mirrors/kaldifeat
1
0
Fork 0
mirror of https://github.com/csukuangfj/kaldifeat.git synced 2025-08-09 10:02:20 +00:00
Code Issues Packages Projects Releases Wiki Activity

Fix ci (#73)

Browse Source
This commit is contained in:
Fangjun Kuang 2023-07-24 09:46:44 +08:00 committed by GitHub
parent 6be131252a
commit 1868136b40
No known key found for this signature in database
GPG Key ID: 4AEE18F83AFDEB23
6 changed files with 105 additions and 6 deletions
Show Stats Download Patch File Download Diff File Expand all files Collapse all files

2
.github/workflows/build-doc.yml vendored
View File

@ -52,6 +52,8 @@ jobs:
cd doc
python3 -m pip install -r ./requirements.txt
make html
cp source/cpu.html build/html/
cp source/cuda.html build/html/
touch build/html/.nojekyll
- name: Deploy

2
doc/requirements.txt
View File

@ -1,6 +1,6 @@
dataclasses
recommonmark
sphinx
sphinx<7.0
sphinx-autodoc-typehints
sphinx_rtd_theme
sphinxcontrib-bibtex

92
doc/source/installation/from_wheels.rst
View File

@ -5,3 +5,95 @@ You can find pre-compiled wheels at
- CPU wheels: `<https://csukuangfj.github.io/kaldifeat/cpu.html>`_
- CUDA wheels: `<https://csukuangfj.github.io/kaldifeat/cuda.html>`_
We give a few examples below to show you how to install `kaldifeat`_ from
pre-compiled wheels.
Linux (CPU)
-----------
Suppose you want to install the following wheel:
.. code-block:: bash
https://huggingface.co/csukuangfj/kaldifeat/resolve/main/macos/kaldifeat-1.24.dev20230722+cpu.torch2.0.1-cp311-cp311-macosx_10_9_x86_64.whl
you can use one of the following methods:
.. code-block:: bash
# method 1
pip install torch==2.0.1+cpu -f https://download.pytorch.org/whl/torch_stable.html
pip install kaldifeat==1.24.dev20230722+cpu.torch2.0.1 -f https://csukuangfj.github.io/kaldifeat/cpu.html
# method 2
pip install torch==2.0.1+cpu -f https://download.pytorch.org/whl/torch_stable.html
wget https://huggingface.co/csukuangfj/kaldifeat/resolve/main/macos/kaldifeat-1.24.dev20230722+cpu.torch2.0.1-cp311-cp311-macosx_10_9_x86_64.whl
pip install ./kaldifeat-1.24.dev20230722+cpu.torch2.0.1-cp311-cp311-macosx_10_9_x86_64.whl
Windows (CPU)
--------------
Suppose you want to install the following wheel:
.. code-block:: bash
https://huggingface.co/csukuangfj/kaldifeat/resolve/main/windows-cpu/kaldifeat-1.24.dev20230722+cpu.torch2.0.1-cp311-cp311-win_amd64.whl
you can use one of the following methods:
.. code-block:: bash
# method 1
pip install torch==2.0.1+cpu -f https://download.pytorch.org/whl/torch_stable.html
pip install kaldifeat==1.24.dev20230722+cpu.torch2.0.1 -f https://csukuangfj.github.io/kaldifeat/cpu.html
# method 2
pip install torch==2.0.1+cpu -f https://download.pytorch.org/whl/torch_stable.html
wget https://huggingface.co/csukuangfj/kaldifeat/resolve/main/windows-cpu/kaldifeat-1.24.dev20230722+cpu.torch2.0.1-cp311-cp311-win_amd64.whl
pip install ./kaldifeat-1.24.dev20230722+cpu.torch2.0.1-cp311-cp311-win_amd64.whl
macOS (CPU)
-----------
Suppose you want to install the following wheel:
.. code-block:: bash
https://huggingface.co/csukuangfj/kaldifeat/resolve/main/macos/kaldifeat-1.24.dev20230722+cpu.torch2.0.1-cp311-cp311-macosx_10_9_x86_64.whl
you can use one of the following methods:
.. code-block:: bash
# method 1
pip install torch==2.0.1
pip install kaldifeat==1.24.dev20230722+cpu.torch2.0.1 -f https://csukuangfj.github.io/kaldifeat/cpu.html
# method 2
pip install torch==2.0.1+cpu -f https://download.pytorch.org/whl/torch_stable.html
wget https://huggingface.co/csukuangfj/kaldifeat/resolve/main/macos/kaldifeat-1.24.dev20230722+cpu.torch2.0.1-cp311-cp311-macosx_10_9_x86_64.whl
pip install ./kaldifeat-1.24.dev20230722+cpu.torch2.0.1-cp311-cp311-macosx_10_9_x86_64.whl
Linux (CUDA)
------------
Suppose you want to install the following wheel:
.. code-block:: bash
https://huggingface.co/csukuangfj/kaldifeat/resolve/main/ubuntu-cuda/kaldifeat-1.24.dev20230722+cuda11.8.torch2.0.1-cp311-cp311-manylinux_2_17_x86_64.manylinux2014_x86_64.whl
you can use one of the following methods:
.. code-block:: bash
# method 1
pip install torch==2.0.1+cu118 -f https://download.pytorch.org/whl/torch_stable.html
pip install kaldifeat==1.24.dev20230722+cuda11.8.torch2.0.1 -f https://csukuangfj.github.io/kaldifeat/cuda.html
# method 2
pip install torch==2.0.1+cu118 -f https://download.pytorch.org/whl/torch_stable.html
wget https://huggingface.co/csukuangfj/kaldifeat/resolve/main/ubuntu-cuda/kaldifeat-1.24.dev20230722+cuda11.8.torch2.0.1-cp311-cp311-manylinux_2_17_x86_64.manylinux2014_x86_64.whl
pip install ./kaldifeat-1.24.dev20230722+cuda11.8.torch2.0.1-cp311-cp311-manylinux_2_17_x86_64.manylinux2014_x86_64.whl

