45], "32g": 6, "v100": [6, 23, 25, 26, 28], "vari": 6, "word": [7, 23, 25, 26, 28, 32, 36, 37, 39], "error": [7, 8, 13, 16, 17, 18, 28], "rate": [7, 23, 25, 26, 28, 31, 32, 33, 34, 36, 37, 39, 43, 44, 45], "These": [7, 23, 25, 26, 28, 31, 32, 33, 34, 36, 37, 39, 43, 44, 45], "alreadi": [7, 13, 14], "But": [7, 16, 31, 33, 34, 43, 44, 45], "long": [7, 16], "true": [7, 13, 14, 16, 17, 18, 23, 25, 28, 29, 31, 32, 33, 34, 36, 37, 39, 43, 44, 45], "either": [7, 11, 23, 25, 26, 28, 44, 45], "choos": [7, 11, 13, 29, 31, 33, 34, 43, 44, 45], "three": [7, 16, 17, 18, 21, 23, 25, 41], "associ": 7, "dimens": [7, 31, 44, 45], "obviou": 7, "rel": 7, "reduct": [7, 13, 16, 17, 33], "around": 7, "A": [7, 14, 16, 17, 18, 23, 25, 26, 28, 31, 32, 33, 34, 43, 44, 45], "few": [7, 16, 17, 18, 29], "paramet": [7, 14, 16, 17, 18, 20, 23, 25, 26, 28, 31, 32, 33, 34, 36, 37, 43, 44, 45], "tune": [7, 16, 17, 18, 23, 25, 26, 28, 29, 31, 33, 34, 43, 44, 45], "control": [7, 23, 25, 26, 28, 29, 31, 32, 33, 34, 36, 37, 39, 43, 44, 45], "too": 7, "small": [7, 25, 36, 37, 39], "fulli": 7, "util": [7, 8, 13, 28], "larg": 7, "domin": 7, "bad": 7, "typic": [7, 23, 25, 26, 28], "activ": [7, 11, 13], "path": [7, 11, 13, 14, 16, 17, 18, 21, 23, 25, 26, 28, 29, 31, 33, 34, 43, 44, 45], "trade": 7, "off": [7, 16], "accuraci": [7, 16, 17, 24], "larger": [7, 17, 23, 25, 26, 28, 31, 33, 34, 43, 44, 45], "slower": 7, "collect": [8, 13], "user": 8, "post": 8, "correspond": [8, 10, 11], "solut": 8, "One": 8, "torch": [8, 12, 14, 15, 22, 23, 25, 28], "torchaudio": [8, 12, 41], "cu111": 8, "torchvis": 8, "f": [8, 13, 36, 37], "org": [8, 13, 24, 25, 31, 43, 44, 45], "whl": [8, 13], "torch_stabl": [8, 13], "throw": [8, 16, 17, 18], "cuda": [8, 12, 14, 16, 17, 18, 23, 25, 26, 28, 31, 32, 33, 34, 36, 37, 43, 44, 45], "while": [8, 13, 16, 17, 18, 23, 25, 26, 28, 29, 31, 33, 34, 43, 44, 45], "That": [8, 16, 17, 29, 31, 43, 44, 45], "cu11": 8, "correct": 8, "traceback": 8, "most": [8, 44, 45], "recent": [8, 16, 17, 18], "last": 8, "line": [8, 16, 17, 18, 31, 44, 45], "14": [8, 13, 14, 16, 17, 20, 23, 28, 31, 32, 33, 36, 43, 44, 45], "yesnoasrdatamodul": 8, "home": [8, 16, 17, 23, 28], "xxx": [8, 14, 16, 17, 18], "next": [8, 11, 13, 16, 17, 18, 28, 29, 31, 32, 33, 34, 43, 44, 45], "gen": [8, 11, 13, 28, 29, 31, 32, 33, 34, 43, 44, 45], "kaldi": [8, 11, 13, 28, 29, 31, 32, 33, 34, 43, 44, 45], "34": [8, 16, 17], "datamodul": 8, "__init__": [8, 13, 14, 16, 17, 18, 23, 25, 28], "add_eo": 8, "add_so": 8, "get_text": 8, "39": [8, 13, 16, 18, 25, 28, 32, 36], "tensorboard": [8, 13, 23, 25, 26, 28, 31, 32, 33, 34, 36, 37, 39, 43, 44, 45], "summarywrit": 8, "miniconda3": 8, "env": 8, "yyi": 8, "lib": [8, 13, 18], "site": [8, 13, 18], "packag": [8, 13, 18], "loosevers": 8, "uninstal": 8, "setuptool": [8, 13], "conda": 8, "dev": [8, 13, 14, 16, 17, 18, 23, 25, 26, 28, 31, 32, 33, 34, 36, 37, 39, 43, 44, 45], "yangyifan": 8, "anaconda3": 8, "dev20230112": 8, "cuda11": [8, 13], "torch1": [8, 13], "13": [8, 13, 14, 16, 17, 18, 25, 26, 28, 32, 33, 36], "py3": [8, 13], "linux": [8, 11, 13, 15, 16, 17, 18, 19], "x86_64": [8, 13, 16], "egg": 8, "_k2": [8, 13], "determinizeweightpushingtyp": 8, "handl": [8, 23, 26, 28, 29, 31, 32, 33, 34, 36, 37, 39, 43, 44, 45], "except": [8, 14], "anoth": 8, "occur": 8, "pruned_transducer_stateless7_ctc_b": [8, 33], "104": [8, 13], "30": [8, 13, 16, 17, 18, 23, 25, 26, 28, 29, 31, 33, 34, 39, 43, 44, 45], "rais": 8, "anaconda": 8, "maco": [8, 11, 15, 16, 17, 18, 19], "probabl": [8, 25, 31, 33, 43, 44, 45], "variabl": [8, 13, 16, 17, 18, 23, 26, 28, 29, 31, 33, 34, 43, 44, 45], "export": [8, 12, 13, 23, 25, 26, 28, 29, 32, 36, 37, 39], "dyld_library_path": 8, "conda_prefix": 8, "find": [8, 9, 10, 11, 14, 16, 17, 18, 21, 23, 25, 26, 28, 31, 32, 33, 34, 36, 37, 39, 43, 44, 45], "locat": [8, 16], "libpython": 8, "abl": 8, "insid": [8, 21], "codna_prefix": 8, "ld_library_path": 8, "within": [9, 11, 16, 17], "anyth": [9, 11], "youtub": [9, 12, 28, 29, 31, 32, 33, 34, 43, 44, 45], "video": [9, 12, 28, 29, 31, 32, 33, 34, 43, 44, 45], "upload": [10, 11, 23, 25, 26, 28, 29, 31, 32, 33, 34, 36, 37, 39, 43, 44, 45], "visit": [10, 11, 31, 33, 34, 43, 44, 45], "specif": [10, 19, 25], "aishel": [10, 12, 23, 25, 26, 27, 46], "wenetspeech": [10, 20], "framework": [11, 31, 44], "sherpa": [11, 15, 20, 21, 22, 43], "window": [11, 15, 16, 17, 18, 19], "ipad": 11, "phone": 11, "start": [11, 13, 14, 18, 23, 25, 26, 28, 31, 32, 33, 34, 36, 37, 39, 43, 44, 45], "recognit": [11, 12, 15, 16, 17, 24, 25, 39, 46], "screenshot": [11, 23, 25, 26, 28, 29, 31, 39, 43, 44], "select": [11, 13, 16, 17, 18, 31, 32, 36, 37, 39, 43, 44, 45], "current": [11, 16, 17, 25, 29, 41, 43, 44, 45, 46], "chines": [11, 24, 25], "english": [11, 39, 43], "greedi": 11, "record": [11, 17, 18, 23, 24, 25, 26, 28, 31, 32, 33, 34, 36, 37, 39, 43, 44, 45], "click": [11, 13, 23, 25, 26, 28, 31, 33, 34, 39, 43, 44], "button": 11, "submit": 11, "wait": 11, "moment": 11, "bottom": [11, 31, 33, 34, 43, 44, 45], "part": [11, 13, 23, 25, 26, 28, 31, 32, 33, 34, 36, 37, 39, 41, 43, 44, 45], "one": [11, 14, 16, 17, 18, 23, 25, 26, 28, 31, 32, 33, 34, 36, 37, 39, 41, 43, 44, 45], "subscrib": [11, 13, 28, 29, 31, 32, 33, 34, 43, 44, 45], "channel": [11, 13, 23, 25, 26, 28, 29, 31, 32, 33, 34, 36, 37, 39, 43, 44, 45], "nadira": [11, 13, 28, 29, 31, 32, 33, 34, 43, 44, 45], "povei": [11, 13, 28, 29, 31, 32, 33, 34, 43, 44, 45], "www": [11, 13, 24, 28, 29, 31, 32, 33, 34, 43, 44, 45], "uc_vaumpkminz1pnkfxan9mw": [11, 13, 28, 29, 31, 32, 33, 34, 43, 44, 45], "toolkit": 12, "cudnn": 12, "frequent": 12, "ask": 12, "question": 12, "faq": 12, "oserror": 12, "libtorch_hip": 12, "cannot": [12, 16, 17, 18], "share": [12, 13], "object": [12, 13, 23, 25, 26, 31, 39, 43, 44], "attributeerror": 12, "modul": [12, 16, 18, 33, 44], "distutil": 12, "attribut": [12, 18, 28], "importerror": 12, "libpython3": 12, "No": [12, 16, 17, 18, 39], "state_dict": [12, 22, 23, 25, 26, 28, 32, 36, 37, 39], "jit": [12, 15, 22, 28], "trace": [12, 15, 20, 22], "onnx": [12, 14, 22], "ncnn": [12, 22], "non": [12, 28, 41, 44, 46], "timit": [12, 27, 36, 37, 46], "introduct": [12, 40, 46], "contribut": 12, "guid": 13, "setup": [13, 16, 23, 25, 26, 28, 29, 31, 32, 36, 37, 39, 44, 45], "matter": [13, 16], "don": [13, 16, 17, 18, 20, 23, 26, 28, 31, 32, 33, 34, 36, 37, 39, 43, 44, 45], "suggest": [13, 31, 33, 34, 43, 44, 45], "compil": [13, 16, 17, 23, 25, 28], "wheel": [13, 16], "from_wheel": 13, "alwai": [13, 14], "strongli": 13, "pythonpath": [13, 16, 17, 18], "point": [13, 14, 23, 26, 28, 29, 31, 33, 34, 43, 44, 45], "folder": [13, 14, 23, 25, 26, 28, 31, 32, 33, 34, 36, 37, 39, 43, 44, 45], "tmp": [13, 23, 25, 26, 28, 29, 31, 32, 33, 34, 36, 37, 39, 43, 44, 45], "put": [13, 16, 17, 33, 44], "sever": [13, 14, 23, 25, 26, 28, 29, 31, 32, 33, 34, 36, 37, 39, 41, 43, 44, 45], "switch": [13, 23, 28, 34], "just": [13, 16, 17, 18, 41], "about": [13, 16, 17, 18, 25, 29, 31, 34, 43, 44, 45], "kuangfangjun": [13, 16, 17, 18], "virtualenv": 13, "cpython3": 13, "final": [13, 14, 16, 17, 28, 32], "64": [13, 14, 16, 25, 44], "9422m": 13, "creator": 13, "cpython3posix": 13, "dest": 13, "star": [13, 16, 17, 18], "fj": [13, 14, 16, 17, 18, 25, 28], "fangjun": [13, 14, 16, 17, 18, 25, 28], "clear": 13, "no_vcs_ignor": 13, "global": 13, "seeder": 13, "fromappdata": 13, "bundl": 13, "copi": [13, 41], "app_data_dir": 13, "ad": [13, 16, 17, 18, 23, 25, 26, 28, 31, 33, 34, 39, 41, 43, 44, 45], "seed": 13, "bashactiv": 13, "cshellactiv": 13, "fishactiv": 13, "nushellactiv": 13, "powershellactiv": 13, "pythonactiv": 13, "bin": [13, 16, 17, 18, 23, 28], "determin": 13, "nvidia": [13, 23, 25, 26, 28], "smi": 13, "head": [13, 25, 41], "wed": [13, 16, 23, 25, 28], "jul": 13, "26": [13, 16, 17, 18, 25, 28, 37], "21": [13, 14, 16, 23, 25, 28, 36, 37], "49": [13, 16, 17, 28, 37, 39], "2023": [13, 16, 17, 18, 33], "510": 13, "47": [13, 16, 17, 18, 23, 28], "03": [13, 14, 17, 25, 28, 36, 37, 43], "driver": 13, "greater": 13, "our": [13, 16, 17, 18, 20, 21, 28, 29, 31, 41, 44, 45], "case": [13, 14, 16, 17, 18, 31, 33, 34, 43, 44, 45], "verifi": 13, "nvcc": 13, "copyright": 13, "c": [13, 25, 26, 31, 33, 34, 39, 43, 44, 45], "2005": 13, "2019": 13, "corpor": 13, "built": 13, "wed_oct_23_19": 13, "38_pdt_2019": 13, "releas": [13, 14, 16, 17, 18, 23, 25, 28], "v10": 13, "89": [13, 23], "cu116": 13, "compat": 13, "audio": [13, 36, 37], "stabl": 13, "matrix": 13, "appropri": 13, "2bcu116": 13, "cp38": 13, "linux_x86_64": 13, "1983": 13, "mb": [13, 16, 17, 18], "________________________________________": 13, "gb": [13, 25], "764": 13, "kb": [13, 16, 17, 18, 36, 37], "eta": 13, "00": [13, 16, 23, 25, 26, 28, 32, 36, 37, 39], "satisfi": 13, "extens": 13, "successfulli": [13, 16, 17, 18], "__version__": 13, "dev20230725": 13, "pypi": 13, "tuna": 13, "tsinghua": 13, "edu": 13, "cn": 13, "csukuangfj": [13, 14, 16, 17, 19, 23, 25, 26, 28, 32, 36, 37, 39, 43], "resolv": 13, "main": [13, 23, 28, 41], "ubuntu": [13, 16, 17, 18], "2bcuda11": 13, "manylinux_2_17_x86_64": 13, "manylinux2014_x86_64": 13, "graphviz": 13, "cach": [13, 18], "de": [13, 14, 16, 17, 18, 25], "5e": 13, "fcbb22c68208d39edff467809d06c9d81d7d27426460ebc598e55130c1aa": 13, "20": [13, 14, 16, 18, 23, 25, 26, 28, 31, 32, 36, 37, 39, 44], "none": [13, 23, 28], "sha1": [13, 14, 16, 17, 18, 23, 25, 28], "4c05309499a08454997adf500b56dcc629e35ae5": 13, "date": [13, 14, 16, 17, 18, 23, 25, 28], "tue": [13, 16, 28], "25": [13, 14, 16, 17, 23, 28, 31, 36, 37, 39, 44], "16": [13, 14, 16, 17, 18, 21, 23, 25, 26, 28, 31, 32, 36, 37, 39, 43, 44, 45], "36": [13, 16, 25, 28, 29], "o": 13, "cento": 13, "2009": 13, "core": 13, "cmake": [13, 16, 17, 23, 28], "27": [13, 16, 17, 18, 23, 25, 32, 37], "gcc": 13, "cmake_cuda_flag": 13, "wno": 13, "deprec": [13, 25], "lineinfo": 13, "expt": 13, "extend": 13, "lambda": 13, "use_fast_math": 13, "xptxa": 13, "w": [13, 28, 36, 37], "gencod": 13, "arch": 13, "compute_35": 13, "sm_35": 13, "compute_50": 13, "sm_50": 13, "compute_60": 13, "sm_60": 13, "compute_61": 13, "sm_61": 13, "compute_70": 13, "sm_70": 13, "compute_75": 13, "sm_75": 13, "compute_80": 13, "sm_80": 13, "compute_86": 13, "sm_86": 13, "donnx_namespac": 13, "onnx_c2": 13, "compute_52": 13, "sm_52": 13, "xcudaf": 13, "diag_suppress": 13, "cc_clobber_ignor": 13, "integer_sign_chang": 13, "useless_using_declar": 13, "set_but_not_us": 13, "field_without_dll_interfac": 13, "base_class_has_different_dll_interfac": 13, "dll_interface_conflict_none_assum": 13, "dll_interface_conflict_dllexport_assum": 13, "implicit_return_from_non_void_funct": 13, "unsigned_compare_with_zero": 13, "declared_but_not_referenc": 13, "bad_friend_decl": 13, "relax": 13, "constexpr": 13, "d_glibcxx_use_cxx11_abi": 13, "option": [13, 15, 19, 22, 25, 29, 32, 36, 37, 39], "wall": 13, "strict": [13, 18, 24], "overflow": 13, "unknown": 13, "pragma": 13, "cmake_cxx_flag": 13, "unus": 13, "nvtx": 13, "enabl": [13, 29], "disabl": [13, 14, 16, 17], "debug": 13, "sync": 13, "kernel": [13, 16, 18, 25], "cpu": [13, 14, 16, 17, 18, 20, 23, 31, 33, 34, 39, 44, 45], "memori": [13, 16, 23, 25, 28, 41], "alloc": 13, "214748364800": 13, "byte": [13, 16, 17, 18], "200": [13, 14, 16, 17, 18, 23, 28, 29, 36, 37, 39], "abort": 13, "__file__": 13, "cpython": [13, 16], "gnu": [13, 16], "req": 13, "vq12fd5i": 13, "filter": 13, "quiet": [13, 24], "7640d663469b22cd0b36f3246ee9b849cd25e3b7": 13, "metadata": [13, 36, 37], "pyproject": 13, "toml": 13, "cytoolz": 13, "1e": 13, "3b": 13, "a7828d575aa17fb7acaf1ced49a3655aa36dad7e16eb7e6a2e4df0dda76f": 13, "33": [13, 16, 17, 23, 24, 25, 28, 36], "pyyaml": 13, "c8": 13, "6b": 13, "6600ac24725c7388255b2f5add93f91e58a5d7efaf4af244fdbcc11a541b": 13, "ma": 13, "nylinux_2_17_x86_64": 13, "736": 13, "dataclass": 13, "2f": 13, "1095cdc2868052dd1e64520f7c0d5c8c550ad297e944e641dbf1ffbb9a5d": 13, "dev0": 13, "7640d66": 13, "lilcom": 13, "a8": 13, "df0a69c52bd085ca1ad4e5c4c1a5c680e25f9477d8e49316c4ff1e5084a4": 13, "linux_2_17_x86_64": 13, "87": [13, 16], "tqdm": 13, "e6": 13, "02": [13, 14, 16, 17, 18, 25, 28, 31, 37, 43, 44], "a2cff6306177ae6bc73bc0665065de51dfb3b9db7373e122e2735faf0d97": 13, "numpi": 13, "18": [13, 16, 17, 18, 23, 25, 26, 28, 31, 32, 36, 37, 43, 44, 45], "audioread": 13, "5d": 13, "cb": 13, "82a002441902dccbe427406785db07af10182245ee639ea9f4d92907c923": 13, "tar": 13, "gz": 13, "377": 13, "tabul": 13, "40": [13, 16, 17, 18, 26, 28, 32, 36, 37], "44": [13, 16, 17, 28, 36, 37], "4a5f08c96eb108af5cb50b41f76142f0afa346dfa99d5296fe7202a11854": 13, "35": [13, 14, 16, 17, 18, 25, 28, 43], "1a": 13, "70": 13, "e63223f8116931d365993d4a6b7ef653a4d920b41d03de7c59499962821f": 13, "97": [13, 16, 23], "ab": [13, 31, 43, 44, 45], "c3": 13, "57f0601a2d4fe15de7a553c00adbc901425661bf048f2a22dfc500caf121": 13, "48": [13, 16, 17, 23, 25], "intervaltre": 13, "50": [13, 14, 16, 17, 18, 28, 31, 36, 43, 44, 45], "fb": 13, "396d568039d21344639db96d940d40eb62befe704ef849b27949ded5c3bb": 13, "soundfil": 13, "bd": 13, "0602167a213d9184fc688b1086dc6d374b7ae8c33eccf169f9b50ce6568c": 13, "py2": 13, "46": [13, 17, 23, 28], "toolz": 13, "7f": 13, "5c": 13, "922a3508f5bda2892be3df86c74f9cf1e01217c2b1f8a0ac4841d903e3e9": 13, "55": [13, 16, 26, 28, 36], "sortedcontain": 13, "9cb0e58b2deb7f82b84065f37f3bffeb12413f947f9388e4cac22c4621c": 13, "cffi": 13, "b7": 13, "8b": 13, "06f30caa03b5b3ac006de4f93478dbd0239e2a16566d81a106c322dc4f79": 13, "15": [13, 14, 16, 17, 18, 25, 26, 28, 36, 39], "442": 13, "pycpars": 13, "d5": 13, "5f610ebe421e85889f2e55e33b7f9a6795bd982198517d912eb1c76e1a53": 13, "118": [13, 28], "filenam": [13, 16, 17, 18, 19, 20, 21, 33, 34, 43, 45], "size": [13, 14, 16, 17, 18, 23, 25, 26, 28, 29, 31, 32, 33, 34, 36, 37, 39, 43, 44, 45], "687627": 13, "sha256": 13, "cbf0a4d2d0b639b33b91637a4175bc251d6a021a069644ecb1a9f2b3a83d072a": 13, "store": [13, 28], "ephem": 13, "wwtk90_m": 13, "7a": 13, "8e": 13, "a0bf241336e2e3cb573e1e21e5600952d49f5162454f2e612f": 13, "23704": 13, "5e2d3537c96ce9cf0f645a654c671163707bf8cb8d9e358d0e2b0939a85ff4c2": 13, "9c": 13, "f19ae5a03f8862d9f0776b0c0570f1fdd60a119d90954e3f39": 13, "26098": 13, "2604170976cfffe0d2f678cb1a6e5b525f561cd50babe53d631a186734fec9f9": 13, "f3": 13, "ed": 13, "2b": 13, "c179ebfad4e15452d6baef59737f27beb9bfb442e0620f7271": 13, "remot": 13, "enumer": 13, "12942": 13, "count": 13, "100": [13, 23, 25, 26, 28, 29, 31, 33, 34, 43, 44, 45], "compress": 13, "56": [13, 16, 17, 28, 36], "total": [13, 17, 18, 23, 25, 26, 28, 29, 31, 32, 39, 43, 44], "delta": 13, "17": [13, 14, 16, 17, 18, 23, 28, 36, 37, 43], "reus": 13, "pack": [13, 44, 45], "12875": 13, "receiv": 13, "mib": 13, "8835": 13, "07": [13, 16, 17, 18, 23, 25, 26, 28], "41": [13, 16, 18, 23, 25, 36, 39], "dl_dir": [13, 23, 26, 28, 29, 31, 33, 34, 43, 44, 45], "waves_yesno": 13, "___________________________________________________": 13, "70m": 13, "1mb": 13, "manifest": [13, 29], "fbank": [13, 14, 16, 17, 18, 23, 25, 26, 28, 32, 36, 37, 39], "718": 13, "info": [13, 14, 16, 17, 18, 23, 25, 26, 28, 32, 36, 37, 39], "compute_fbank_yesno": 13, "process": [13, 14, 16, 17, 23, 25, 26, 28, 31, 33, 34, 43, 44, 45], "extract": [13, 23, 25, 26, 28, 29, 31, 32, 33, 34, 36, 37, 39, 43, 44, 45], "featur": [13, 23, 25, 26, 28, 29, 31, 32, 33, 34, 36, 37, 39, 41, 43, 44, 45], "_______________________________________________________________________________": 13, "90": [13, 16], "01": [13, 16, 25, 26, 28, 29, 33], "82it": 13, "778": 13, "______________________________________________________________________________": 13, "256": [13, 18, 36, 37], "92it": 13, "51": [13, 16, 23, 28, 39], "lang": [13, 14, 25, 28, 34], "66": [13, 17], "project": 13, "kaldilm": 13, "csrc": [13, 28], "arpa_file_pars": 13, "cc": 13, "void": 13, "arpafilepars": 13, "std": 13, "istream": 13, "79": 13, "140": [13, 26], "92": [13, 28], "hlg": [13, 32, 36, 37, 39], "275": [13, 23], "compile_hlg": 13, "124": [13, 23, 28], "lang_phon": [13, 26, 32, 36, 37, 39], "276": 13, "lexicon": [13, 23, 25, 26, 28, 29, 31, 33, 34, 39, 43, 44, 45], "171": [13, 26, 28, 36, 37], "convert": [13, 16, 17, 18, 28], "l": [13, 16, 17, 18, 25, 36, 37, 39], "linv": [13, 25, 28, 39], "309": 13, "ctc_topo": 13, "max_token_id": 13, "310": 13, "314": 13, "intersect": [13, 31, 44, 45], "323": 13, "lg": [13, 31, 34, 44, 45], "shape": [13, 18], "connect": [13, 14, 28, 31, 32, 43, 44, 45], "68": [13, 28], "class": [13, 28], "tensor": [13, 17, 18, 23, 25, 26, 28, 31, 39, 43, 44], "71": [13, 28, 32], "341": 13, "rag": 13, "raggedtensor": 13, "76": [13, 39], "remov": [13, 23, 25, 26, 28, 32, 36, 37], "disambigu": 13, "354": 13, "91": 13, "remove_epsilon": 13, "445": 13, "arc": 13, "compos": 13, "h": 13, "446": 13, "106": [13, 17, 28], "109": [13, 23, 28], "447": 13, "111": [13, 28], "127": [13, 16, 17, 39], "cuda_visible_devic": [13, 23, 25, 26, 28, 29, 31, 32, 33, 34, 36, 37, 39, 43, 44, 45], "segment": 13, "fault": 13, "dump": 13, "protocol_buffers_python_implement": 13, "674": 13, "interest": [13, 29, 31, 33, 34, 43, 44, 45], "936": 13, "481": 13, "482": 13, "posixpath": [13, 16, 17, 18, 25, 28], "lang_dir": [13, 25, 28], "lr": [13, 25, 43], "feature_dim": [13, 14, 16, 17, 18, 23, 25, 28, 39], "weight_decai": 13, "06": [13, 14, 16, 26, 28, 32, 39], "start_epoch": 13, "best_train_loss": [13, 14, 16, 17, 18], "inf": [13, 14, 16, 17, 18], "best_valid_loss": [13, 14, 16, 17, 18], "best_train_epoch": [13, 14, 16, 17, 18], "best_valid_epoch": [13, 14, 17, 18], "batch_idx_train": [13, 14, 16, 17, 18], "log_interv": [13, 14, 16, 17, 18], "reset_interv": [13, 14, 16, 17, 18], "valid_interv": [13, 14, 16, 17, 18], "beam_siz": [13, 14, 25], "sum": 13, "use_double_scor": [13, 23, 28, 39], "world_siz": [13, 29], "master_port": 13, "12354": 13, "num_epoch": 13, "feature_dir": [13, 28], "max_dur": [13, 28], "bucketing_sampl": [13, 28], "num_bucket": [13, 28], "concatenate_cut": [13, 28], "duration_factor": [13, 28], "gap": [13, 28], "on_the_fly_feat": [13, 28], "shuffl": [13, 28], "return_cut": [13, 28], "num_work": [13, 28], "env_info": [13, 14, 16, 17, 18, 23, 25, 28], "branch": [13, 14, 16, 17, 18, 23, 25, 28, 33], "3fb0a43": 13, "thu": [13, 14, 16, 17, 18, 25, 28, 32], "05": [13, 14, 16, 17, 23, 25, 26, 28, 37], "hostnam": [13, 14, 16, 17, 18, 25], "74279": [13, 14, 16, 17, 18, 25], "1220091118": 13, "57c4d55446": 13, "sph26": 13, "ip": [13, 14, 16, 17, 18, 25], "941": 13, "168": [13, 32], "949": 13, "495": 13, "devic": [13, 14, 16, 17, 18, 23, 25, 26, 28, 31, 32, 33, 34, 36, 37, 39, 44, 45], "965": [13, 23], "146": 13, "cut": [13, 28], "244": 13, "967": 13, "149": [13, 16, 28], "199": [13, 28, 32], "singlecutsampl": 13, "205": [13, 28], "968": 13, "218": [13, 17], "252": 13, "565": [13, 28], "422": 13, "batch": [13, 16, 17, 18, 23, 25, 26, 28, 31, 33, 34, 43, 44, 45], "loss": [13, 16, 17, 23, 26, 28, 29, 31, 32, 33, 34, 36, 37, 39, 43, 44, 45], "065": 13, "over": [13, 23, 25, 26, 28, 31, 33, 34, 43, 44, 45], "2436": 13, "frame": [13, 25, 31, 33, 44, 45], "tot_loss": 13, "53": [13, 18, 23, 31, 32, 37, 43, 44], "681": [13, 16], "4561": 13, "2828": 13, "7076": 13, "22192": 13, "54": [13, 17, 18, 28, 32, 36, 37], "167": 13, "444": 13, "9002": 13, "18067": 13, "011": 13, "2555": 13, "2695": 13, "484": 13, "34971": 13, "331": [13, 16, 17, 28, 32], "4688": 13, "368": 13, "75": [13, 16], "633": 13, "2532": 13, "242": [13, 23, 28], "1139": 13, "1592": 13, "522": [13, 28], "1627": 13, "209": [13, 32], "07055": 13, "1175": 13, "07091": 13, "640": [13, 18], "847": 13, "07731": 13, "427": [13, 17, 28], "04391": 13, "05341": 13, "884": 13, "04384": 13, "387": [13, 37], "03458": 13, "04616": 13, "707": [13, 23, 28], "03379": 13, "758": [13, 28], "433": [13, 28], "01054": 13, "980": [13, 28], "009014": 13, "009974": 13, "489": [13, 23], "01085": 13, "258": [13, 36, 37], "01172": 13, "01055": 13, "621": [13, 39], "01074": 13, "699": 13, "866": 13, "01044": 13, "844": 13, "008942": 13, "221": [13, 28], "01082": 13, "970": [13, 28], "01169": 13, "28": [13, 16, 17, 25, 28, 32], "247": 13, "01073": 13, "326": [13, 17], "555": 13, "840": 13, "264": [13, 18], "search_beam": [13, 23, 28, 39], "output_beam": [13, 23, 28, 39], "min_active_st": [13, 23, 28, 39], "max_active_st": [13, 23, 28, 39], "10000": [13, 23, 28, 39], "841": 13, "855": 13, "273": [13, 14, 25], "868": 13, "291": 13, "882": 13, "883": 13, "157": 13, "204": [13, 18, 28], "until": [13, 28, 33], "701": 13, "241": [13, 23], "transcript": [13, 23, 24, 25, 26, 28, 31, 32, 36, 37, 43, 44, 45], "recog": [13, 25, 28], "test_set": [13, 39], "702": [13, 28], "564": 13, "240": [13, 23, 39], "ins": [13, 28, 39], "del": [13, 28, 39], "sub": [13, 28, 39], "704": [13, 23, 36], "249": [13, 17], "wrote": [13, 28], "detail": [13, 15, 19, 23, 25, 26, 28, 29, 31, 32, 33, 34, 36, 37, 39, 41, 43, 44, 45], "stat": [13, 28], "err": [13, 25, 28], "316": [13, 28], "congratul": [13, 16, 17, 18, 23, 26, 28, 32, 36, 37, 39], "fun": [13, 16, 17], "variou": [13, 19, 22, 46], "period": [14, 16], "disk": 14, "optim": [14, 23, 25, 26, 28, 31, 32, 33, 34, 36, 37, 39, 43, 44, 45], "relat": [14, 23, 25, 28, 32, 36, 37, 39], "resum": [14, 23, 25, 26, 28, 31, 32, 33, 34, 36, 37, 39, 43, 44, 45], "strip": 14, "reduc": [14, 23, 25, 26, 28, 31, 32, 33, 34, 36, 37, 39, 43, 44, 45], "each": [14, 16, 17, 19, 23, 25, 26, 28, 31, 33, 34, 41, 43, 44, 45], "well": [14, 39, 46], "usag": [14, 16, 17, 18, 20, 21, 32, 36, 37, 39], "pruned_transducer_stateless3": [14, 20, 41], "almost": [14, 31, 41, 44, 45], "dict": [14, 18], "host": 14, "stateless3": [14, 16], "repo": [14, 19], "prefix": 14, "those": 14, "wave": [14, 16, 17, 18, 23, 28], "iter": [14, 16, 17, 18, 21, 31, 33, 34, 43, 44, 45], "1224000": 14, "greedy_search": [14, 25, 31, 33, 43, 44, 45], "test_wav": [14, 16, 17, 18, 19, 23, 25, 26, 28, 32, 36, 37, 39], "1089": [14, 16, 17, 18, 19, 28, 32], "134686": [14, 16, 17, 18, 19, 28, 32], "0001": [14, 16, 17, 18, 19, 28, 32], "wav": [14, 16, 17, 18, 19, 21, 23, 25, 26, 28, 31, 33, 34, 36, 37, 39, 43, 44, 45], "1221": [14, 16, 17, 28, 32], "135766": [14, 16, 17, 28, 32], "0002": [14, 16, 17, 28, 32], "multipl": [14, 23, 25, 26, 28, 32, 36, 37, 39], "sound": [14, 16, 17, 18, 21, 22, 23, 25, 26, 28, 32, 36, 37, 39], "Its": [14, 16, 17, 18, 28], "output": [14, 16, 17, 18, 23, 25, 26, 28, 29, 31, 32, 33, 34, 36, 37, 39, 41, 43, 44, 45], "19": [14, 16, 17, 18, 23, 28, 32, 36, 37], "09": [14, 17, 23, 25, 26, 28, 43], "233": [14, 16, 17], "265": 14, "3000": [14, 16, 17, 18], "80": [14, 16, 17, 18, 23, 25, 28], "subsampling_factor": [14, 17, 18, 23, 25, 28], "encoder_dim": [14, 16, 17, 18], "512": [14, 16, 17, 18, 23, 25, 28], "nhead": [14, 16, 18, 23, 25, 28, 31, 44], "dim_feedforward": [14, 16, 17, 25], "num_encoder_lay": [14, 16, 17, 18, 25], "decoder_dim": [14, 16, 17, 18], "joiner_dim": [14, 16, 17, 18], "model_warm_step": [14, 16, 17], "4810e00d8738f1a21278b0156a42ff396a2d40ac": 14, "fri": 14, "oct": [14, 28], "miss": [14, 16, 17, 18, 25, 28], "cu102": [14, 16, 17, 18], "1013": 14, "c39cba5": 14, "dirti": [14, 16, 17, 23, 28], "ceph": [14, 23, 25, 28], "jsonl": 14, "0324160024": 14, "65bfd8b584": 14, "jjlbn": 14, "bpe_model": [14, 16, 17, 18, 28], "sound_fil": [14, 23, 25, 28, 39], "sample_r": [14, 23, 25, 28, 39], "16000": [14, 23, 25, 26, 28, 32, 33, 36, 37], "max_context": 14, "max_stat": 14, "context_s": [14, 16, 17, 18, 25], "max_sym_per_fram": [14, 25], "simulate_stream": 14, "decode_chunk_s": 14, "left_context": 14, "dynamic_chunk_train": 14, "causal_convolut": 14, "short_chunk_s": [14, 18, 44, 45], "num_left_chunk": [14, 18], "blank_id": [14, 16, 17, 18, 25], "unk_id": 14, "vocab_s": [14, 16, 17, 18, 25], "271": [14, 17], "612": 14, "458": 14, "giga": [14, 17, 43], "623": 14, "277": 14, "78648040": 14, "951": [14, 28], "285": [14, 25, 28], "construct": [14, 16, 17, 18, 23, 25, 26, 28, 32, 36, 37, 39], "952": 14, "295": [14, 23, 25, 26, 28], "957": 14, "301": [14, 28], "700": 14, "329": [14, 17, 28], "912": 14, "388": 14, "earli": [14, 16, 17, 18, 28, 32], "nightfal": [14, 16, 17, 18, 28, 32], "THE": [14, 16, 17, 18, 28, 32], "yellow": [14, 16, 17, 18, 28, 32], "lamp": [14, 16, 17, 18, 28, 32], "light": [14, 16, 17, 18, 28, 32], "AND": [14, 16, 17, 18, 28, 32], "THERE": [14, 16, 17, 18, 28, 32], "squalid": [14, 16, 17, 18, 28, 32], "quarter": [14, 16, 17, 18, 28, 32], "OF": [14, 16, 17, 18, 28, 32], "brothel": [14, 16, 17, 18, 28, 32], "god": [14, 28, 32], "AS": [14, 28, 32], "direct": [14, 28, 32], "consequ": [14, 28, 32], "sin": [14, 28, 32], "man": [14, 28, 32], "punish": [14, 28, 32], "had": [14, 28, 32], "her": [14, 28, 32], "love": [14, 28, 32], "child": [14, 28, 32], "whose": [14, 25, 28, 32], "ON": [14, 16, 28, 32], "THAT": [14, 28, 32], "dishonor": [14, 28, 32], "bosom": [14, 28, 32], "TO": [14, 28, 32], "parent": [14, 28, 32], "forev": [14, 28, 32], "WITH": [14, 28, 32], "race": [14, 28, 32], "descent": [14, 28, 32], "mortal": [14, 28, 32], "BE": [14, 28, 32], "bless": [14, 28, 32], "soul": [14, 28, 32], "IN": [14, 28, 32], "heaven": [14, 28, 32], "yet": [14, 16, 17, 28, 32], "THESE": [14, 28, 32], "thought": [14, 28, 32], "affect": [14, 28, 32], "hester": [14, 28, 32], "prynn": [14, 28, 32], "hope": [14, 24, 28, 32], "apprehens": [14, 28, 32], "390": 14, "down": [14, 23, 28, 31, 33, 34, 43, 44, 45], "reproduc": [14, 28], "9999": [14, 33, 34, 43], "symlink": 14, "pass": [14, 18, 23, 25, 26, 28, 31, 33, 34, 41, 43, 44, 45], "reason": [14, 16, 17, 18, 44], "support": [15, 16, 17, 18, 23, 25, 28, 31, 33, 34, 41, 43, 44, 45], "zipform": [15, 19, 22, 27, 30, 40, 42], "convemform": [15, 22, 41], "platform": [15, 19], "android": [15, 16, 17, 18, 19], "raspberri": [15, 19], "pi": [15, 19], "\u7231\u82af\u6d3e": 15, "maix": 15, "iii": 15, "axera": 15, "rv1126": 15, "static": 15, "produc": [15, 31, 33, 34, 43, 44, 45], "binari": [15, 16, 17, 18, 23, 25, 26, 28, 31, 39, 43, 44], "everyth": 15, "pnnx": [15, 22], "torchscript": [15, 20, 21, 22], "encod": [15, 19, 21, 22, 23, 25, 26, 28, 31, 32, 33, 39, 41, 43, 44, 45], "int8": [15, 22], "quantiz": [15, 22, 29], "conv": [16, 17], "emform": [16, 17, 20], "stateless2": [16, 17, 43], "cpp": [16, 20], "pretrained_model": [16, 17, 18], "online_transduc": 16, "continu": [16, 17, 18, 19, 23, 25, 26, 28, 31, 33, 34, 39, 43, 44], "jit_xxx": [16, 17, 18], "anywher": [16, 17], "submodul": 16, "updat": [16, 17, 18], "recurs": 16, "init": 16, "dcmake_build_typ": [16, 23, 28], "dncnn_python": 16, "dncnn_build_benchmark": 16, "dncnn_build_exampl": 16, "dncnn_build_tool": 16, "j4": 16, "pwd": 16, "src": [16, 18], "compon": [16, 41], "ncnn2int8": [16, 17], "am": 16, "sai": [16, 17, 18, 23, 25, 26, 28, 29, 31, 32, 33, 34, 36, 37, 39, 43, 44, 45], "later": [16, 17, 18, 23, 26, 28, 31, 32, 33, 34, 36, 37, 43, 44, 45], "termin": 16, "tencent": [16, 17], "modif": [16, 25], "offic": 16, "synchron": 16, "offici": 16, "renam": [16, 17, 18], "conv_emformer_transducer_stateless2": [16, 41], "length": [16, 18, 25, 44, 45], "cnn": [16, 18], "31": [16, 17, 18, 28], "context": [16, 25, 31, 41, 43, 44, 45], "configur": [16, 18, 25, 29, 32, 36, 37, 39], "accordingli": [16, 17, 18], "yourself": [16, 17, 18, 29, 44, 45], "combin": [16, 17, 18], "677": 16, "220": [16, 25, 26, 28], "229": [16, 23], "best_v": 16, "alid_epoch": 16, "subsampl": [16, 44, 45], "ing_factor": 16, "a34171ed85605b0926eebbd0463d059431f4f74a": 16, "dec": 16, "ver": 16, "ion": 16, "530e8a1": 16, "op": 16, "1220120619": [16, 17, 18], "7695ff496b": [16, 17, 18], "s9n4w": [16, 17, 18], "icefa": 16, "ll": 16, "transdu": 16, "cer": 16, "use_averaged_model": [16, 17, 18], "cnn_module_kernel": [16, 18], "left_context_length": 16, "chunk_length": 16, "right_context_length": 16, "memory_s": 16, "231": [16, 17, 18], "053": 16, "112": [16, 17, 18], "022": 16, "708": [16, 23, 25, 28, 39], "315": [16, 23, 25, 26, 28, 32], "75490012": 16, "318": [16, 17], "320": [16, 25], "682": 16, "lh": [16, 17, 18], "rw": [16, 17, 18], "root": [16, 17, 18], "289m": 16, "jan": [16, 17, 18], "289": 16, "roughli": [16, 17, 18], "equal": [16, 17, 18, 44, 45], "1024": [16, 17, 18, 43], "287": [16, 39], "1010k": [16, 17], "decoder_jit_trac": [16, 17, 18, 21, 43, 45], "283m": 16, "encoder_jit_trac": [16, 17, 18, 21, 43, 45], "0m": [16, 17], "joiner_jit_trac": [16, 17, 18, 21, 43, 45], "sure": [16, 17, 18], "could": [16, 17, 18, 23, 26], "found": [16, 17, 18, 23, 25, 26, 28, 31, 33, 34, 39, 43, 44], "param": [16, 17, 18], "503k": [16, 17], "437": [16, 17, 18], "142m": 16, "79k": 16, "5m": [16, 17], "architectur": [16, 17, 18, 43], "editor": [16, 17, 18], "content": [16, 17, 18], "283": [16, 18], "1010": [16, 17], "142": [16, 23, 26, 28], "503": [16, 17], "convers": [16, 17, 18], "half": [16, 17, 18, 31, 44, 45], "v": [16, 17, 18, 28, 36, 37], "default": [16, 17, 18, 23, 25, 26, 28, 29, 31, 32, 33, 34, 36, 37, 39, 43, 44, 45], "float32": [16, 17, 18], "float16": [16, 17, 18], "occupi": [16, 17, 18], "twice": [16, 17, 18], "smaller": [16, 17, 18, 23, 25, 26, 28, 31, 33, 34, 43, 44, 45], "fp16": [16, 17, 18, 31, 33, 34, 43, 44, 45], "won": [16, 17, 18, 19, 23, 26, 28, 29, 31, 33, 34, 43, 44, 45], "accept": [16, 17, 18], "216": [16, 23, 28, 36, 37], "encoder_param_filenam": [16, 17, 18], "encoder_bin_filenam": [16, 17, 18], "decoder_param_filenam": [16, 17, 18], "decoder_bin_filenam": [16, 17, 18], "joiner_param_filenam": [16, 17, 18], "joiner_bin_filenam": [16, 17, 18], "sound_filenam": [16, 17, 18], "141": 16, "328": 16, "151": 16, "176": [16, 25, 28], "336": 16, "106000": [16, 17, 18, 28, 32], "581": [16, 32], "381": 16, "7767517": [16, 17, 18], "1060": 16, "1342": 16, "in0": [16, 17, 18], "explan": [16, 17, 18], "magic": [16, 17, 18], "intermedi": [16, 17, 18], "mean": [16, 17, 18, 23, 25, 26, 28, 31, 32, 33, 34, 36, 37, 39, 41, 43, 44, 45], "increment": [16, 17, 18], "1061": 16, "sherpametadata": [16, 17, 18], "sherpa_meta_data1": [16, 17, 18], "still": [16, 17, 18], "newli": [16, 17, 18], "must": [16, 17, 18, 44], "kei": [16, 17, 18, 28], "eas": [16, 17, 18], "list": [16, 17, 18, 23, 25, 26, 28, 32, 36, 37], "pair": [16, 17, 18], "sad": [16, 17, 18], "rememb": [16, 17, 18], "anymor": [16, 17, 18], "flexibl": [16, 17, 18], "edit": [16, 17, 18], "arm": [16, 17, 18], "aarch64": [16, 17, 18], "onc": [16, 17], "mayb": [16, 17], "year": [16, 17], "_jit_trac": [16, 17], "fp32": [16, 17], "doubl": [16, 17], "j": [16, 17, 23, 28], "py38": [16, 17, 18], "arg": [16, 17], "wave_filenam": [16, 17], "16k": [16, 17], "hz": [16, 17, 36, 37], "mono": [16, 17], "calibr": [16, 17], "cat": [16, 17], "eof": [16, 17], "calcul": [16, 17, 33, 44, 45], "has_gpu": [16, 17], "config": [16, 17], "use_vulkan_comput": [16, 17], "88": [16, 25], "conv_87": 16, "942385": [16, 17], "threshold": [16, 17, 33], "938493": 16, "968131": 16, "conv_88": 16, "442448": 16, "549335": 16, "167552": 16, "conv_89": 16, "228289": 16, "001738": 16, "871552": 16, "linear_90": 16, "976146": 16, "101789": 16, "115": [16, 17, 23, 28], "267128": 16, "linear_91": 16, "962030": 16, "162033": 16, "602713": 16, "linear_92": 16, "323041": 16, "853959": 16, "953129": 16, "linear_94": 16, "905416": 16, "648006": 16, "323545": 16, "linear_93": 16, "474093": 16, "200188": 16, "linear_95": 16, "888012": 16, "403563": 16, "483986": 16, "linear_96": 16, "856741": 16, "398679": 16, "524273": 16, "linear_97": 16, "635942": 16, "613655": 16, "590950": 16, "linear_98": 16, "460340": 16, "670146": 16, "398010": 16, "linear_99": 16, "532276": 16, "585537": 16, "119396": 16, "linear_101": 16, "585871": 16, "719224": 16, "205809": 16, "linear_100": 16, "751382": 16, "081648": 16, "linear_102": 16, "593344": 16, "450581": 16, "551147": 16, "linear_103": 16, "592681": 16, "705824": 16, "257959": 16, "linear_104": 16, "752957": 16, "980955": 16, "110489": 16, "linear_105": 16, "696240": 16, "877193": 16, "608953": 16, "linear_106": 16, "059659": 16, "643138": 16, "048950": 16, "linear_108": 16, "975461": 16, "589567": 16, "671457": 16, "linear_107": 16, "190381": 16, "515701": 16, "linear_109": 16, "710759": 16, "305635": 16, "082436": 16, "linear_110": 16, "531228": 16, "731162": 16, "159557": 16, "linear_111": 16, "528083": 16, "259322": 16, "211544": 16, "linear_112": 16, "148807": 16, "500842": 16, "087374": 16, "linear_113": 16, "592566": 16, "948851": 16, "166611": 16, "linear_115": 16, "437109": 16, "608947": 16, "642395": 16, "linear_114": 16, "193942": 16, "503904": 16, "linear_116": 16, "966980": 16, "200896": 16, "676392": 16, "linear_117": 16, "451303": 16, "061664": 16, "951344": 16, "linear_118": 16, "077262": 16, "965800": 16, "023804": 16, "linear_119": 16, "671615": 16, "847613": 16, "198460": 16, "linear_120": 16, "625638": 16, "131427": 16, "556595": 16, "linear_122": 16, "274080": 16, "888716": 16, "978189": 16, "linear_121": 16, "420480": 16, "429659": 16, "linear_123": 16, "826197": 16, "599617": 16, "281532": 16, "linear_124": 16, "396383": 16, "325849": 16, "335875": 16, "linear_125": 16, "337198": 16, "941410": 16, "221970": 16, "linear_126": 16, "699965": 16, "842878": 16, "224073": 16, "linear_127": 16, "775370": 16, "884215": 16, "696438": 16, "linear_129": 16, "872276": 16, "837319": 16, "254213": 16, "linear_128": 16, "180057": 16, "687883": 16, "linear_130": 16, "150427": 16, "454298": 16, "765789": 16, "linear_131": 16, "112692": 16, "924847": 16, "025545": 16, "linear_132": 16, "852893": 16, "116593": 16, "749626": 16, "linear_133": 16, "517084": 16, "024665": 16, "275314": 16, "linear_134": 16, "683807": 16, "878618": 16, "743618": 16, "linear_136": 16, "421055": 16, "322729": 16, "086264": 16, "linear_135": 16, "309880": 16, "917679": 16, "linear_137": 16, "827781": 16, "744595": 16, "915554": 16, "linear_138": 16, "422395": 16, "742882": 16, "402161": 16, "linear_139": 16, "527538": 16, "866123": 16, "849449": 16, "linear_140": 16, "128619": 16, "657793": 16, "266134": 16, "linear_141": 16, "839593": 16, "845993": 16, "021378": 16, "linear_143": 16, "442304": 16, "099039": 16, "889746": 16, "linear_142": 16, "325038": 16, "849592": 16, "linear_144": 16, "929444": 16, "618206": 16, "605080": 16, "linear_145": 16, "382126": 16, "321095": 16, "625010": 16, "linear_146": 16, "894987": 16, "867645": 16, "836517": 16, "linear_147": 16, "915313": 16, "906028": 16, "886522": 16, "linear_148": 16, "614287": 16, "908151": 16, "496181": 16, "linear_150": 16, "724932": 16, "485588": 16, "312899": 16, "linear_149": 16, "161146": 16, "606939": 16, "linear_151": 16, "164453": 16, "847355": 16, "719223": 16, "linear_152": 16, "086471": 16, "984121": 16, "222834": 16, "linear_153": 16, "099524": 16, "991601": 16, "816805": 16, "linear_154": 16, "054585": 16, "489706": 16, "286930": 16, "linear_155": 16, "389185": 16, "100321": 16, "963501": 16, "linear_157": 16, "982999": 16, "154796": 16, "637253": 16, "linear_156": 16, "537706": 16, "875190": 16, "linear_158": 16, "420287": 16, "502287": 16, "531588": 16, "linear_159": 16, "014746": 16, "423280": 16, "477261": 16, "linear_160": 16, "633553": 16, "715335": 16, "220921": 16, "linear_161": 16, "371849": 16, "117830": 16, "815203": 16, "linear_162": 16, "492933": 16, "126283": 16, "623318": 16, "linear_164": 16, "697504": 16, "825712": 16, "317358": 16, "linear_163": 16, "078367": 16, "008038": 16, "linear_165": 16, "023975": 16, "836278": 16, "577358": 16, "linear_166": 16, "860619": 16, "259792": 16, "493614": 16, "linear_167": 16, "380934": 16, "496160": 16, "107042": 16, "linear_168": 16, "691216": 16, "733317": 16, "831076": 16, "linear_169": 16, "723948": 16, "952728": 16, "129707": 16, "linear_171": 16, "034811": 16, "366547": 16, "665123": 16, "linear_170": 16, "356277": 16, "710501": 16, "linear_172": 16, "556884": 16, "729481": 16, "166058": 16, "linear_173": 16, "033039": 16, "207264": 16, "442120": 16, "linear_174": 16, "597379": 16, "658676": 16, "768131": 16, "linear_2": [16, 17], "293503": 16, "305265": 16, "877850": 16, "linear_1": [16, 17], "812222": 16, "766452": 16, "487047": 16, "linear_3": [16, 17], "999999": 16, "999755": 16, "031174": 16, "wish": [16, 17], "955k": 16, "18k": 16, "inparam": [16, 17], "inbin": [16, 17], "outparam": [16, 17], "outbin": [16, 17], "99m": 16, "78k": 16, "774k": [16, 17], "496": [16, 17, 28, 32], "replac": [16, 17], "774": [16, 17], "linear": [16, 17, 25], "convolut": [16, 17, 33, 41, 44], "exact": [16, 17], "4x": [16, 17], "comparison": 16, "468000": [17, 21, 43], "lstm_transducer_stateless2": [17, 21, 43], "862": 17, "222": [17, 26, 28], "865": 17, "is_pnnx": 17, "62e404dd3f3a811d73e424199b3408e309c06e1a": [17, 18], "mon": [17, 18], "6d7a559": [17, 18], "feb": [17, 18, 25], "147": [17, 18], "rnn_hidden_s": 17, "aux_layer_period": 17, "235": 17, "239": [17, 25], "472": 17, "595": 17, "324": 17, "83137520": 17, "596": 17, "325": 17, "257024": 17, "781812": 17, "327": 17, "84176356": 17, "182": [17, 18, 23, 32], "158": 17, "183": [17, 36, 37], "335": 17, "101": 17, "tracerwarn": [17, 18], "boolean": [17, 18], "caus": [17, 18, 23, 25, 26, 28, 31, 33, 34, 43, 44, 45], "incorrect": [17, 18, 25], "flow": [17, 18], "constant": [17, 18], "futur": [17, 18, 25, 46], "need_pad": 17, "bool": 17, "259": [17, 23], "180": [17, 23, 28], "339": 17, "304": 17, "207": [17, 26, 28], "84": [17, 23], "324m": 17, "321": [17, 23], "107": [17, 32], "318m": 17, "159m": 17, "21k": 17, "159": [17, 28, 39], "37": [17, 23, 25, 28, 36], "861": 17, "425": [17, 28], "266": [17, 18, 28, 32], "431": 17, "342": 17, "343": 17, "267": [17, 25, 36, 37], "379": 17, "268": [17, 28, 32], "317m": 17, "317": 17, "conv_15": 17, "930708": 17, "972025": 17, "conv_16": 17, "978855": 17, "031788": 17, "456645": 17, "conv_17": 17, "868437": 17, "830528": 17, "218575": 17, "linear_18": 17, "107259": 17, "194808": 17, "293236": 17, "linear_19": 17, "193777": 17, "634748": 17, "401705": 17, "linear_20": 17, "259933": 17, "606617": 17, "722160": 17, "linear_21": 17, "186600": 17, "790260": 17, "512129": 17, "linear_22": 17, "759041": 17, "265832": 17, "050053": 17, "linear_23": 17, "931209": 17, "099090": 17, "979767": 17, "linear_24": 17, "324160": 17, "215561": 17, "321835": 17, "linear_25": 17, "800708": 17, "599352": 17, "284134": 17, "linear_26": 17, "492444": 17, "153369": 17, "274391": 17, "linear_27": 17, "660161": 17, "720994": 17, "674126": 17, "linear_28": 17, "415265": 17, "174434": 17, "007133": 17, "linear_29": 17, "038418": 17, "118534": 17, "724262": 17, "linear_30": 17, "072084": 17, "936867": 17, "259155": 17, "linear_31": 17, "342712": 17, "599489": 17, "282787": 17, "linear_32": 17, "340535": 17, "120308": 17, "701103": 17, "linear_33": 17, "846987": 17, "630030": 17, "985939": 17, "linear_34": 17, "686298": 17, "204571": 17, "607586": 17, "linear_35": 17, "904821": 17, "575518": 17, "756420": 17, "linear_36": 17, "806659": 17, "585589": 17, "118401": 17, "linear_37": 17, "402340": 17, "047157": 17, "162680": 17, "linear_38": 17, "174589": 17, "923361": 17, "030258": 17, "linear_39": 17, "178576": 17, "556058": 17, "807705": 17, "linear_40": 17, "901954": 17, "301267": 17, "956539": 17, "linear_41": 17, "839805": 17, "597429": 17, "716181": 17, "linear_42": 17, "178945": 17, "651595": 17, "895699": 17, "829245": 17, "627592": 17, "637907": 17, "746186": 17, "255032": 17, "167313": 17, "000000": 17, "999756": 17, "031013": 17, "345k": 17, "17k": 17, "218m": 17, "counterpart": 17, "bit": [17, 23, 25, 26, 28, 32, 39], "4532": 17, "feedforward": [18, 25, 31, 44], "384": [18, 28], "192": [18, 28], "unmask": 18, "downsampl": [18, 24], "factor": [18, 23, 25, 26, 28, 29, 31, 33, 34, 43, 44, 45], "473": [18, 28], "246": [18, 25, 28, 36, 37], "477": 18, "warm_step": 18, "2000": [18, 26], "feedforward_dim": 18, "attention_dim": [18, 23, 25, 28], "encoder_unmasked_dim": 18, "zipformer_downsampling_factor": 18, "decode_chunk_len": 18, "257": [18, 25, 36, 37], "023": 18, "zipformer2": 18, "419": 18, "At": [18, 23, 28], "stack": 18, "downsampling_factor": 18, "037": 18, "655": 18, "346": 18, "68944004": 18, "347": 18, "260096": 18, "348": [18, 36], "716276": 18, "656": [18, 28], "349": 18, "69920376": 18, "351": 18, "353": 18, "174": [18, 28], "175": 18, "1344": 18, "assert": 18, "cached_len": 18, "num_lay": 18, "1348": 18, "cached_avg": 18, "1352": 18, "cached_kei": 18, "1356": 18, "cached_v": 18, "1360": 18, "cached_val2": 18, "1364": 18, "cached_conv1": 18, "1368": 18, "cached_conv2": 18, "1373": 18, "left_context_len": 18, "1884": 18, "x_size": 18, "2442": 18, "2449": 18, "2469": 18, "2473": 18, "2483": 18, "kv_len": 18, "k": [18, 31, 36, 37, 43, 44, 45], "2570": 18, "attn_output": 18, "bsz": 18, "num_head": 18, "seq_len": 18, "head_dim": 18, "2926": 18, "lorder": 18, "2652": 18, "2653": 18, "embed_dim": 18, "2666": 18, "1543": 18, "in_x_siz": 18, "1637": 18, "1643": 18, "in_channel": 18, "1571": 18, "1763": 18, "src1": 18, "src2": 18, "1779": 18, "dim1": 18, "1780": 18, "dim2": 18, "_trace": 18, "958": 18, "tracer": 18, "instead": [18, 25, 44], "tupl": 18, "namedtupl": 18, "absolut": 18, "know": [18, 29], "side": 18, "allow": [18, 31, 44], "behavior": [18, 25], "_c": 18, "_create_method_from_trac": 18, "646": 18, "357": 18, "102": [18, 23], "embedding_out": 18, "686": 18, "361": [18, 28, 32], "735": 18, "69": 18, "269m": 18, "269": [18, 23, 36, 37], "725": [18, 32], "1022k": 18, "266m": 18, "8m": 18, "509k": 18, "133m": 18, "152k": 18, "4m": 18, "1022": 18, "133": 18, "509": 18, "260": [18, 28], "360": 18, "365": 18, "280": [18, 28], "372": [18, 23], "state": [18, 23, 25, 26, 28, 31, 33, 34, 43, 44, 45], "026": 18, "410": 18, "411": [18, 28], "2028": 18, "2547": 18, "2029": 18, "23316": 18, "23317": 18, "23318": 18, "23319": 18, "23320": 18, "amount": [18, 24], "pad": [18, 23, 25, 26, 28, 31, 33, 34, 43, 44, 45], "conv2dsubsampl": 18, "v2": [18, 23, 28], "arrai": 18, "23300": 18, "element": 18, "onnx_pretrain": 19, "onnxruntim": 19, "separ": 19, "deploi": [19, 23, 28], "repo_url": 19, "basenam": 19, "tree": [20, 21, 23, 25, 26, 28, 32, 36, 37, 39, 43], "cpu_jit": [20, 23, 28, 31, 33, 34, 44, 45], "confus": 20, "move": [20, 31, 33, 34, 44, 45], "why": 20, "streaming_asr": [20, 21, 43, 44, 45], "conv_emform": 20, "offline_asr": [20, 31], "jit_pretrain": [21, 33, 34, 43], "baz": 21, "1best": [23, 26, 28, 32, 33, 34, 36, 37], "automag": [23, 26, 28, 29, 31, 32, 33, 34, 36, 37, 39, 43, 44, 45], "stop": [23, 25, 26, 28, 29, 31, 32, 33, 34, 36, 37, 39, 43, 44, 45], "By": [23, 26, 28, 29, 31, 32, 33, 34, 36, 37, 39, 43, 44, 45], "musan": [23, 26, 28, 29, 31, 33, 34, 43, 44, 45], "thei": [23, 25, 26, 28, 29, 31, 33, 34, 43, 44, 45], "intal": [23, 26], "sudo": [23, 26], "apt": [23, 26], "permiss": [23, 26], "commandlin": [23, 25, 26, 28, 31, 33, 34, 43, 44, 45], "quit": [23, 25, 26, 28, 31, 33, 34, 43, 44, 45], "experi": [23, 25, 26, 28, 29, 31, 33, 34, 39, 43, 44, 45], "world": [23, 25, 26, 28, 29, 31, 32, 33, 34, 43, 44, 45], "multi": [23, 25, 26, 28, 29, 31, 33, 34, 41, 43, 44, 45], "machin": [23, 25, 26, 28, 31, 33, 34, 43, 44, 45], "ddp": [23, 25, 26, 28, 31, 33, 34, 43, 44, 45], "implement": [23, 25, 26, 28, 29, 31, 33, 34, 41, 43, 44, 45], "present": [23, 25, 26, 28, 31, 33, 34, 43, 44, 45], "second": [23, 25, 26, 28, 29, 31, 33, 34, 39, 43, 44, 45], "utter": [23, 25, 26, 28, 31, 33, 34, 43, 44, 45], "oom": [23, 25, 26, 28, 31, 33, 34, 43, 44, 45], "due": [23, 25, 26, 28, 31, 33, 34, 43, 44, 45], "decai": [23, 26, 28, 33, 34, 43], "warmup": [23, 25, 26, 28, 31, 33, 34, 43, 44, 45], "function": [23, 25, 26, 28, 31, 32, 33, 34, 36, 37, 39, 43, 44, 45], "get_param": [23, 25, 26, 28, 31, 32, 33, 34, 36, 37, 39, 43, 44, 45], "directli": [23, 25, 26, 28, 29, 31, 33, 34, 43, 44, 45], "perturb": [23, 25, 26, 28, 31, 33, 34, 43, 44, 45], "actual": [23, 25, 26, 28, 31, 33, 34, 43, 44, 45], "3x150": [23, 25, 26], "450": [23, 25, 26], "visual": [23, 25, 26, 28, 31, 32, 33, 34, 36, 37, 39, 43, 44, 45], "logdir": [23, 25, 26, 28, 31, 32, 33, 34, 36, 37, 39, 43, 44, 45], "labelsmooth": 23, "someth": [23, 25, 26, 28, 31, 33, 34, 39, 43, 44], "tensorflow": [23, 25, 26, 28, 31, 33, 34, 39, 43, 44], "press": [23, 25, 26, 28, 31, 33, 34, 39, 43, 44, 45], "ctrl": [23, 25, 26, 28, 31, 33, 34, 39, 43, 44, 45], "engw8ksktzqs24zbv5dgcg": 23, "2021": [23, 26, 28, 32, 36, 37, 39], "22t11": 23, "scan": [23, 25, 26, 28, 31, 39, 43, 44], "116068": 23, "scalar": [23, 25, 26, 28, 31, 39, 43, 44], "listen": [23, 25, 26, 31, 39, 43, 44], "url": [23, 25, 26, 28, 31, 33, 34, 39, 43, 44], "xxxx": [23, 25, 26, 28, 31, 32, 33, 34, 36, 37, 39, 43, 44, 45], "saw": [23, 25, 26, 28, 31, 32, 33, 34, 36, 37, 39, 43, 44, 45], "consol": [23, 25, 26, 28, 31, 32, 33, 34, 36, 37, 39, 43, 44, 45], "avoid": [23, 25, 28], "nbest": [23, 28, 34], "lattic": [23, 26, 28, 31, 32, 36, 37, 44, 45], "uniqu": [23, 28, 31, 44, 45], "pkufool": [23, 26, 32], "icefall_asr_aishell_conformer_ctc": 23, "transcrib": [23, 25, 26, 28], "v1": [23, 26, 28, 32, 36, 37], "lang_char": [23, 25], "bac009s0764w0121": [23, 25, 26], "bac009s0764w0122": [23, 25, 26], "bac009s0764w0123": [23, 25, 26], "tran": [23, 26, 28, 32, 36, 37], "graph": [23, 26, 28, 31, 32, 36, 37, 44, 45], "id": [23, 26, 28, 32, 36, 37], "conveni": [23, 26, 28, 29], "eo": [23, 26, 28], "soxi": [23, 25, 26, 28, 32, 39], "sampl": [23, 25, 26, 28, 32, 33, 39, 44, 45], "precis": [23, 25, 26, 28, 31, 32, 39, 44, 45], "67263": [23, 25, 26], "cdda": [23, 25, 26, 28, 32, 39], "sector": [23, 25, 26, 28, 32, 39], "135k": [23, 25, 26], "256k": [23, 25, 26, 28], "sign": [23, 25, 26, 28, 39], "integ": [23, 25, 26, 28, 39], "pcm": [23, 25, 26, 28, 39], "65840": [23, 25, 26], "308": [23, 25, 26], "625": [23, 25, 26], "132k": [23, 25, 26], "64000": [23, 25, 26], "300": [23, 25, 26, 28, 29, 31, 44], "128k": [23, 25, 26, 39], "displai": [23, 25, 26, 28], "topologi": [23, 28], "num_decoder_lay": [23, 28], "vgg_frontend": [23, 25, 28], "use_feat_batchnorm": [23, 28], "f2fd997f752ed11bbef4c306652c433e83f9cf12": 23, "sun": 23, "sep": 23, "33cfe45": 23, "d57a873": 23, "nov": [23, 28], "hw": 23, "kangwei": 23, "icefall_aishell3": 23, "k2_releas": 23, "tokens_fil": 23, "words_fil": [23, 28, 39], "num_path": [23, 28, 31, 44, 45], "ngram_lm_scal": [23, 28], "attention_decoder_scal": [23, 28], "nbest_scal": [23, 28], "sos_id": [23, 28], "eos_id": [23, 28], "num_class": [23, 28, 39], "4336": [23, 25], "131": [23, 28], "134": 23, "138": [23, 25], "293": [23, 28], "369": [23, 28], "\u751a": [23, 25], "\u81f3": [23, 25], "\u51fa": [23, 25], "\u73b0": [23, 25], "\u4ea4": [23, 25], "\u6613": [23, 25], "\u51e0": [23, 25], "\u4e4e": [23, 25], "\u505c": [23, 25], "\u6b62": 23, "\u7684": [23, 25, 26], "\u60c5": [23, 25], "\u51b5": [23, 25], "\u4e00": [23, 25], "\u4e8c": [23, 25], "\u7ebf": [23, 25, 26], "\u57ce": [23, 25], "\u5e02": [23, 25], "\u867d": [23, 25], "\u7136": [23, 25], "\u4e5f": [23, 25, 26], "\u5904": [23, 25], "\u4e8e": [23, 25], "\u8c03": [23, 25], "\u6574": [23, 25], "\u4e2d": [23, 25, 26], "\u4f46": [23, 25, 26], "\u56e0": [23, 25], "\u4e3a": [23, 25], "\u805a": [23, 25], "\u96c6": [23, 25], "\u4e86": [23, 25, 26], "\u8fc7": [23, 25], "\u591a": [23, 25], "\u516c": [23, 25], "\u5171": [23, 25], "\u8d44": [23, 25], "\u6e90": [23, 25], "371": 23, "683": 23, "684": [23, 39], "651": [23, 39], "654": 23, "659": 23, "752": 23, "887": 23, "340": 23, "370": 23, "\u751a\u81f3": [23, 26], "\u51fa\u73b0": [23, 26], "\u4ea4\u6613": [23, 26], "\u51e0\u4e4e": [23, 26], "\u505c\u6b62": 23, "\u60c5\u51b5": [23, 26], "\u4e00\u4e8c": [23, 26], "\u57ce\u5e02": [23, 26], "\u867d\u7136": [23, 26], "\u5904\u4e8e": [23, 26], "\u8c03\u6574": [23, 26], "\u56e0\u4e3a": [23, 26], "\u805a\u96c6": [23, 26], "\u8fc7\u591a": [23, 26], "\u516c\u5171": [23, 26], "\u8d44\u6e90": [23, 26], "recor": [23, 28], "highest": [23, 28], "966": 23, "821": 23, "822": 23, "826": 23, "916": 23, "345": 23, "888": 23, "889": 23, "limit": [23, 25, 28, 41, 44], "upgrad": [23, 28], "pro": [23, 28], "finish": [23, 25, 26, 28, 29, 31, 32, 36, 37, 39, 44, 45], "NOT": [23, 25, 28, 39], "checkout": [23, 28], "hlg_decod": [23, 28], "four": [23, 28], "messag": [23, 28, 31, 33, 34, 43, 44, 45], "nn_model": [23, 28], "use_gpu": [23, 28], "word_tabl": [23, 28], "caution": [23, 28], "forward": [23, 28, 33], "cu": [23, 28], "int": [23, 28], "char": [23, 28], "98": 23, "150": [23, 28], "693": [23, 36], "165": [23, 28], "nnet_output": [23, 28], "185": [23, 28, 39], "217": [23, 28], "mandarin": 24, "beij": 24, "shell": 24, "technologi": 24, "ltd": 24, "peopl": 24, "accent": 24, "area": 24, "china": 24, "invit": 24, "particip": 24, "conduct": 24, "indoor": 24, "high": 24, "fidel": 24, "microphon": 24, "16khz": 24, "manual": 24, "through": 24, "profession": 24, "annot": 24, "inspect": 24, "free": [24, 29, 43], "academ": 24, "moder": 24, "research": 24, "field": 24, "openslr": 24, "ctc": [24, 27, 30, 34, 35, 38], "stateless": [24, 27, 31, 43, 44, 45], "conv1d": [25, 31, 43, 44, 45], "nn": [25, 31, 33, 34, 43, 44, 45], "tanh": 25, "borrow": 25, "ieeexplor": 25, "ieee": 25, "stamp": 25, "jsp": 25, "arnumb": 25, "9054419": 25, "predict": [25, 29, 31, 43, 44, 45], "charact": 25, "unit": 25, "vocabulari": 25, "87939824": 25, "optimized_transduc": 25, "technqiu": 25, "end": [25, 31, 33, 34, 39, 43, 44, 45], "furthermor": 25, "maximum": 25, "emit": 25, "per": [25, 31, 44, 45], "simplifi": [25, 41], "significantli": 25, "degrad": 25, "exactli": 25, "unprun": 25, "advantag": 25, "minim": 25, "pruned_transducer_stateless": [25, 31, 41, 44], "altern": 25, "though": 25, "transducer_stateless_modifi": 25, "pr": 25, "ram": 25, "tri": 25, "prob": [25, 43], "219": [25, 28], "lagz6hrcqxoigbfd5e0y3q": 25, "03t14": 25, "8477": 25, "250": [25, 32], "sym": [25, 31, 44, 45], "beam_search": [25, 31, 44, 45], "decoding_method": 25, "beam_4": 25, "ensur": 25, "give": 25, "poor": 25, "531": [25, 26], "994": [25, 28], "027": 25, "encoder_out_dim": 25, "f4fefe4882bc0ae59af951da3f47335d5495ef71": 25, "50d2281": 25, "mar": 25, "0815224919": 25, "75d558775b": 25, "mmnv8": 25, "72": [25, 28], "248": 25, "878": [25, 37], "880": 25, "891": 25, "113": [25, 28], "userwarn": 25, "__floordiv__": 25, "round": 25, "toward": 25, "trunc": 25, "floor": 25, "keep": [25, 31, 44, 45], "div": 25, "b": [25, 28, 36, 37], "rounding_mod": 25, "divis": 25, "x_len": 25, "163": [25, 28], "\u6ede": 25, "322": 25, "759": 25, "760": 25, "919": 25, "922": 25, "929": 25, "046": 25, "047": 25, "319": [25, 28], "798": 25, "831": [25, 37], "214": [25, 28], "215": [25, 28, 32], "402": 25, "topk_hyp_index": 25, "topk_index": 25, "logit": 25, "583": [25, 37], "lji9mwuorlow3jkdhxwk8a": 26, "13t11": 26, "4454": 26, "icefall_asr_aishell_tdnn_lstm_ctc": 26, "858": [26, 28], "389": [26, 28], "154": 26, "161": [26, 28], "536": 26, "539": 26, "917": 26, "129": 26, "\u505c\u6ede": 26, "mmi": [27, 30], "blank": [27, 30], "skip": [27, 29, 30, 31, 43, 44, 45], "distil": [27, 30], "hubert": [27, 30], "ligru": [27, 35], "full": [28, 29, 31, 33, 34, 43, 44, 45], "libri": [28, 29, 31, 33, 34, 43, 44, 45], "subset": [28, 31, 33, 34, 43, 44, 45], "3x960": [28, 31, 33, 34, 43, 44, 45], "2880": [28, 31, 33, 34, 43, 44, 45], "lzgnetjwrxc3yghnmd4kpw": 28, "24t16": 28, "4540": 28, "sentenc": 28, "piec": 28, "And": [28, 31, 33, 34, 43, 44, 45], "neither": 28, "nor": 28, "5000": 28, "033": 28, "537": 28, "538": 28, "full_libri": [28, 29], "406": 28, "464": 28, "548": 28, "776": 28, "652": [28, 39], "109226120": 28, "714": [28, 36], "206": 28, "944": 28, "1328": 28, "443": [28, 32], "2563": 28, "494": 28, "592": 28, "1715": 28, "52576": 28, "128": 28, "1424": 28, "807": 28, "506": 28, "808": [28, 36], "362": 28, "1477": 28, "2922": 28, "208": 28, "4295": 28, "52343": 28, "396": 28, "3584": 28, "432": 28, "680": [28, 36], "_pickl": 28, "unpicklingerror": 28, "invalid": 28, "hlg_modifi": 28, "g_4_gram": [28, 32, 36, 37], "sentencepiec": 28, "875": [28, 32], "212k": 28, "267440": [28, 32], "1253": [28, 32], "535k": 28, "83": [28, 32], "77200": [28, 32], "154k": 28, "554": 28, "7178d67e594bc7fa89c2b331ad7bd1c62a6a9eb4": 28, "8d93169": 28, "601": 28, "025": 28, "broffel": 28, "osom": 28, "723": 28, "775": 28, "881": 28, "234": 28, "571": 28, "whole": [28, 32, 36, 37, 44, 45], "857": 28, "979": 28, "055": 28, "117": 28, "051": 28, "363": 28, "959": [28, 37], "546": 28, "598": 28, "599": [28, 32], "833": 28, "834": 28, "915": 28, "076": 28, "110": 28, "397": 28, "999": [28, 31, 44, 45], "concaten": 28, "bucket": 28, "sampler": 28, "1000": 28, "ctc_decod": 28, "ngram_lm_rescor": 28, "attention_rescor": 28, "kind": [28, 31, 33, 34, 43, 44, 45], "105": 28, "125": [28, 39], "136": 28, "228": 28, "144": 28, "543": 28, "topo": 28, "547": 28, "729": 28, "703": 28, "545": 28, "279": 28, "122": 28, "126": 28, "135": [28, 39], "153": [28, 39], "945": 28, "475": 28, "191": [28, 36, 37], "398": 28, "515": 28, "deseri": 28, "441": 28, "fsaclass": 28, "loadfsa": 28, "const": 28, "string": 28, "c10": 28, "ignor": 28, "dummi": 28, "589": 28, "attention_scal": 28, "162": 28, "169": [28, 36, 37], "188": 28, "984": 28, "624": 28, "519": [28, 37], "632": 28, "645": [28, 39], "243": 28, "303": 28, "179": 28, "knowledg": 29, "vector": 29, "mvq": 29, "kd": 29, "pruned_transducer_stateless4": [29, 31, 41, 44], "theoret": 29, "applic": 29, "minor": 29, "out": 29, "thing": 29, "distillation_with_hubert": 29, "Of": 29, "cours": 29, "xl": 29, "proce": 29, "960h": [29, 33], "use_extracted_codebook": 29, "augment": 29, "th": [29, 36, 37], "fine": 29, "embedding_lay": 29, "num_codebook": 29, "under": 29, "vq_fbank_layer36_cb8": 29, "whola": 29, "snippet": 29, "echo": 29, "awk": 29, "split": 29, "_": 29, "pruned_transducer_stateless6": 29, "12359": 29, "spec": 29, "aug": 29, "warp": 29, "paid": 29, "suitabl": [31, 43, 44, 45], "pruned_transducer_stateless2": [31, 41, 44], "pruned_transducer_stateless5": [31, 41, 44], "scroll": [31, 33, 34, 43, 44, 45], "arxiv": [31, 43, 44, 45], "2206": [31, 43, 44, 45], "13236": [31, 43, 44, 45], "rework": [31, 41, 44], "daniel": [31, 44, 45], "joint": [31, 43, 44, 45], "contrari": [31, 43, 44, 45], "convent": [31, 43, 44, 45], "recurr": [31, 43, 44, 45], "2x": [31, 44, 45], "littl": [31, 44], "436000": [31, 33, 34, 43, 44, 45], "438000": [31, 33, 34, 43, 44, 45], "qogspbgsr8kzcrmmie9jgw": 31, "20t15": [31, 43, 44], "4468": [31, 43, 44], "210171": [31, 43, 44], "access": [31, 33, 34, 43, 44, 45], "googl": [31, 33, 34, 43, 44, 45], "6008": [31, 33, 34, 43, 44, 45], "localhost": [31, 33, 34, 43, 44, 45], "expos": [31, 33, 34, 43, 44, 45], "proxi": [31, 33, 34, 43, 44, 45], "bind_al": [31, 33, 34, 43, 44, 45], "fast_beam_search": [31, 33, 43, 44, 45], "474000": [31, 43, 44, 45], "largest": [31, 44, 45], "posterior": [31, 33, 44, 45], "algorithm": [31, 44, 45], "pdf": [31, 34, 44, 45], "1211": [31, 44, 45], "3711": [31, 44, 45], "espnet": [31, 44, 45], "net": [31, 44, 45], "beam_search_transduc": [31, 44, 45], "basicli": [31, 44, 45], "topk": [31, 44, 45], "expand": [31, 44, 45], "mode": [31, 44, 45], "being": [31, 44, 45], "hardcod": [31, 44, 45], "composit": [31, 44, 45], "log_prob": [31, 44, 45], "hard": [31, 41, 44, 45], "2211": [31, 44, 45], "00484": [31, 44, 45], "fast_beam_search_lg": [31, 44, 45], "trivial": [31, 44, 45], "fast_beam_search_nbest": [31, 44, 45], "random_path": [31, 44, 45], "shortest": [31, 44, 45], "fast_beam_search_nbest_lg": [31, 44, 45], "logic": [31, 44, 45], "smallest": [31, 43, 44, 45], "normal": [32, 36, 37, 39, 44], "icefall_asr_librispeech_tdnn": 32, "lstm_ctc": 32, "flac": 32, "116k": 32, "140k": 32, "343k": 32, "164k": 32, "105k": 32, "174k": 32, "pretraind": 32, "170": 32, "584": [32, 37], "791": 32, "245": 32, "098": 32, "099": 32, "methond": [32, 36, 37], "403": 32, "631": 32, "190": 32, "121": 32, "010": 32, "guidanc": 33, "bigger": 33, "simpli": 33, "discard": 33, "prevent": 33, "lconv": 33, "encourag": [33, 34, 43], "stabil": [33, 34], "doesn": 33, "warm": [33, 34], "xyozukpeqm62hbilud4upa": [33, 34], "ctc_guide_decode_b": 33, "pretrained_ctc": 33, "jit_pretrained_ctc": 33, "100h": 33, "yfyeung": 33, "wechat": 34, "zipformer_mmi": 34, "worker": [34, 43], "hp": 34, "tdnn_ligru_ctc": 36, "enough": [36, 37, 39], "luomingshuang": [36, 37], "icefall_asr_timit_tdnn_ligru_ctc": 36, "pretrained_average_9_25": 36, "fdhc0_si1559": [36, 37], "felc0_si756": [36, 37], "fmgd0_si1564": [36, 37], "ffprobe": [36, 37], "show_format": [36, 37], "nistspher": [36, 37], "database_id": [36, 37], "database_vers": [36, 37], "utterance_id": [36, 37], "dhc0_si1559": [36, 37], "sample_min": [36, 37], "4176": [36, 37], "sample_max": [36, 37], "5984": [36, 37], "bitrat": [36, 37], "pcm_s16le": [36, 37], "s16": [36, 37], "elc0_si756": [36, 37], "1546": [36, 37], "1989": [36, 37], "mgd0_si1564": [36, 37], "7626": [36, 37], "10573": [36, 37], "660": 36, "695": 36, "697": 36, "819": 36, "829": 36, "sil": [36, 37], "dh": [36, 37], "ih": [36, 37], "uw": [36, 37], "ah": [36, 37], "ii": [36, 37], "z": [36, 37], "aa": [36, 37], "ei": [36, 37], "dx": [36, 37], "d": [36, 37], "uh": [36, 37], "ng": [36, 37], "eh": [36, 37], "jh": [36, 37], "er": [36, 37], "ai": [36, 37], "hh": [36, 37], "aw": 36, "ae": [36, 37], "705": 36, "715": 36, "720": 36, "251": [36, 37], "ch": 36, "icefall_asr_timit_tdnn_lstm_ctc": 37, "pretrained_average_16_25": 37, "816": 37, "827": 37, "unk": 37, "739": 37, "971": 37, "977": 37, "978": 37, "981": 37, "ow": 37, "ykubhb5wrmosxykid1z9eg": 39, "23t23": 39, "icefall_asr_yesno_tdnn": 39, "l_disambig": 39, "lexicon_disambig": 39, "0_0_0_1_0_0_0_1": 39, "0_0_1_0_0_0_1_0": 39, "0_0_1_0_0_1_1_1": 39, "0_0_1_0_1_0_0_1": 39, "0_0_1_1_0_0_0_1": 39, "0_0_1_1_0_1_1_0": 39, "0_0_1_1_1_0_0_0": 39, "0_0_1_1_1_1_0_0": 39, "0_1_0_0_0_1_0_0": 39, "0_1_0_0_1_0_1_0": 39, "0_1_0_1_0_0_0_0": 39, "0_1_0_1_1_1_0_0": 39, "0_1_1_0_0_1_1_1": 39, "0_1_1_1_0_0_1_0": 39, "0_1_1_1_1_0_1_0": 39, "1_0_0_0_0_0_0_0": 39, "1_0_0_0_0_0_1_1": 39, "1_0_0_1_0_1_1_1": 39, "1_0_1_1_0_1_1_1": 39, "1_0_1_1_1_1_0_1": 39, "1_1_0_0_0_1_1_1": 39, "1_1_0_0_1_0_1_1": 39, "1_1_0_1_0_1_0_0": 39, "1_1_0_1_1_0_0_1": 39, "1_1_0_1_1_1_1_0": 39, "1_1_1_0_0_1_0_1": 39, "1_1_1_0_1_0_1_0": 39, "1_1_1_1_0_0_1_0": 39, "1_1_1_1_1_0_0_0": 39, "1_1_1_1_1_1_1_1": 39, "54080": 39, "507": 39, "108k": 39, "ye": 39, "hebrew": 39, "NO": 39, "119": 39, "650": 39, "139": 39, "143": 39, "198": 39, "181": 39, "186": 39, "187": 39, "213": 39, "correctli": 39, "simplest": 39, "former": 41, "idea": 41, "mask": [41, 44, 45], "wenet": 41, "did": 41, "request": 41, "metion": 41, "complic": 41, "techniqu": 41, "bank": 41, "memor": 41, "histori": 41, "introduc": 41, "variant": 41, "pruned_stateless_emformer_rnnt2": 41, "conv_emformer_transducer_stateless": 41, "ourself": 41, "mechan": 41, "onlin": 43, "lstm_transducer_stateless": 43, "lower": 43, "prepare_giga_speech": 43, "cj2vtpiwqhkn9q1tx6ptpg": 43, "dynam": [44, 45], "causal": 44, "short": [44, 45], "2012": 44, "05481": 44, "flag": 44, "indic": [44, 45], "whether": 44, "sequenc": [44, 45], "uniformli": [44, 45], "seen": [44, 45], "97vkxf80ru61cnp2alwzzg": 44, "streaming_decod": [44, 45], "wise": [44, 45], "parallel": [44, 45], "bath": [44, 45], "parallelli": [44, 45], "seem": 44, "benefit": 44, "mdoel": 44, "320m": 45, "550": 45, "scriptmodul": 45, "jit_trace_export": 45, "jit_trace_pretrain": 45, "task": 46}, "objects": {}, "objtypes": {}, "objnames": {}, "titleterms": {"follow": 0, "code": 0, "style": 0, "contribut": [1, 3], "document": 1, "how": [2, 14, 20, 21], "creat": [2, 13], "recip": [2, 46], "data": [2, 13, 23, 25, 26, 28, 29, 31, 32, 33, 34, 36, 37, 39, 43, 44, 45], "prepar": [2, 13, 23, 25, 26, 28, 29, 31, 32, 33, 34, 36, 37, 39, 43, 44, 45], "train": [2, 10, 13, 16, 17, 18, 19, 23, 25, 26, 28, 29, 31, 32, 33, 34, 36, 37, 39, 43, 44, 45], "decod": [2, 5, 6, 7, 13, 14, 19, 23, 25, 26, 28, 29, 31, 32, 33, 34, 36, 37, 39, 43, 44, 45], "pre": [2, 10, 16, 17, 18, 19, 23, 25, 26, 28, 31, 32, 33, 34, 36, 37, 39, 43, 44, 45], "model": [2, 5, 10, 14, 16, 17, 18, 19, 20, 21, 22, 23, 25, 26, 28, 31, 32, 33, 34, 36, 37, 39, 43, 44, 45], "lodr": [4, 6], "rnn": 4, "transduc": [4, 6, 7, 16, 17, 18, 25, 31, 43, 44, 45], "wer": [4, 5, 6, 7, 28], "differ": [4, 6, 7], "beam": [4, 6, 7, 25], "size": [4, 6, 7], "languag": 5, "lm": [5, 6, 28], "rescor": [5, 6, 23, 28], "base": [5, 6], "method": [5, 6], "v": [5, 6], "shallow": [5, 6, 7], "fusion": [5, 6, 7], "The": [5, 6, 25], "number": [5, 6], "each": [5, 6], "field": [5, 6], "i": [5, 6], "test": [5, 6, 7, 13, 16, 17, 18], "clean": [5, 6, 7], "other": [5, 6], "time": [5, 6, 7], "frequent": 8, "ask": 8, "question": 8, "faq": 8, "oserror": 8, "libtorch_hip": 8, "so": 8, "cannot": 8, "open": 8, "share": 8, "object": 8, "file": [8, 19], "directori": 8, "attributeerror": 8, "modul": 8, "distutil": 8, "ha": 8, "attribut": 8, "version": 8, "importerror": 8, "libpython3": 8, "10": 8, "1": [8, 13, 16, 17, 18, 23, 25, 26, 28], "0": [8, 13], "No": 8, "huggingfac": [9, 11], "space": 11, "youtub": [11, 13], "video": [11, 13], "icefal": [12, 13, 16, 17, 18], "content": [12, 46], "instal": [13, 16, 17, 18, 23, 25, 26, 28, 32, 36, 37], "cuda": 13, "toolkit": 13, "cudnn": 13, "torch": [13, 16, 17, 18, 20, 21, 31, 33, 34, 43, 44, 45], "torchaudio": 13, "2": [13, 16, 17, 18, 23, 25, 26, 28], "k2": 13, "3": [13, 16, 17, 18, 23, 25, 28], "lhots": 13, "4": [13, 16, 17, 18], "download": [13, 16, 17, 18, 19, 23, 25, 26, 28, 31, 32, 33, 34, 36, 37, 39, 43, 44, 45], "exampl": [13, 19, 23, 25, 26, 28, 31, 33, 34, 43, 44, 45], "virtual": 13, "environ": 13, "5": [13, 16, 17, 18], "6": [13, 16, 17, 18], "your": 13, "export": [14, 15, 16, 17, 18, 19, 20, 21, 22, 31, 33, 34, 43, 44, 45], "state_dict": [14, 31, 33, 34, 43, 44, 45], "when": [14, 20, 21], "us": [14, 20, 21, 31, 33, 34, 43, 44, 45], "run": 14, "py": 14, "ncnn": [15, 16, 17, 18], "convemform": 16, "pnnx": [16, 17, 18], "via": [16, 17, 18], "jit": [16, 17, 18, 20, 21, 31, 33, 34, 43, 44, 45], "trace": [16, 17, 18, 21, 43, 45], "torchscript": [16, 17, 18], "modifi": [16, 17, 18, 25], "encod": [16, 17, 18], "sherpa": [16, 17, 18, 19, 31, 44, 45], "7": [16, 17], "option": [16, 17, 23, 26, 28, 31, 33, 34, 43, 44, 45], "int8": [16, 17], "quantiz": [16, 17], "lstm": [17, 26, 32, 37, 43], "stream": [18, 27, 40, 41, 44, 45], "zipform": [18, 33, 34, 45], "onnx": 19, "sound": 19, "script": [20, 31, 33, 34, 44, 45], "conform": [23, 28, 41], "ctc": [23, 26, 28, 32, 33, 36, 37, 39], "configur": [23, 26, 28, 31, 33, 34, 43, 44, 45], "log": [23, 25, 26, 28, 31, 33, 34, 43, 44, 45], "usag": [23, 25, 26, 28, 31, 33, 34, 43, 44, 45], "case": [23, 25, 26, 28], "kaldifeat": [23, 25, 26, 28, 32, 36, 37, 39], "hlg": [23, 26, 28], "attent": [23, 28], "colab": [23, 25, 26, 28, 32, 36, 37, 39], "notebook": [23, 25, 26, 28, 32, 36, 37, 39], "deploy": [23, 28], "c": [23, 28], "aishel": 24, "stateless": 25, "loss": 25, "todo": 25, "greedi": 25, "search": 25, "tdnn": [26, 32, 36, 37, 39], "non": 27, "asr": [27, 40], "comput": 28, "n": 28, "gram": 28, "distil": 29, "hubert": 29, "codebook": 29, "index": 29, "librispeech": [30, 42], "prune": [31, 44], "statelessx": [31, 44], "pretrain": [31, 33, 34, 43, 44, 45], "deploi": [31, 44, 45], "infer": [32, 36, 37, 39], "blank": 33, "skip": 33, "mmi": 34, "timit": 35, "ligru": 36, "yesno": 38, "introduct": 41, "emform": 41, "which": 43, "simul": [44, 45], "real": [44, 45], "tabl": 46}, "envversion": {"sphinx.domains.c": 2, "sphinx.domains.changeset": 1, "sphinx.domains.citation": 1, "sphinx.domains.cpp": 8, "sphinx.domains.index": 1, "sphinx.domains.javascript": 2, "sphinx.domains.math": 2, "sphinx.domains.python": 3, "sphinx.domains.rst": 2, "sphinx.domains.std": 2, "sphinx.ext.todo": 2, "sphinx": 57}, "alltitles": {"Follow the code style": [[0, "follow-the-code-style"]], "Contributing to Documentation": [[1, "contributing-to-documentation"]], "How to create a recipe": [[2, "how-to-create-a-recipe"]], "Data Preparation": [[2, "data-preparation"], [25, "data-preparation"]], "Training": [[2, "training"], [13, "training"], [23, "training"], [25, "training"], [26, "training"], [28, "training"], [29, "training"], [31, "training"], [32, "training"], [33, "training"], [34, "training"], [36, "training"], [37, "training"], [39, "training"], [43, "training"], [44, "training"], [45, "training"]], "Decoding": [[2, "decoding"], [13, "decoding"], [23, "decoding"], [25, "decoding"], [26, "decoding"], [28, "decoding"], [29, "decoding"], [31, "decoding"], [32, "decoding"], [33, "decoding"], [34, "decoding"], [36, "decoding"], [37, "decoding"], [39, "decoding"], [43, "decoding"], [44, "decoding"], [45, "decoding"]], "Pre-trained model": [[2, "pre-trained-model"]], "Contributing": [[3, "contributing"]], "LODR for RNN Transducer": [[4, "lodr-for-rnn-transducer"]], "WER of LODR with different beam sizes": [[4, "id1"]], "Decoding with language models": [[5, "decoding-with-language-models"]], "LM-rescoring-based methods vs shallow-fusion-based methods (The numbers in each field is WER on test-clean, WER on test-other and decoding time on test-clean)": [[5, "id1"], [6, "id3"]], "LM rescoring for Transducer": [[6, "lm-rescoring-for-transducer"]], "WERs of LM rescoring with different beam sizes": [[6, "id1"]], "WERs of LM rescoring + LODR with different beam sizes": [[6, "id2"]], "Shallow fusion for Transducer": [[7, "shallow-fusion-for-transducer"]], "WERs and decoding time (on test-clean) of shallow fusion with different beam sizes": [[7, "id2"]], "Frequently Asked Questions (FAQs)": [[8, "frequently-asked-questions-faqs"]], "OSError: libtorch_hip.so: cannot open shared object file: no such file or directory": [[8, "oserror-libtorch-hip-so-cannot-open-shared-object-file-no-such-file-or-directory"]], "AttributeError: module \u2018distutils\u2019 has no attribute \u2018version\u2019": [[8, "attributeerror-module-distutils-has-no-attribute-version"]], "ImportError: libpython3.10.so.1.0: cannot open shared object file: No such file or directory": [[8, "importerror-libpython3-10-so-1-0-cannot-open-shared-object-file-no-such-file-or-directory"]], "Huggingface": [[9, "huggingface"]], "Pre-trained models": [[10, "pre-trained-models"]], "Huggingface spaces": [[11, "huggingface-spaces"]], "YouTube Video": [[11, "youtube-video"], [13, "youtube-video"]], "Icefall": [[12, "icefall"]], "Contents:": [[12, null]], "Installation": [[13, "installation"]], "(0) Install CUDA toolkit and cuDNN": [[13, "install-cuda-toolkit-and-cudnn"]], "(1) Install torch and torchaudio": [[13, "install-torch-and-torchaudio"]], "(2) Install k2": [[13, "install-k2"]], "(3) Install lhotse": [[13, "install-lhotse"]], "(4) Download icefall": [[13, "download-icefall"]], "Installation example": [[13, "installation-example"]], "(1) Create a virtual environment": [[13, "create-a-virtual-environment"]], "(2) Install CUDA toolkit and cuDNN": [[13, "id1"]], "(3) Install torch and torchaudio": [[13, "id2"]], "(4) Install k2": [[13, "id3"]], "(5) Install lhotse": [[13, "id5"]], "(6) Download icefall": [[13, "id6"]], "Test Your Installation": [[13, "test-your-installation"]], "Data preparation": [[13, "data-preparation"], [23, "data-preparation"], [26, "data-preparation"], [28, "data-preparation"], [29, "data-preparation"], [31, "data-preparation"], [32, "data-preparation"], [33, "data-preparation"], [34, "data-preparation"], [36, "data-preparation"], [37, "data-preparation"], [39, "data-preparation"], [43, "data-preparation"], [44, "data-preparation"], [45, "data-preparation"]], "Export model.state_dict()": [[14, "export-model-state-dict"], [31, "export-model-state-dict"], [33, "export-model-state-dict"], [34, "export-model-state-dict"], [43, "export-model-state-dict"], [44, "export-model-state-dict"], [45, "export-model-state-dict"]], "When to use it": [[14, "when-to-use-it"], [20, "when-to-use-it"], [21, "when-to-use-it"]], "How to export": [[14, "how-to-export"], [20, "how-to-export"], [21, "how-to-export"]], "How to use the exported model": [[14, "how-to-use-the-exported-model"], [20, "how-to-use-the-exported-model"]], "Use the exported model to run decode.py": [[14, "use-the-exported-model-to-run-decode-py"]], "Export to ncnn": [[15, "export-to-ncnn"]], "Export ConvEmformer transducer models to ncnn": [[16, "export-convemformer-transducer-models-to-ncnn"]], "1. Download the pre-trained model": [[16, "download-the-pre-trained-model"], [17, "download-the-pre-trained-model"], [18, "download-the-pre-trained-model"]], "2. Install ncnn and pnnx": [[16, "install-ncnn-and-pnnx"], [17, "install-ncnn-and-pnnx"], [18, "install-ncnn-and-pnnx"]], "3. Export the model via torch.jit.trace()": [[16, "export-the-model-via-torch-jit-trace"], [17, "export-the-model-via-torch-jit-trace"], [18, "export-the-model-via-torch-jit-trace"]], "4. Export torchscript model via pnnx": [[16, "export-torchscript-model-via-pnnx"], [17, "export-torchscript-model-via-pnnx"], [18, "export-torchscript-model-via-pnnx"]], "5. Test the exported models in icefall": [[16, "test-the-exported-models-in-icefall"], [17, "test-the-exported-models-in-icefall"], [18, "test-the-exported-models-in-icefall"]], "6. Modify the exported encoder for sherpa-ncnn": [[16, "modify-the-exported-encoder-for-sherpa-ncnn"], [17, "modify-the-exported-encoder-for-sherpa-ncnn"], [18, "modify-the-exported-encoder-for-sherpa-ncnn"]], "7. (Optional) int8 quantization with sherpa-ncnn": [[16, "optional-int8-quantization-with-sherpa-ncnn"], [17, "optional-int8-quantization-with-sherpa-ncnn"]], "Export LSTM transducer models to ncnn": [[17, "export-lstm-transducer-models-to-ncnn"]], "Export streaming Zipformer transducer models to ncnn": [[18, "export-streaming-zipformer-transducer-models-to-ncnn"]], "Export to ONNX": [[19, "export-to-onnx"]], "sherpa-onnx": [[19, "sherpa-onnx"]], "Example": [[19, "example"]], "Download the pre-trained model": [[19, "download-the-pre-trained-model"], [23, "download-the-pre-trained-model"], [25, "download-the-pre-trained-model"], [26, "download-the-pre-trained-model"], [28, "download-the-pre-trained-model"], [32, "download-the-pre-trained-model"], [36, "download-the-pre-trained-model"], [37, "download-the-pre-trained-model"], [39, "download-the-pre-trained-model"]], "Export the model to ONNX": [[19, "export-the-model-to-onnx"]], "Decode sound files with exported ONNX models": [[19, "decode-sound-files-with-exported-onnx-models"]], "Export model with torch.jit.script()": [[20, "export-model-with-torch-jit-script"]], "Export model with torch.jit.trace()": [[21, "export-model-with-torch-jit-trace"]], "How to use the exported models": [[21, "how-to-use-the-exported-models"]], "Model export": [[22, "model-export"]], "Conformer CTC": [[23, "conformer-ctc"], [28, "conformer-ctc"]], "Configurable options": [[23, "configurable-options"], [26, "configurable-options"], [28, "configurable-options"], [31, "configurable-options"], [33, "configurable-options"], [34, "configurable-options"], [43, "configurable-options"], [44, "configurable-options"], [45, "configurable-options"]], "Pre-configured options": [[23, "pre-configured-options"], [26, "pre-configured-options"], [28, "pre-configured-options"], [31, "pre-configured-options"], [33, "pre-configured-options"], [34, "pre-configured-options"], [43, "pre-configured-options"], [44, "pre-configured-options"], [45, "pre-configured-options"]], "Training logs": [[23, "training-logs"], [25, "training-logs"], [26, "training-logs"], [28, "training-logs"], [31, "training-logs"], [33, "training-logs"], [34, "training-logs"], [43, "training-logs"], [44, "training-logs"], [45, "training-logs"]], "Usage examples": [[23, "usage-examples"], [25, "usage-examples"], [26, "usage-examples"], [28, "usage-examples"]], "Case 1": [[23, "case-1"], [25, "case-1"], [26, "case-1"], [28, "case-1"]], "Case 2": [[23, "case-2"], [25, "case-2"], [26, "case-2"], [28, "case-2"]], "Case 3": [[23, "case-3"], [25, "case-3"], [28, "case-3"]], "Pre-trained Model": [[23, "pre-trained-model"], [25, "pre-trained-model"], [26, "pre-trained-model"], [28, "pre-trained-model"], [32, "pre-trained-model"], [36, "pre-trained-model"], [37, "pre-trained-model"], [39, "pre-trained-model"]], "Install kaldifeat": [[23, "install-kaldifeat"], [25, "install-kaldifeat"], [26, "install-kaldifeat"], [28, "install-kaldifeat"], [32, "install-kaldifeat"], [36, "install-kaldifeat"], [37, "install-kaldifeat"]], "Usage": [[23, "usage"], [25, "usage"], [26, "usage"], [28, "usage"]], "CTC decoding": [[23, "ctc-decoding"], [28, "ctc-decoding"], [28, "id2"]], "HLG decoding": [[23, "hlg-decoding"], [23, "id2"], [26, "hlg-decoding"], [28, "hlg-decoding"], [28, "id3"]], "HLG decoding + attention decoder rescoring": [[23, "hlg-decoding-attention-decoder-rescoring"]], "Colab notebook": [[23, "colab-notebook"], [25, "colab-notebook"], [26, "colab-notebook"], [28, "colab-notebook"], [32, "colab-notebook"], [36, "colab-notebook"], [37, "colab-notebook"], [39, "colab-notebook"]], "Deployment with C++": [[23, "deployment-with-c"], [28, "deployment-with-c"]], "aishell": [[24, "aishell"]], "Stateless Transducer": [[25, "stateless-transducer"]], "The Model": [[25, "the-model"]], "The Loss": [[25, "the-loss"]], "Todo": [[25, "id1"]], "Greedy search": [[25, "greedy-search"]], "Beam search": [[25, "beam-search"]], "Modified Beam search": [[25, "modified-beam-search"]], "TDNN-LSTM CTC": [[26, "tdnn-lstm-ctc"]], "Non Streaming ASR": [[27, "non-streaming-asr"]], "HLG decoding + LM rescoring": [[28, "hlg-decoding-lm-rescoring"]], "HLG decoding + LM rescoring + attention decoder rescoring": [[28, "hlg-decoding-lm-rescoring-attention-decoder-rescoring"]], "Compute WER with the pre-trained model": [[28, "compute-wer-with-the-pre-trained-model"]], "HLG decoding + n-gram LM rescoring": [[28, "hlg-decoding-n-gram-lm-rescoring"]], "HLG decoding + n-gram LM rescoring + attention decoder rescoring": [[28, "hlg-decoding-n-gram-lm-rescoring-attention-decoder-rescoring"]], "Distillation with HuBERT": [[29, "distillation-with-hubert"]], "Codebook index preparation": [[29, "codebook-index-preparation"]], "LibriSpeech": [[30, "librispeech"], [42, "librispeech"]], "Pruned transducer statelessX": [[31, "pruned-transducer-statelessx"], [44, "pruned-transducer-statelessx"]], "Usage example": [[31, "usage-example"], [33, "usage-example"], [34, "usage-example"], [43, "usage-example"], [44, "usage-example"], [45, "usage-example"]], "Export Model": [[31, "export-model"], [44, "export-model"], [45, "export-model"]], "Export model using torch.jit.script()": [[31, "export-model-using-torch-jit-script"], [33, "export-model-using-torch-jit-script"], [34, "export-model-using-torch-jit-script"], [44, "export-model-using-torch-jit-script"], [45, "export-model-using-torch-jit-script"]], "Download pretrained models": [[31, "download-pretrained-models"], [33, "download-pretrained-models"], [34, "download-pretrained-models"], [43, "download-pretrained-models"], [44, "download-pretrained-models"], [45, "download-pretrained-models"]], "Deploy with Sherpa": [[31, "deploy-with-sherpa"], [44, "deploy-with-sherpa"], [45, "deploy-with-sherpa"]], "TDNN-LSTM-CTC": [[32, "tdnn-lstm-ctc"], [37, "tdnn-lstm-ctc"]], "Inference with a pre-trained model": [[32, "inference-with-a-pre-trained-model"], [36, "inference-with-a-pre-trained-model"], [37, "inference-with-a-pre-trained-model"], [39, "inference-with-a-pre-trained-model"]], "Zipformer CTC Blank Skip": [[33, "zipformer-ctc-blank-skip"]], "Export models": [[33, "export-models"], [34, "export-models"], [43, "export-models"]], "Zipformer MMI": [[34, "zipformer-mmi"]], "TIMIT": [[35, "timit"]], "TDNN-LiGRU-CTC": [[36, "tdnn-ligru-ctc"]], "YesNo": [[38, "yesno"]], "TDNN-CTC": [[39, "tdnn-ctc"]], "Download kaldifeat": [[39, "download-kaldifeat"]], "Streaming ASR": [[40, "streaming-asr"]], "Introduction": [[41, "introduction"]], "Streaming Conformer": [[41, "streaming-conformer"]], "Streaming Emformer": [[41, "streaming-emformer"]], "LSTM Transducer": [[43, "lstm-transducer"]], "Which model to use": [[43, "which-model-to-use"]], "Export model using torch.jit.trace()": [[43, "export-model-using-torch-jit-trace"], [45, "export-model-using-torch-jit-trace"]], "Simulate streaming decoding": [[44, "simulate-streaming-decoding"], [45, "simulate-streaming-decoding"]], "Real streaming decoding": [[44, "real-streaming-decoding"], [45, "real-streaming-decoding"]], "Zipformer Transducer": [[45, "zipformer-transducer"]], "Recipes": [[46, "recipes"]], "Table of Contents": [[46, null]]}, "indexentries": {}}) \ No newline at end of file