2
doc/source/installation/index.rst
View File

@ -2,7 +2,7 @@ Installation
============
.. toctree::
:maxdepth: 2
:maxdepth: 3
./from_wheels.rst
./from_source.rst

1
doc/source/usage/fbank_options.rst
View File

@ -10,7 +10,6 @@ The following code shows how to construct an instance of `kaldifeat.FbankOptions
.. literalinclude:: ./code/fbank_options-1.txt
:caption: Usage of `kaldifeat.FbankOptions`_
:emphasize-lines: 6,8,22,37
:language: python
Note that we reuse the same option name with `compute-fbank-feats`_ from `Kaldi`_:

12
scripts/github_actions/generate_build_matrix.py
View File

@ -131,15 +131,21 @@ def generate_build_matrix(enable_cuda, for_windows, for_macos, test_only_latest_
},
"1.13.1": {
"python-version": ["3.7", "3.8", "3.9", "3.10", "3.11"],
"cuda": ["11.6", "11.7"], # default 11.7
"cuda": ["11.6", "11.7"] # default 11.7
if not for_windows
else ["11.6.2", "11.7.1"],
},
"2.0.0": {
"python-version": ["3.8", "3.9", "3.10", "3.11"],
"cuda": ["11.7", "11.8"], # default 11.7
"cuda": ["11.7", "11.8"] # default 11.7
if not for_windows
else ["11.7.1", "11.8.0"],
},
"2.0.1": {
"python-version": ["3.8", "3.9", "3.10", "3.11"],
"cuda": ["11.7", "11.8"], # default 11.7
"cuda": ["11.7", "11.8"] # default 11.7
if not for_windows
else ["11.7.1", "11.8.0"],
},
}
if test_only_latest_torch: