{ai}[foss/2024a] WhisperX v3.7.4, ONNX-Runtime v1.23.2, ONNX v1.20.0, pyannote.audio v3.4.0, torchaudio v2.7.1, ... w/ CUDA 12.6.0 #24922

pavelToman · 2025-12-18T15:11:13Z

(created using eb --new-pr)
resolves vscentrum/vsc-software-stack#613
requires:

{ai}[foss/2024a] Lightning v2.5.6, PyTorch-Lightning v2.5.6, bitsandbytes v0.46.1 w/ CUDA 12.6.0 #25262
{ai}[foss/2024a] PyTorch v2.7.1 w/ CUDA 12.6.0 #23923
Extend list of easyconfig parameters that extensions should not inherit: sources, patches, checksums, skipsteps easybuild-framework#5075
- fix installation of bitsandbytes-0.46.1: {ai}[foss/2024a] bitsandbytes v0.46.1, PyTorch-Lightning v2.5.2 #23446 (comment)

…ntime-1.23.2-foss-2024a-CUDA-12.6.0.eb, ONNX-1.20.0-gfbf-2024a.eb, CTranslate2-4.5.0-foss-2024a-CUDA-12.6.0.eb, cpu_features-0.10.1-GCCcore-13.3.0.eb, pyannote.audio-3.4.0-foss-2024a-CUDA-12.6.0.eb, torchaudio-2.7.1-foss-2024a-CUDA-12.6.0.eb, Lightning-2.5.6-foss-2024a-CUDA-12.6.0.eb, PyTorch-Lightning-2.5.6-foss-2024a-CUDA-12.6.0.eb, bitsandbytes-0.46.1-foss-2024a-CUDA-12.6.0.eb and patches: ONNX-Runtime-1.23.2_gpu-package-name.patch, torchaudio-2.6.0_fix_tests_gpu.patch

github-actions · 2025-12-18T15:11:49Z

Updated software `CTranslate2-4.5.0-foss-2024a-CUDA-12.6.0.eb`

Diff against CTranslate2-4.5.0-foss-2023a-CUDA-12.6.0.eb

easybuild/easyconfigs/c/CTranslate2/CTranslate2-4.5.0-foss-2023a-CUDA-12.6.0.eb

diff --git a/easybuild/easyconfigs/c/CTranslate2/CTranslate2-4.5.0-foss-2023a-CUDA-12.6.0.eb b/easybuild/easyconfigs/c/CTranslate2/CTranslate2-4.5.0-foss-2024a-CUDA-12.6.0.eb
index a6376dba2e..4f54771108 100644
--- a/easybuild/easyconfigs/c/CTranslate2/CTranslate2-4.5.0-foss-2023a-CUDA-12.6.0.eb
+++ b/easybuild/easyconfigs/c/CTranslate2/CTranslate2-4.5.0-foss-2024a-CUDA-12.6.0.eb
@@ -7,7 +7,13 @@ versionsuffix = '-CUDA-%(cudaver)s'
 homepage = 'https://opennmt.net/CTranslate2/'
 description = "Fast inference engine for Transformer models."
 
-toolchain = {'name': 'foss', 'version': '2023a'}
+toolchain = {'name': 'foss', 'version': '2024a'}
+toolchainopts = {'extra_cxxflags': '-D_AMXTILEINTRIN_H_INCLUDED'}
+# '-D_AMXTILEINTRIN_H_INCLUDED' flag is required to avoid the following errors,
+# likely due to an incompatibility between the GCC 13.3.0 and NVCC CUDA 12.6.0:
+#   [...]/GCCcore/13.3.0/lib/gcc/x86_64-pc-linux-gnu/13.3.0/include/amxtileintrin.h(42):
+#     error: identifier "__builtin_ia32_ldtilecfg" is undefined
+# from ec: LAMMPS-29Aug2024_update2-foss-2024a-kokkos-CUDA-12.6.0.eb
 
 source_urls = ['https://github.com/OpenNMT/CTranslate2/archive/']
 sources = [{
@@ -27,20 +33,20 @@ checksums = [
 ]
 
 builddependencies = [
-    ('CMake', '3.26.3'),
-    ('pybind11', '2.11.1'),
+    ('CMake', '3.31.8'),
+    ('pybind11', '2.12.0'),
     ('cxxopts', '3.0.0', '', SYSTEM),
-    ('spdlog', '1.11.0'),
-    ('cpu_features', '0.9.0'),
+    ('spdlog', '1.15.3'),
+    ('cpu_features', '0.10.1'),
 ]
 
 dependencies = [
     ('CUDA', '12.6.0', '', SYSTEM),
-    ('Python', '3.11.3'),
-    ('SciPy-bundle', '2023.07'),
-    ('googletest', '1.13.0'),
-    ('PyYAML', '6.0'),
-    ('cuDNN', '9.5.0.50', versionsuffix, SYSTEM),
+    ('Python', '3.12.3'),
+    ('SciPy-bundle', '2024.05'),
+    ('googletest', '1.15.2'),
+    ('PyYAML', '6.0.2'),
+    ('cuDNN', '9.5.1.17', versionsuffix, SYSTEM),
 ]
 
 # make sure that CTranslate2 libraries are linked to FlexiBLAS, not OpenBLAS
@@ -76,8 +82,8 @@ sanity_check_paths = {
 
 sanity_check_commands = [
     "ct2-translator --help",
-    "python -c 'import ctranslate2'",
-    "python -m pip check",
+    "python -s -c 'import ctranslate2'",
+    "python -s -m pip check",
 ]
 
 moduleclass = 'ai'

Updated software `cpu_features-0.10.1-GCCcore-13.3.0.eb`

Diff against cpu_features-0.9.0-GCCcore-12.3.0.eb

easybuild/easyconfigs/c/cpu_features/cpu_features-0.9.0-GCCcore-12.3.0.eb

diff --git a/easybuild/easyconfigs/c/cpu_features/cpu_features-0.9.0-GCCcore-12.3.0.eb b/easybuild/easyconfigs/c/cpu_features/cpu_features-0.10.1-GCCcore-13.3.0.eb
index 1d352eaf0a..dcd84f2faf 100644
--- a/easybuild/easyconfigs/c/cpu_features/cpu_features-0.9.0-GCCcore-12.3.0.eb
+++ b/easybuild/easyconfigs/c/cpu_features/cpu_features-0.10.1-GCCcore-13.3.0.eb
@@ -3,21 +3,21 @@
 easyblock = 'CMakeMake'
 
 name = 'cpu_features'
-version = '0.9.0'
+version = '0.10.1'
 
 homepage = 'https://github.com/google/cpu_features'
 description = """A cross-platform C library to retrieve CPU features (such as available instructions) at runtime."""
 
-toolchain = {'name': 'GCCcore', 'version': '12.3.0'}
+toolchain = {'name': 'GCCcore', 'version': '13.3.0'}
 toolchainopts = {'pic': True}
 
 source_urls = ['https://github.com/google/cpu_features/archive/']
 sources = ['v%(version)s.tar.gz']
-checksums = ['bdb3484de8297c49b59955c3b22dba834401bc2df984ef5cfc17acbe69c5018e']
+checksums = ['52639b380fced11d738f8b151dbfee63fb94957731d07f1966c812e5b90cbad4']
 
 builddependencies = [
-    ('CMake', '3.26.3'),
-    ('binutils', '2.40'),
+    ('CMake', '3.31.8'),
+    ('binutils', '2.42'),
 ]
 
 modextrapaths = {MODULE_LOAD_ENV_HEADERS: 'include/cpu_features'}

Updated software `ONNX-Runtime-1.23.2-foss-2024a-CUDA-12.6.0.eb`

Diff against ONNX-Runtime-1.19.2-foss-2023a-CUDA-12.6.0.eb

easybuild/easyconfigs/o/ONNX-Runtime/ONNX-Runtime-1.19.2-foss-2023a-CUDA-12.6.0.eb

diff --git a/easybuild/easyconfigs/o/ONNX-Runtime/ONNX-Runtime-1.19.2-foss-2023a-CUDA-12.6.0.eb b/easybuild/easyconfigs/o/ONNX-Runtime/ONNX-Runtime-1.23.2-foss-2024a-CUDA-12.6.0.eb
index dc5c855893..59546f1660 100644
--- a/easybuild/easyconfigs/o/ONNX-Runtime/ONNX-Runtime-1.19.2-foss-2023a-CUDA-12.6.0.eb
+++ b/easybuild/easyconfigs/o/ONNX-Runtime/ONNX-Runtime-1.23.2-foss-2024a-CUDA-12.6.0.eb
@@ -1,7 +1,7 @@
 easyblock = 'PythonBundle'
 
 name = 'ONNX-Runtime'
-version = '1.19.2'
+version = '1.23.2'
 versionsuffix = '-CUDA-%(cudaver)s'
 
 homepage = 'https://onnxruntime.ai'
@@ -13,16 +13,52 @@ hardware, drivers, and operating systems, and provides optimal performance by
 leveraging hardware accelerators where applicable alongside graph optimizations
 and transforms."""
 
-toolchain = {'name': 'foss', 'version': '2023a'}
+toolchain = {'name': 'foss', 'version': '2024a'}
 
+builddependencies = [
+    ('CMake', '3.31.8'),
+    ('Ninja', '1.12.1'),
+    ('Eigen', '3.4.0'),
+]
 dependencies = [
     ('CUDA', '12.6.0', '', SYSTEM),
-    ('Python', '3.11.3'),
-    ('ONNX', '1.15.0'),
-    ('flatbuffers-python', '23.5.26'),
-    ('sympy', '1.12'),
+    ('cuDNN', '9.5.1.17', versionsuffix, SYSTEM),
+    ('Python', '3.12.3'),
+    ('SciPy-bundle', '2024.05'),
+    ('ONNX', '1.20.0'),
+    ('flatbuffers-python', '24.3.25'),
+    ('sympy', '1.13.3'),
+    ('Abseil', '20240722.0'),
 ]
 
+# CUSTOM BUILD
+local_onnx_buildcmd = ""
+# do not fetch and install Abseil - use installed one from the module
+# https://github.com/microsoft/onnxruntime/blob/v1.23.2/cmake/external/abseil-cpp.cmake#L39
+local_onnx_buildcmd += "sed -i 's/20250512/20240722/g' %(start_dir)s/cmake/external/abseil-cpp.cmake && "
+# the build command for onnx-runtime wheels
+# creates /build/Linux/Release/dist/onnxruntime-1.23.2-cp312-cp312-linux_x86_64.whl
+local_onnx_buildcmd += """
+./build.sh \
+--config Release \
+--update --build \
+--parallel %(parallel)s \
+--use_cuda \
+--cuda_home="$EBROOTCUDA" \
+--cudnn_home="$EBROOTCUDNN" \
+--cuda_version=%(cudashortver)s \
+--skip_tests \
+--build_shared_lib \
+--build_wheel \
+--skip_submodule_sync \
+--cmake_generator Ninja \
+--compile_no_warning_as_error \
+--cmake_extra_defines \
+absl_DIR=$EBROOTABSEIL/lib/cmake/absl \
+"CMAKE_CUDA_ARCHITECTURES=%(cuda_cc_cmake)s" \
+ONNXRUNTIME_VERSION=%(version)s
+"""
+
 exts_list = [
     ('humanfriendly', '10.0', {
         'checksums': ['6b0b831ce8f15f7300721aa49829fc4e83921a9a301cc7f606be6686a2288ddc'],
@@ -32,27 +68,24 @@ exts_list = [
     }),
     (name, version, {
         'modulename': 'onnxruntime',
-        'source_urls': ['http://pypi.python.org/packages/source/o/onnxruntime-gpu'],
-        'sources': ['onnxruntime_gpu-%(version)s-cp311-cp311-manylinux_2_27_%(arch)s.manylinux_2_28_%(arch)s.whl'],
-        'checksums': ['562fc7c755393eaad9751e56149339dd201ffbfdb3ef5f43ff21d0619ba9045f'],
+        'buildcmd': local_onnx_buildcmd,
+        'install_src': '%(start_dir)s/build/Linux/Release/dist/*.whl',
+        'source_urls': ['https://github.com/microsoft/onnxruntime/archive/'],
+        'sources': [{'download_filename': 'v%(version)s.tar.gz', 'filename': SOURCE_TAR_GZ}],
+        'patches': ['ONNX-Runtime-1.23.2_gpu-package-name.patch'],
+        'checksums': [
+            {'ONNX-Runtime-1.23.2.tar.gz': '99bcf964ce4e869d823c99b2294562a9050cbfa8e76ec81c8683cb3c7e19c2b4'},
+            {'ONNX-Runtime-1.23.2_gpu-package-name.patch':
+             'f0b80ae45878be371a1c5ef2b917dc34095af9351dfb379450b7be798f6d43bd'},
+        ],
     }),
 ]
 
-# Due to its name 'onnxruntime-gpu', this package does not fullfil requirements on 'onnxruntime' although it provides
-# the 'onnxruntime' python module. Fix this dependency issue in pip by creating a stub 'onnxruntime' dist-info folder
-_py_sitepkgs = '%(installdir)s/lib/python%(pyshortver)s/site-packages'
-postinstallcmds = [
-    "cp -r %s/onnxruntime{_gpu,}-%%(version)s.dist-info" % _py_sitepkgs,
-    "sed -i 's/onnxruntime.gpu/onnxruntime/g' %s/onnxruntime-%%(version)s.dist-info/{METADATA,RECORD}" % _py_sitepkgs,
-]
-
-_py_sitepkgs_dirs = ['onnxruntime', 'onnxruntime-%(version)s.dist-info', 'onnxruntime_gpu-%(version)s.dist-info']
+sanity_check_commands = ["python -c 'import onnxruntime; onnxruntime.get_available_providers()'"]
 
 sanity_check_paths = {
-    'files': [],
-    'dirs': ['lib/python%%(pyshortver)s/site-packages/%s' % x for x in _py_sitepkgs_dirs],
+    'files': ['bin/onnxruntime_test'],
+    'dirs': ['lib'],
 }
 
-options = {'modulename': 'onnxruntime'}
-
 moduleclass = 'devel'

Diff against ONNX-Runtime-1.19.2-foss-2023a-CUDA-12.1.1.eb

easybuild/easyconfigs/o/ONNX-Runtime/ONNX-Runtime-1.19.2-foss-2023a-CUDA-12.1.1.eb

diff --git a/easybuild/easyconfigs/o/ONNX-Runtime/ONNX-Runtime-1.19.2-foss-2023a-CUDA-12.1.1.eb b/easybuild/easyconfigs/o/ONNX-Runtime/ONNX-Runtime-1.23.2-foss-2024a-CUDA-12.6.0.eb
index 4623e4898d..59546f1660 100644
--- a/easybuild/easyconfigs/o/ONNX-Runtime/ONNX-Runtime-1.19.2-foss-2023a-CUDA-12.1.1.eb
+++ b/easybuild/easyconfigs/o/ONNX-Runtime/ONNX-Runtime-1.23.2-foss-2024a-CUDA-12.6.0.eb
@@ -1,7 +1,7 @@
 easyblock = 'PythonBundle'
 
 name = 'ONNX-Runtime'
-version = '1.19.2'
+version = '1.23.2'
 versionsuffix = '-CUDA-%(cudaver)s'
 
 homepage = 'https://onnxruntime.ai'
@@ -13,17 +13,51 @@ hardware, drivers, and operating systems, and provides optimal performance by
 leveraging hardware accelerators where applicable alongside graph optimizations
 and transforms."""
 
-toolchain = {'name': 'foss', 'version': '2023a'}
+toolchain = {'name': 'foss', 'version': '2024a'}
 
+builddependencies = [
+    ('CMake', '3.31.8'),
+    ('Ninja', '1.12.1'),
+    ('Eigen', '3.4.0'),
+]
 dependencies = [
-    ('CUDA', '12.1.1', '', SYSTEM),
-    ('Python', '3.11.3'),
-    ('ONNX', '1.15.0'),
-    ('flatbuffers-python', '23.5.26'),
-    ('sympy', '1.12'),
+    ('CUDA', '12.6.0', '', SYSTEM),
+    ('cuDNN', '9.5.1.17', versionsuffix, SYSTEM),
+    ('Python', '3.12.3'),
+    ('SciPy-bundle', '2024.05'),
+    ('ONNX', '1.20.0'),
+    ('flatbuffers-python', '24.3.25'),
+    ('sympy', '1.13.3'),
+    ('Abseil', '20240722.0'),
 ]
 
-local_whl_tmpl = 'onnxruntime_gpu-%%(version)s-cp311-cp311-manylinux_2_27_%s.manylinux_2_28_%s.whl'
+# CUSTOM BUILD
+local_onnx_buildcmd = ""
+# do not fetch and install Abseil - use installed one from the module
+# https://github.com/microsoft/onnxruntime/blob/v1.23.2/cmake/external/abseil-cpp.cmake#L39
+local_onnx_buildcmd += "sed -i 's/20250512/20240722/g' %(start_dir)s/cmake/external/abseil-cpp.cmake && "
+# the build command for onnx-runtime wheels
+# creates /build/Linux/Release/dist/onnxruntime-1.23.2-cp312-cp312-linux_x86_64.whl
+local_onnx_buildcmd += """
+./build.sh \
+--config Release \
+--update --build \
+--parallel %(parallel)s \
+--use_cuda \
+--cuda_home="$EBROOTCUDA" \
+--cudnn_home="$EBROOTCUDNN" \
+--cuda_version=%(cudashortver)s \
+--skip_tests \
+--build_shared_lib \
+--build_wheel \
+--skip_submodule_sync \
+--cmake_generator Ninja \
+--compile_no_warning_as_error \
+--cmake_extra_defines \
+absl_DIR=$EBROOTABSEIL/lib/cmake/absl \
+"CMAKE_CUDA_ARCHITECTURES=%(cuda_cc_cmake)s" \
+ONNXRUNTIME_VERSION=%(version)s
+"""
 
 exts_list = [
     ('humanfriendly', '10.0', {
@@ -33,31 +67,25 @@ exts_list = [
         'checksums': ['7c991aa71a4577af2f82600d8f8f3a89f936baeaf9b50a9c197da014e5bf16b0'],
     }),
     (name, version, {
-        'source_urls': ['http://pypi.python.org/packages/source/o/onnxruntime-gpu'],
-        'sources': [local_whl_tmpl % ('%(arch)s', '%(arch)s')],
-        'checksums': [{
-            local_whl_tmpl % ('x86_64', 'x86_64'):
-                '562fc7c755393eaad9751e56149339dd201ffbfdb3ef5f43ff21d0619ba9045f',
-        }],
         'modulename': 'onnxruntime',
+        'buildcmd': local_onnx_buildcmd,
+        'install_src': '%(start_dir)s/build/Linux/Release/dist/*.whl',
+        'source_urls': ['https://github.com/microsoft/onnxruntime/archive/'],
+        'sources': [{'download_filename': 'v%(version)s.tar.gz', 'filename': SOURCE_TAR_GZ}],
+        'patches': ['ONNX-Runtime-1.23.2_gpu-package-name.patch'],
+        'checksums': [
+            {'ONNX-Runtime-1.23.2.tar.gz': '99bcf964ce4e869d823c99b2294562a9050cbfa8e76ec81c8683cb3c7e19c2b4'},
+            {'ONNX-Runtime-1.23.2_gpu-package-name.patch':
+             'f0b80ae45878be371a1c5ef2b917dc34095af9351dfb379450b7be798f6d43bd'},
+        ],
     }),
 ]
 
-# Due to its name 'onnxruntime-gpu', this package does not fullfil requirements on 'onnxruntime' although it provides
-# the 'onnxruntime' python module. Fix this dependency issue in pip by creating a stub 'onnxruntime' dist-info folder
-_py_sitepkgs = '%(installdir)s/lib/python%(pyshortver)s/site-packages'
-postinstallcmds = [
-    "cp -r %s/onnxruntime{_gpu,}-%%(version)s.dist-info" % _py_sitepkgs,
-    "sed -i 's/onnxruntime.gpu/onnxruntime/g' %s/onnxruntime-%%(version)s.dist-info/{METADATA,RECORD}" % _py_sitepkgs,
-]
-
-_py_sitepkgs_dirs = ['onnxruntime', 'onnxruntime-%(version)s.dist-info', 'onnxruntime_gpu-%(version)s.dist-info']
+sanity_check_commands = ["python -c 'import onnxruntime; onnxruntime.get_available_providers()'"]
 
 sanity_check_paths = {
-    'files': [],
-    'dirs': ['lib/python%%(pyshortver)s/site-packages/%s' % x for x in _py_sitepkgs_dirs],
+    'files': ['bin/onnxruntime_test'],
+    'dirs': ['lib'],
 }
 
-options = {'modulename': 'onnxruntime'}
-
 moduleclass = 'devel'

Updated software `pyannote.audio-3.4.0-foss-2024a-CUDA-12.6.0.eb`

Diff against pyannote.audio-3.3.2-foss-2023a.eb

easybuild/easyconfigs/p/pyannote.audio/pyannote.audio-3.3.2-foss-2023a.eb

diff --git a/easybuild/easyconfigs/p/pyannote.audio/pyannote.audio-3.3.2-foss-2023a.eb b/easybuild/easyconfigs/p/pyannote.audio/pyannote.audio-3.4.0-foss-2024a-CUDA-12.6.0.eb
index 2475c5d2ae..09fe6b4980 100644
--- a/easybuild/easyconfigs/p/pyannote.audio/pyannote.audio-3.3.2-foss-2023a.eb
+++ b/easybuild/easyconfigs/p/pyannote.audio/pyannote.audio-3.4.0-foss-2024a-CUDA-12.6.0.eb
@@ -1,40 +1,50 @@
 easyblock = 'PythonBundle'
 
 name = 'pyannote.audio'
-version = '3.3.2'
-local_pytorch_version = '2.1.2'
+version = '3.4.0'
+versionsuffix = '-CUDA-%(cudaver)s'
+local_pytorch_version = '2.7.1'
 
 homepage = 'https://github.com/pyannote'
 description = "pyannote is an open-source toolkit for speaker diarization."
 
-toolchain = {'name': 'foss', 'version': '2023a'}
+toolchain = {'name': 'foss', 'version': '2024a'}
 
 builddependencies = [
-    ('PDM', '2.12.4'),
+    ('PDM', '2.18.2'),
     ('Java', '21', '', SYSTEM),
+    ('CMake', '3.31.8'),
 ]
 dependencies = [
-    ('Python', '3.11.3'),
-    ('PyTorch-bundle', '2.1.2'),
-    ('Lightning', '2.2.1'),
-    ('SciPy-bundle', '2023.07'),
-    ('tensorboardX', '2.6.2.2'),
-    ('scikit-learn', '1.3.1'),
-    ('Optuna', '3.5.0'),
-    ('einops', '0.7.0'),
-    ('matplotlib', '3.7.2'),
-    ('ruamel.yaml', '0.17.32'),
-    ('cffi', '1.15.1'),
-    ('libsndfile', '1.2.2'),
+    ('CUDA', '12.6.0', '', SYSTEM),
+    ('Python', '3.12.3'),
+    ('PyTorch', local_pytorch_version, versionsuffix),
+    ('torchaudio', local_pytorch_version, versionsuffix),
+    ('Lightning', '2.5.6', versionsuffix),
+    ('SciPy-bundle', '2024.05'),
+    ('tensorboardX', '2.6.4'),
+    ('Optuna', '4.1.0'),
+    ('einops', '0.8.1'),
+    ('ruamel.yaml', '0.18.6'),
+    ('huggingface_hub', '0.34.4'),
+    ('scikit-learn', '1.5.2'),
+    ('tqdm', '4.66.5'),
+    ('PyYAML', '6.0.2'),
+    ('SentencePiece', '0.2.1'),
 ]
 
+# fix versioneer.py for python 3.12
+local_preinstallopts_versioneer = (
+    "sed -i -e 's/SafeConfigParser/ConfigParser/g' -e 's/readfp/read_file/g' versioneer.py && "
+)
+
 exts_list = [
     ('primePy', '1.3', {
         'modulename': 'primePy',
         'checksums': ['25fd7e25344b0789a5984c75d89f054fcf1f180bef20c998e4befbac92de4669'],
     }),
-    ('typer', '0.9.4', {
-        'checksums': ['f714c2d90afae3a7929fcd72a3abb08df305e1ff61719381384211c4070af57f'],
+    ('typer', '0.20.0', {
+        'checksums': ['1aaf6494031793e4876fb0bacfa6a912b551cf43c1e63c800df8b1a866720c37'],
     }),
     ('torch_pitch_shift', '1.2.5', {
         'checksums': ['6e1c7531f08d0f407a4c55e5ff8385a41355c5c5d27ab7fa08632e51defbd0ed'],
@@ -42,34 +52,23 @@ exts_list = [
     ('julius', '0.2.7', {
         'checksums': ['3c0f5f5306d7d6016fcc95196b274cae6f07e2c9596eed314e4e7641554fbb08'],
     }),
-    ('HyperPyYAML', '1.2.2', {
-        'checksums': ['bdb734210d18770a262f500fe5755c7a44a5d3b91521b06e24f7a00a36ee0f87'],
-    }),
-    ('antlr4-python3-runtime', '4.9.3', {
-        'modulename': 'antlr4',
-        'checksums': ['f224469b4168294902bb1efa80a8bf7855f24c99aef99cbefc1bcd3cce77881b'],
-    }),
-    ('torchmetrics', '1.5.2', {
-        'checksums': ['2d0e4957af0ea76438d2779fe1a626d8cba6cda8607eadb54267598153e7ea63'],
-    }),
     ('torch_audiomentations', '0.12.0', {
         'checksums': ['b02d4c5eb86376986a53eb405cca5e34f370ea9284411237508e720c529f7888'],
     }),
-    ('speechbrain', '1.0.2', {
-        'checksums': ['b293d665161d7cd34caa3ee8966acce6c43b611b7fc9cd0b24637ce871bd7a73'],
+    ('HyperPyYAML', '1.2.2', {
+        'checksums': ['bdb734210d18770a262f500fe5755c7a44a5d3b91521b06e24f7a00a36ee0f87'],
     }),
-    ('soundfile', '0.12.1', {
-        'checksums': ['e8e1017b2cf1dda767aef19d2fd9ee5ebe07e050d430f77a0a7c66ba08b8cdae'],
+    ('speechbrain', '1.0.3', {
+        'checksums': ['fcab3c6e90012cecb1eed40ea235733b550137e73da6bfa2340ba191ec714052'],
     }),
-    ('semver', '3.0.2', {
-        'checksums': ['6253adb39c70f6e51afed2fa7152bcd414c411286088fb4b9effb133885ab4cc'],
+    ('soundfile', '0.13.1', {
+        'checksums': ['b2c68dab1e30297317080a5b43df57e302584c49e2942defdde0acccc53f0e5b'],
     }),
-    ('pytorch-metric-learning', '2.8.1', {
-        'checksums': ['fcc4d3b4a805e5fce25fb2e67505c47ba6fea0563fc09c5655ea1f08d1e8ed93'],
+    ('semver', '3.0.4', {
+        'checksums': ['afc7d8c584a5ed0a11033af086e8af226a9c0b206f313e0301f8dd7b6b589602'],
     }),
-    ('huggingface-hub', '0.26.5', {
-        'sources': ['huggingface_hub-%(version)s.tar.gz'],
-        'checksums': ['1008bd18f60bfb65e8dbc0a97249beeeaa8c99d3c2fa649354df9fa5a13ed83b'],
+    ('pytorch-metric-learning', '2.9.0', {
+        'checksums': ['27a626caf5e2876a0fd666605a78cb67ef7597e25d7a68c18053dd503830701f'],
     }),
     ('asteroid-filterbanks', '0.4.0', {
         'checksums': ['415f89d1dcf2b13b35f03f7a9370968ac4e6fa6800633c522dac992b283409b9'],
@@ -78,22 +77,24 @@ exts_list = [
         'checksums': ['d5d4b6d29955cc50ad50c46dc269bcd92c6e00f5f90d23ab5fee7bfca4ba4cc7'],
     }),
     ('pyannote.core', '5.0.0', {
+        'preinstallopts': local_preinstallopts_versioneer,
         'checksums': ['1a55bcc8bd680ba6be5fa53efa3b6f3d2cdd67144c07b6b4d8d66d5cb0d2096f'],
     }),
-    ('pyannote.database', '5.0.1', {
-        'checksums': ['980bdc910615380a17d8bd82805f9770006d050b2c7a946a520ade4627192681'],
+    ('pyannote.database', '5.1.3', {
+        'checksums': ['0eaf64c1cc506718de60d2d702f1359b1ae7ff252ee3e4799f1c5e378cd52c31'],
     }),
     ('pyannote.metrics', '3.2.1', {
+        'preinstallopts': local_preinstallopts_versioneer,
         'checksums': ['08024255a3550e96a8e9da4f5f4af326886548480de891414567c8900920ee5c'],
     }),
     ('pyannote.pipeline', '3.0.1', {
+        'preinstallopts': local_preinstallopts_versioneer,
         'checksums': ['021794e26a2cf5d8fb5bb1835951e71f5fac33eb14e23dfb7468e16b1b805151'],
     }),
     (name, version, {
-        # unpin torchaudio version to let it works with v2.1.2
-        'preinstallopts': "sed -i 's/torchaudio >= 2.2.0/torchaudio/' requirements.txt && ",
-        'source_tmpl': 'pyannote_audio-%(version)s.tar.gz',
-        'checksums': ['b2115e86b0db5faedb9f36ee1a150cebd07f7758e65e815accdac1a12ca9c777'],
+        'source_urls': ['https://github.com/pyannote/pyannote-audio/archive/'],
+        'sources': [{'download_filename': '%(version)s.tar.gz', 'filename': SOURCE_TAR_GZ}],
+        'checksums': ['9fc8b4c96457733945d399559c328e07237a037ab7151443a81fe8473ea104ab'],
     }),
 ]

Diff against pyannote.audio-3.3.2-foss-2023a-CUDA-12.1.1.eb

easybuild/easyconfigs/p/pyannote.audio/pyannote.audio-3.3.2-foss-2023a-CUDA-12.1.1.eb

diff --git a/easybuild/easyconfigs/p/pyannote.audio/pyannote.audio-3.3.2-foss-2023a-CUDA-12.1.1.eb b/easybuild/easyconfigs/p/pyannote.audio/pyannote.audio-3.4.0-foss-2024a-CUDA-12.6.0.eb
index 27f3e3ea4e..09fe6b4980 100644
--- a/easybuild/easyconfigs/p/pyannote.audio/pyannote.audio-3.3.2-foss-2023a-CUDA-12.1.1.eb
+++ b/easybuild/easyconfigs/p/pyannote.audio/pyannote.audio-3.4.0-foss-2024a-CUDA-12.6.0.eb
@@ -1,42 +1,50 @@
 easyblock = 'PythonBundle'
 
 name = 'pyannote.audio'
-version = '3.3.2'
+version = '3.4.0'
 versionsuffix = '-CUDA-%(cudaver)s'
-local_pytorch_version = '2.1.2'
+local_pytorch_version = '2.7.1'
 
 homepage = 'https://github.com/pyannote'
 description = "pyannote is an open-source toolkit for speaker diarization."
 
-toolchain = {'name': 'foss', 'version': '2023a'}
+toolchain = {'name': 'foss', 'version': '2024a'}
 
 builddependencies = [
-    ('PDM', '2.12.4'),
+    ('PDM', '2.18.2'),
     ('Java', '21', '', SYSTEM),
+    ('CMake', '3.31.8'),
 ]
 dependencies = [
-    ('CUDA', '12.1.1', '', SYSTEM),
-    ('Python', '3.11.3'),
-    ('PyTorch-bundle', '2.1.2', versionsuffix),
-    ('Lightning', '2.2.1', versionsuffix),
-    ('SciPy-bundle', '2023.07'),
-    ('tensorboardX', '2.6.2.2'),
-    ('scikit-learn', '1.3.1'),
-    ('Optuna', '3.5.0'),
-    ('einops', '0.7.0'),
-    ('matplotlib', '3.7.2'),
-    ('ruamel.yaml', '0.17.32'),
-    ('cffi', '1.15.1'),
-    ('libsndfile', '1.2.2'),
+    ('CUDA', '12.6.0', '', SYSTEM),
+    ('Python', '3.12.3'),
+    ('PyTorch', local_pytorch_version, versionsuffix),
+    ('torchaudio', local_pytorch_version, versionsuffix),
+    ('Lightning', '2.5.6', versionsuffix),
+    ('SciPy-bundle', '2024.05'),
+    ('tensorboardX', '2.6.4'),
+    ('Optuna', '4.1.0'),
+    ('einops', '0.8.1'),
+    ('ruamel.yaml', '0.18.6'),
+    ('huggingface_hub', '0.34.4'),
+    ('scikit-learn', '1.5.2'),
+    ('tqdm', '4.66.5'),
+    ('PyYAML', '6.0.2'),
+    ('SentencePiece', '0.2.1'),
 ]
 
+# fix versioneer.py for python 3.12
+local_preinstallopts_versioneer = (
+    "sed -i -e 's/SafeConfigParser/ConfigParser/g' -e 's/readfp/read_file/g' versioneer.py && "
+)
+
 exts_list = [
     ('primePy', '1.3', {
         'modulename': 'primePy',
         'checksums': ['25fd7e25344b0789a5984c75d89f054fcf1f180bef20c998e4befbac92de4669'],
     }),
-    ('typer', '0.9.4', {
-        'checksums': ['f714c2d90afae3a7929fcd72a3abb08df305e1ff61719381384211c4070af57f'],
+    ('typer', '0.20.0', {
+        'checksums': ['1aaf6494031793e4876fb0bacfa6a912b551cf43c1e63c800df8b1a866720c37'],
     }),
     ('torch_pitch_shift', '1.2.5', {
         'checksums': ['6e1c7531f08d0f407a4c55e5ff8385a41355c5c5d27ab7fa08632e51defbd0ed'],
@@ -44,34 +52,23 @@ exts_list = [
     ('julius', '0.2.7', {
         'checksums': ['3c0f5f5306d7d6016fcc95196b274cae6f07e2c9596eed314e4e7641554fbb08'],
     }),
-    ('HyperPyYAML', '1.2.2', {
-        'checksums': ['bdb734210d18770a262f500fe5755c7a44a5d3b91521b06e24f7a00a36ee0f87'],
-    }),
-    ('antlr4-python3-runtime', '4.9.3', {
-        'modulename': 'antlr4',
-        'checksums': ['f224469b4168294902bb1efa80a8bf7855f24c99aef99cbefc1bcd3cce77881b'],
-    }),
-    ('torchmetrics', '1.5.2', {
-        'checksums': ['2d0e4957af0ea76438d2779fe1a626d8cba6cda8607eadb54267598153e7ea63'],
-    }),
     ('torch_audiomentations', '0.12.0', {
         'checksums': ['b02d4c5eb86376986a53eb405cca5e34f370ea9284411237508e720c529f7888'],
     }),
-    ('speechbrain', '1.0.2', {
-        'checksums': ['b293d665161d7cd34caa3ee8966acce6c43b611b7fc9cd0b24637ce871bd7a73'],
+    ('HyperPyYAML', '1.2.2', {
+        'checksums': ['bdb734210d18770a262f500fe5755c7a44a5d3b91521b06e24f7a00a36ee0f87'],
     }),
-    ('soundfile', '0.12.1', {
-        'checksums': ['e8e1017b2cf1dda767aef19d2fd9ee5ebe07e050d430f77a0a7c66ba08b8cdae'],
+    ('speechbrain', '1.0.3', {
+        'checksums': ['fcab3c6e90012cecb1eed40ea235733b550137e73da6bfa2340ba191ec714052'],
     }),
-    ('semver', '3.0.2', {
-        'checksums': ['6253adb39c70f6e51afed2fa7152bcd414c411286088fb4b9effb133885ab4cc'],
+    ('soundfile', '0.13.1', {
+        'checksums': ['b2c68dab1e30297317080a5b43df57e302584c49e2942defdde0acccc53f0e5b'],
     }),
-    ('pytorch-metric-learning', '2.8.1', {
-        'checksums': ['fcc4d3b4a805e5fce25fb2e67505c47ba6fea0563fc09c5655ea1f08d1e8ed93'],
+    ('semver', '3.0.4', {
+        'checksums': ['afc7d8c584a5ed0a11033af086e8af226a9c0b206f313e0301f8dd7b6b589602'],
     }),
-    ('huggingface-hub', '0.26.5', {
-        'sources': ['huggingface_hub-%(version)s.tar.gz'],
-        'checksums': ['1008bd18f60bfb65e8dbc0a97249beeeaa8c99d3c2fa649354df9fa5a13ed83b'],
+    ('pytorch-metric-learning', '2.9.0', {
+        'checksums': ['27a626caf5e2876a0fd666605a78cb67ef7597e25d7a68c18053dd503830701f'],
     }),
     ('asteroid-filterbanks', '0.4.0', {
         'checksums': ['415f89d1dcf2b13b35f03f7a9370968ac4e6fa6800633c522dac992b283409b9'],
@@ -80,21 +77,24 @@ exts_list = [
         'checksums': ['d5d4b6d29955cc50ad50c46dc269bcd92c6e00f5f90d23ab5fee7bfca4ba4cc7'],
     }),
     ('pyannote.core', '5.0.0', {
+        'preinstallopts': local_preinstallopts_versioneer,
         'checksums': ['1a55bcc8bd680ba6be5fa53efa3b6f3d2cdd67144c07b6b4d8d66d5cb0d2096f'],
     }),
-    ('pyannote.database', '5.0.1', {
-        'checksums': ['980bdc910615380a17d8bd82805f9770006d050b2c7a946a520ade4627192681'],
+    ('pyannote.database', '5.1.3', {
+        'checksums': ['0eaf64c1cc506718de60d2d702f1359b1ae7ff252ee3e4799f1c5e378cd52c31'],
     }),
     ('pyannote.metrics', '3.2.1', {
+        'preinstallopts': local_preinstallopts_versioneer,
         'checksums': ['08024255a3550e96a8e9da4f5f4af326886548480de891414567c8900920ee5c'],
     }),
     ('pyannote.pipeline', '3.0.1', {
+        'preinstallopts': local_preinstallopts_versioneer,
         'checksums': ['021794e26a2cf5d8fb5bb1835951e71f5fac33eb14e23dfb7468e16b1b805151'],
     }),
     (name, version, {
-        'preinstallopts': "sed -i 's/torchaudio >= 2.2.0/torchaudio/' requirements.txt && ",
-        'source_tmpl': 'pyannote_audio-%(version)s.tar.gz',
-        'checksums': ['b2115e86b0db5faedb9f36ee1a150cebd07f7758e65e815accdac1a12ca9c777'],
+        'source_urls': ['https://github.com/pyannote/pyannote-audio/archive/'],
+        'sources': [{'download_filename': '%(version)s.tar.gz', 'filename': SOURCE_TAR_GZ}],
+        'checksums': ['9fc8b4c96457733945d399559c328e07237a037ab7151443a81fe8473ea104ab'],
     }),
 ]

Updated software `torchaudio-2.7.1-foss-2024a-CUDA-12.6.0.eb`

Diff against torchaudio-2.6.0-foss-2024a.eb

easybuild/easyconfigs/t/torchaudio/torchaudio-2.6.0-foss-2024a.eb

diff --git a/easybuild/easyconfigs/t/torchaudio/torchaudio-2.6.0-foss-2024a.eb b/easybuild/easyconfigs/t/torchaudio/torchaudio-2.7.1-foss-2024a-CUDA-12.6.0.eb
index ec6c17a46a..30e08c913e 100644
--- a/easybuild/easyconfigs/t/torchaudio/torchaudio-2.6.0-foss-2024a.eb
+++ b/easybuild/easyconfigs/t/torchaudio/torchaudio-2.7.1-foss-2024a-CUDA-12.6.0.eb
@@ -1,25 +1,26 @@
 easyblock = 'PythonBundle'
 
 name = 'torchaudio'
-version = '2.6.0'
+version = '2.7.1'
+versionsuffix = '-CUDA-%(cudaver)s'
 
 homepage = 'https://github.com/pytorch/audio'
-description = """ Data manipulation and transformation for audio signal
-processing, powered by PyTorch """
+description = """Data manipulation and transformation for audio signal
+processing, powered by PyTorch."""
 
 toolchain = {'name': 'foss', 'version': '2024a'}
 
 builddependencies = [
-    ('CMake', '3.29.3'),
+    ('CMake', '3.31.8'),
     ('Ninja', '1.12.1'),
     ('parameterized', '0.9.0'),  # for tests
     ('scikit-learn', '1.5.2'),  # for tests
     ('librosa', '0.10.2.post1'),  # for tests
 ]
-
 dependencies = [
+    ('CUDA', '12.6.0', '', SYSTEM),
     ('Python', '3.12.3'),
-    ('PyTorch', version),
+    ('PyTorch', version, versionsuffix),
     ('FFmpeg', '7.0.2'),
     ('SoX', '14.4.2'),
 ]
@@ -27,7 +28,9 @@ dependencies = [
 local_preinstall_opts = ' '.join([
     'USE_SYSTEM_LIBS=1',
     'USE_OPENMP=1',
-    'USE_CUDA=0',
+    'USE_CUDA=1',
+    'USE_CUDNN=1',
+    'TORCH_CUDA_ARCH_LIST="%(cuda_cc_semicolon_sep)s"',
     'USE_FFMPEG=1', 'FFMPEG_ROOT="$EBROOTFFMPEG"',
     'CMAKE_BUILD_PARALLEL_LEVEL=%(parallel)s',
 ])
@@ -36,16 +39,18 @@ exts_list = [
     (name, version, {
         'installopts': '-v',
         'patches': [
-            'torchaudio-%(version)s_use_ffmpeg7.patch',
-            'torchaudio-%(version)s_fix_tests_cpu.patch',
+            'torchaudio-2.6.0_use_ffmpeg7.patch',
+            'torchaudio-2.6.0_fix_tests_gpu.patch',
         ],
         'preinstallopts': (
             'unset BUILD_VERSION && rm -r third_party/{sox,ffmpeg/multi}; '  # runs twice when testinstall
-            'BUILD_SOX=0 ') + local_preinstall_opts,
+            'BUILD_SOX=0 '
+        ) + local_preinstall_opts,
         'source_urls': ['https://github.com/pytorch/audio/archive'],
-        'sources': [{'download_filename': 'v%(version)s.tar.gz', 'filename': '%(name)s-%(version)s.tar.gz'}],
+        'sources': [{'download_filename': 'v%(version)s.tar.gz', 'filename': SOURCE_TAR_GZ}],
         'runtest': (
             'export OMP_NUM_THREADS=%(parallel)s && '
+            'export NVIDIA_TF32_OVERRIDE=0 && '
             'pytest test/torchaudio_unittest/'
             ' -k "not TestProcessPoolExecutor"'  # hang maybe related https://github.com/pytorch/audio/issues/1021
             '" and not FilterGraphWithCudaAccel"'  # requires FFmpeg with CUDA support
@@ -53,13 +58,14 @@ exts_list = [
             '" and not test_dup_hw_acel"'  # requires special render device permissions
             '" and not test_h264_cuvid"'  # requires special render device permissions
             '" and not test_hevc_cuvid"'  # requires special render device permissions
+            '" and not TestAutogradLfilterCUDA"'  # requires PyTorch’s nondeterministic algorithms mode
         ),
         'testinstall': True,
         'checksums': [
-            {'torchaudio-2.6.0.tar.gz': '3335d8fcf58c26acf3c628d751103b59226e01c91847ce56efb2a4e7ae8351ef'},
+            {'torchaudio-2.7.1.tar.gz': 'fc8159476d1b3b5978d5e66746fc34be168170800ff4c5e356433d8c9c57cbea'},
             {'torchaudio-2.6.0_use_ffmpeg7.patch': '1a2f7505efee9852ef393e6f4583cef209ad302db241171bf41be8d4a88920bd'},
-            {'torchaudio-2.6.0_fix_tests_cpu.patch':
-             'f44d1ab12fd47dfd83ef423bd1a00f668285f871b4205cbcfae5aa5c110894d5'},
+            {'torchaudio-2.6.0_fix_tests_gpu.patch':
+             '3a728b033c4d568934a127ac1d573a385eeea8326e111ecc5aafaff7b5407d59'},
         ],
     }),
 ]

Diff against torchaudio-0.12.0-foss-2022a-PyTorch-1.12.0-CUDA-11.7.0.eb

easybuild/easyconfigs/t/torchaudio/torchaudio-0.12.0-foss-2022a-PyTorch-1.12.0-CUDA-11.7.0.eb

diff --git a/easybuild/easyconfigs/t/torchaudio/torchaudio-0.12.0-foss-2022a-PyTorch-1.12.0-CUDA-11.7.0.eb b/easybuild/easyconfigs/t/torchaudio/torchaudio-2.7.1-foss-2024a-CUDA-12.6.0.eb
index 87dc789392..30e08c913e 100644
--- a/easybuild/easyconfigs/t/torchaudio/torchaudio-0.12.0-foss-2022a-PyTorch-1.12.0-CUDA-11.7.0.eb
+++ b/easybuild/easyconfigs/t/torchaudio/torchaudio-2.7.1-foss-2024a-CUDA-12.6.0.eb
@@ -1,45 +1,73 @@
-easyblock = 'PythonPackage'
+easyblock = 'PythonBundle'
 
 name = 'torchaudio'
-version = '0.12.0'
-local_pytorch_version = '1.12.0'
-versionsuffix = '-PyTorch-' + local_pytorch_version + '-CUDA-%(cudaver)s'
+version = '2.7.1'
+versionsuffix = '-CUDA-%(cudaver)s'
 
 homepage = 'https://github.com/pytorch/audio'
-description = """ Data manipulation and transformation for audio signal
-processing, powered by PyTorch """
+description = """Data manipulation and transformation for audio signal
+processing, powered by PyTorch."""
 
-toolchain = {'name': 'foss', 'version': '2022a'}
-
-sources = [{
-    'filename': '%(name)s-%(version)s.tar.gz',
-    'git_config': {
-        'url': 'https://github.com/pytorch',
-        'repo_name': 'audio',
-        'tag': 'v%(version)s',
-        'recursive': True,
-        'keep_git_dir': True,
-    }
-}]
-
-checksums = [None]
+toolchain = {'name': 'foss', 'version': '2024a'}
 
 builddependencies = [
-    ('binutils', '2.38'),
-    ('CMake', '3.23.1'),
-    ('Ninja', '1.10.2'),
+    ('CMake', '3.31.8'),
+    ('Ninja', '1.12.1'),
+    ('parameterized', '0.9.0'),  # for tests
+    ('scikit-learn', '1.5.2'),  # for tests
+    ('librosa', '0.10.2.post1'),  # for tests
 ]
-
 dependencies = [
-    ('CUDA', '11.7.0', '', SYSTEM),
-    ('Python', '3.10.4'),
-    ('SciPy-bundle', '2022.05'),
-    ('PyTorch', local_pytorch_version, '-CUDA-%(cudaver)s'),
+    ('CUDA', '12.6.0', '', SYSTEM),
+    ('Python', '3.12.3'),
+    ('PyTorch', version, versionsuffix),
+    ('FFmpeg', '7.0.2'),
     ('SoX', '14.4.2'),
-    ('FFmpeg', '4.4.2'),
 ]
 
-preinstallopts = 'USE_FFMPEG=1'
-installopts = '--no-use-pep517'
+local_preinstall_opts = ' '.join([
+    'USE_SYSTEM_LIBS=1',
+    'USE_OPENMP=1',
+    'USE_CUDA=1',
+    'USE_CUDNN=1',
+    'TORCH_CUDA_ARCH_LIST="%(cuda_cc_semicolon_sep)s"',
+    'USE_FFMPEG=1', 'FFMPEG_ROOT="$EBROOTFFMPEG"',
+    'CMAKE_BUILD_PARALLEL_LEVEL=%(parallel)s',
+])
+
+exts_list = [
+    (name, version, {
+        'installopts': '-v',
+        'patches': [
+            'torchaudio-2.6.0_use_ffmpeg7.patch',
+            'torchaudio-2.6.0_fix_tests_gpu.patch',
+        ],
+        'preinstallopts': (
+            'unset BUILD_VERSION && rm -r third_party/{sox,ffmpeg/multi}; '  # runs twice when testinstall
+            'BUILD_SOX=0 '
+        ) + local_preinstall_opts,
+        'source_urls': ['https://github.com/pytorch/audio/archive'],
+        'sources': [{'download_filename': 'v%(version)s.tar.gz', 'filename': SOURCE_TAR_GZ}],
+        'runtest': (
+            'export OMP_NUM_THREADS=%(parallel)s && '
+            'export NVIDIA_TF32_OVERRIDE=0 && '
+            'pytest test/torchaudio_unittest/'
+            ' -k "not TestProcessPoolExecutor"'  # hang maybe related https://github.com/pytorch/audio/issues/1021
+            '" and not FilterGraphWithCudaAccel"'  # requires FFmpeg with CUDA support
+            '" and not kaldi_io_test"'  # requires kaldi_io
+            '" and not test_dup_hw_acel"'  # requires special render device permissions
+            '" and not test_h264_cuvid"'  # requires special render device permissions
+            '" and not test_hevc_cuvid"'  # requires special render device permissions
+            '" and not TestAutogradLfilterCUDA"'  # requires PyTorch’s nondeterministic algorithms mode
+        ),
+        'testinstall': True,
+        'checksums': [
+            {'torchaudio-2.7.1.tar.gz': 'fc8159476d1b3b5978d5e66746fc34be168170800ff4c5e356433d8c9c57cbea'},
+            {'torchaudio-2.6.0_use_ffmpeg7.patch': '1a2f7505efee9852ef393e6f4583cef209ad302db241171bf41be8d4a88920bd'},
+            {'torchaudio-2.6.0_fix_tests_gpu.patch':
+             '3a728b033c4d568934a127ac1d573a385eeea8326e111ecc5aafaff7b5407d59'},
+        ],
+    }),
+]
 
-moduleclass = 'tools'
+moduleclass = 'ai'

Updated software `WhisperX-3.7.4-foss-2024a-CUDA-12.6.0.eb`

Diff against WhisperX-3.3.1-foss-2023a-CUDA-12.1.1.eb

easybuild/easyconfigs/w/WhisperX/WhisperX-3.3.1-foss-2023a-CUDA-12.1.1.eb

diff --git a/easybuild/easyconfigs/w/WhisperX/WhisperX-3.3.1-foss-2023a-CUDA-12.1.1.eb b/easybuild/easyconfigs/w/WhisperX/WhisperX-3.7.4-foss-2024a-CUDA-12.6.0.eb
index 8fbe9b241a..2ab104e110 100644
--- a/easybuild/easyconfigs/w/WhisperX/WhisperX-3.3.1-foss-2023a-CUDA-12.1.1.eb
+++ b/easybuild/easyconfigs/w/WhisperX/WhisperX-3.7.4-foss-2024a-CUDA-12.6.0.eb
@@ -1,40 +1,49 @@
 easyblock = 'PythonBundle'
 
 name = 'WhisperX'
-version = '3.3.1'
+version = '3.7.4'
 versionsuffix = '-CUDA-%(cudaver)s'
-local_pytorch_version = '2.1.2'
+local_pytorch_version = '2.7.1'
 
 homepage = 'https://github.com/m-bain/whisperx'
 description = "Automatic Speech Recognition with Word-level Timestamps (& Diarization)."
 
-toolchain = {'name': 'foss', 'version': '2023a'}
+toolchain = {'name': 'foss', 'version': '2024a'}
 
+builddependencies = [('Cython', '3.0.10')]
 dependencies = [
-    ('CUDA', '12.1.1', '', SYSTEM),
-    ('Python', '3.11.3'),
-    ('SciPy-bundle', '2023.07'),
-    ('PyTorch-bundle', local_pytorch_version, versionsuffix),
-    ('NLTK', '3.8.1'),
-    ('Transformers', '4.39.3'),
-    ('FFmpeg', '6.0'),
-    ('PyAV', '11.0.0'),
-    ('ONNX-Runtime', '1.19.2', versionsuffix),
-    ('pyannote.audio', '3.3.2', versionsuffix),
+    ('CUDA', '12.6.0', '', SYSTEM),
+    ('Python', '3.12.3'),
+    ('PyTorch', '2.7.1', versionsuffix),
+    ('SciPy-bundle', '2024.05'),
+    ('CTranslate2', '4.5.0', versionsuffix),
+    ('NLTK', '3.9.1'),
+    ('Transformers', '4.55.0'),
+    ('FFmpeg', '7.0.2'),
+    ('ONNX-Runtime', '1.23.2', versionsuffix),
+    ('pyannote.audio', '3.4.0', versionsuffix),
 ]
 
+# unpin versions of dependencies in WhisperX
+local_whisperx_preinstallopts = (
+    "sed -E -i "
+    r""" '/^[[:space:]]*dependencies[[:space:]]*=[[:space:]]*[[]/"""
+    r""",/^[[:space:]]*]/{s/"([^"<>!=~;[:space:]]+)[^"]*"/"\1"/g}' """
+    "pyproject.toml && "
+)
+
 exts_list = [
-    ('faster-whisper', '1.1.0', {
-        'checksums': ['cea4bba5d4527173fdbacafa56f2ffb17dd322688f6c3fdf5fd7b6b6c193ce17'],
+    ('faster-whisper', '1.2.0', {
+        'checksums': ['56b20d616a575049a79f33b04f02db0868ce38c5d057a0b816d36ca59a6d2598'],
     }),
-    ('ctranslate2', '4.4.0', {
-        'source_tmpl': '%(name)s-%(version)s-cp311-cp311-manylinux_2_17_x86_64.manylinux2014_x86_64.whl',
-        'checksums': ['7653f1f22da29bbb0ba5704ac27b1a3fbde1f34b2b65a703dd20f7e2331eb03c'],
+    ('av', '14.0.1', {
+        'checksums': ['2b0a17301af469ddaea46b5c1c982df1b7b5de8bc6c94cdc98cad4a67178c82a'],
     }),
     (name, version, {
+        'preinstallopts': local_whisperx_preinstallopts,
         'source_urls': ['https://github.com/m-bain/whisperX/archive/'],
         'sources': [{'download_filename': 'v%(version)s.tar.gz', 'filename': SOURCE_TAR_GZ}],
-        'checksums': ['37136d07f991d684744129a98096ca0f300a5098a1a387aa0b6c7e5eb815a2aa'],
+        'checksums': ['ffe9ce94d8895e7ba6030f3cc35a357788a458462051422dda6428d2f95324a7'],
     }),
 ]

pavelToman · 2025-12-18T19:25:43Z

Test report by @pavelToman
FAILED
Build succeeded for 26 out of 31 (total: 3 hours 45 mins 16 secs) (10 easyconfigs in total)
node3900.accelgor.os - Linux RHEL 9.6, x86_64, AMD EPYC 7413 24-Core Processor, 1 x NVIDIA NVIDIA A100-SXM4-80GB, 580.95.05, Python 3.9.21
See https://gist.github.com/pavelToman/e4ef4fde73ce1f9e802e5124b6c3151d for a full test report.

EDIT:

-> probably low memory (6GB)
CTranslate2 build failed withgcc: fatal error: Killed signal terminated program cc1plus
ONNX-runtime build failed with g++: fatal error: Killed signal terminated program cc1plus
torchaudio build failed with nvcc error : '"$CICC_PATH/cicc"' died due to signal 9 (Kill signal)

pavelToman · 2025-12-18T23:17:25Z

Test report by @pavelToman
SUCCESS
Build succeeded for 10 out of 10 (total: 8 hours 1 min 55 secs) (10 easyconfigs in total)
node4015.donphan.os - Linux RHEL 9.6, x86_64, Intel(R) Xeon(R) Gold 6240 CPU @ 2.60GHz, 1 x NVIDIA NVIDIA A2, 580.95.05, Python 3.9.21
See https://gist.github.com/pavelToman/c978c1f9348dd21b927b40e3114398fb for a full test report.

pavelToman · 2026-01-06T15:52:11Z

Test report by @pavelToman
SUCCESS
Build succeeded for 22 out of 22 (total: 6 hours 53 mins 51 secs) (10 easyconfigs in total)
node3307.joltik.os - Linux RHEL 9.6, x86_64, Intel(R) Xeon(R) Gold 6242 CPU @ 2.80GHz, 1 x NVIDIA Tesla V100-SXM2-32GB, 580.95.05, Python 3.9.21
See https://gist.github.com/pavelToman/e8073277ef2b6d190d6d9f28e2f1bb41 for a full test report.

pavelToman · 2026-01-06T16:02:53Z

Test report by @pavelToman
SUCCESS
Build succeeded for 3 out of 3 (total: 5 hours 10 mins 6 secs) (10 easyconfigs in total)
node3900.accelgor.os - Linux RHEL 9.6, x86_64, AMD EPYC 7413 24-Core Processor, 1 x NVIDIA NVIDIA A100-SXM4-80GB, 580.95.05, Python 3.9.21
See https://gist.github.com/pavelToman/22da5ef60f966c405a3d13d35a6cfe8b for a full test report.

…32-override

pavelToman · 2026-01-07T15:12:32Z

Test report by @pavelToman
SUCCESS
Build succeeded for 10 out of 10 (total: 5 hours 37 mins 52 secs) (10 easyconfigs in total)
node4308.litleo.os - Linux RHEL 9.6, x86_64, AMD EPYC 9454P 48-Core Processor, 1 x NVIDIA NVIDIA H100 NVL, 580.95.05, Python 3.9.21
See https://gist.github.com/pavelToman/304499abce603fd729fd1c03efaf314f for a full test report.

WilleBell · 2026-01-26T14:42:50Z

@boegelbot please test @ jsc-zen3-a100

boegelbot · 2026-01-26T14:50:09Z

@WilleBell: Request for testing this PR well received on jsczen3l1.int.jsc-zen3.fz-juelich.de

PR test command 'if [[ develop != 'develop' ]]; then EB_BRANCH=develop ./easybuild_develop.sh 2> /dev/null 1>&2; EB_PREFIX=/home/boegelbot/easybuild/develop source init_env_easybuild_develop.sh; fi; EB_PR=24922 EB_ARGS= EB_CONTAINER= EB_REPO=easybuild-easyconfigs EB_BRANCH=develop /opt/software/slurm/bin/sbatch --job-name test_PR_24922 --ntasks=8 --partition=jsczen3g --gres=gpu:1 ~/boegelbot/eb_from_pr_upload_jsc-zen3.sh' executed!

exit code: 0
output:

Submitted batch job 9493

Test results coming soon (I hope)...

Details

- notification for comment with ID 3799953306 processed

Message to humans: this is just bookkeeping information for me,
it is of no use to you (unless you think I have a bug, which I don't).

pavelToman · 2026-01-26T15:52:47Z

Test report by @pavelToman
SUCCESS
Build succeeded for 9 out of 9 (total: 6 hours 9 mins 0 secs) (9 easyconfigs in total)
node3305.joltik.os - Linux RHEL 9.6, x86_64, Intel(R) Xeon(R) Gold 6242 CPU @ 2.80GHz, 1 x NVIDIA Tesla V100-SXM2-32GB, 580.95.05, Python 3.9.21
See https://gist.github.com/pavelToman/65db4c91e617201402c5e590030b64af for a full test report.

boegelbot · 2026-01-28T08:20:03Z

Test report by @boegelbot
FAILED
Build succeeded for 3 out of 10 (total: 34 hours 45 mins 16 secs) (9 easyconfigs in total)
jsczen3g1.int.jsc-zen3.fz-juelich.de - Linux Rocky Linux 9.7, x86_64, AMD EPYC-Milan Processor (zen3), 1 x NVIDIA NVIDIA A100 80GB PCIe, 590.44.01, Python 3.9.23
See https://gist.github.com/boegelbot/c1035c0444989f740d504efdc7bdb0f0 for a full test report.

pavelToman · 2026-01-28T10:43:36Z

Test report by @boegelbot FAILED Build succeeded for 3 out of 10 (total: 34 hours 45 mins 16 secs) (9 easyconfigs in total) jsczen3g1.int.jsc-zen3.fz-juelich.de - Linux Rocky Linux 9.7, x86_64, AMD EPYC-Milan Processor (zen3), 1 x NVIDIA NVIDIA A100 80GB PCIe, 590.44.01, Python 3.9.23 See https://gist.github.com/boegelbot/c1035c0444989f740d504efdc7bdb0f0 for a full test report.

Failed on PyTorch-2.7.1-foss-2024a-CUDA-12.6.0 (not in this PR)

boegel · 2026-02-07T15:57:37Z

@boegelbot please test @ jsc-zen3-a100
CORE_CNT=16

boegelbot · 2026-02-07T16:00:09Z

@boegel: Request for testing this PR well received on jsczen3l1.int.jsc-zen3.fz-juelich.de

PR test command 'if [[ develop != 'develop' ]]; then EB_BRANCH=develop ./easybuild_develop.sh 2> /dev/null 1>&2; EB_PREFIX=/home/boegelbot/easybuild/develop source init_env_easybuild_develop.sh; fi; EB_PR=24922 EB_ARGS= EB_CONTAINER= EB_REPO=easybuild-easyconfigs EB_BRANCH=develop /opt/software/slurm/bin/sbatch --job-name test_PR_24922 --ntasks="16" --partition=jsczen3g --gres=gpu:1 ~/boegelbot/eb_from_pr_upload_jsc-zen3.sh' executed!

exit code: 0
output:

Submitted batch job 9604

Test results coming soon (I hope)...

Details

- notification for comment with ID 3864742736 processed

Message to humans: this is just bookkeeping information for me,
it is of no use to you (unless you think I have a bug, which I don't).

boegelbot · 2026-02-09T20:29:43Z

Test report by @boegelbot
FAILED
Build succeeded for 3 out of 10 (total: 52 hours 28 mins 58 secs) (9 easyconfigs in total)
jsczen3g1.int.jsc-zen3.fz-juelich.de - Linux Rocky Linux 9.7, x86_64, AMD EPYC-Milan Processor (zen3), 1 x NVIDIA NVIDIA A100 80GB PCIe, 590.44.01, Python 3.9.23
See https://gist.github.com/boegelbot/92603006cf7d1339606f066ff414d198 for a full test report.

pavelToman · 2026-02-10T10:14:13Z

Test report by @boegelbot FAILED Build succeeded for 3 out of 10 (total: 52 hours 28 mins 58 secs) (9 easyconfigs in total) jsczen3g1.int.jsc-zen3.fz-juelich.de - Linux Rocky Linux 9.7, x86_64, AMD EPYC-Milan Processor (zen3), 1 x NVIDIA NVIDIA A100 80GB PCIe, 590.44.01, Python 3.9.23 See https://gist.github.com/boegelbot/92603006cf7d1339606f066ff414d198 for a full test report.

Again, failed on PyTorch-2.7.1-foss-2024a-CUDA-12.6.0 (not in this PR)

boegel · 2026-02-10T19:16:00Z

Test report by @boegelbot FAILED Build succeeded for 3 out of 10 (total: 52 hours 28 mins 58 secs) (9 easyconfigs in total) jsczen3g1.int.jsc-zen3.fz-juelich.de - Linux Rocky Linux 9.7, x86_64, AMD EPYC-Milan Processor (zen3), 1 x NVIDIA NVIDIA A100 80GB PCIe, 590.44.01, Python 3.9.23 See https://gist.github.com/boegelbot/92603006cf7d1339606f066ff414d198 for a full test report.

Again, failed on PyTorch-2.7.1-foss-2024a-CUDA-12.6.0 (not in this PR)

This update to PyTorch easyblock should help a lot here...

Add option to allow missing or additional detected PyTorch test suites easybuild-easyblocks#4052

…asyconfigs into 20251218161110_new_pr_WhisperX374

pavelToman added the update label Dec 18, 2025

github-actions bot added the 2024a issues & PRs related to 2024a common toolchains label Dec 18, 2025

pavelToman changed the title ~~{ai}[foss/2024a] WhisperX v3.7.4, ONNX-Runtime v1.23.2, ONNX v1.20.0, ... w/ CUDA 12.6.0~~ {ai}[foss/2024a] WhisperX v3.7.4, ONNX-Runtime v1.23.2, ONNX v1.20.0, PyTorch-Lightning v2.5.6, pyannote.audio v3.4.0, torchaudio v2.7.1, ... w/ CUDA 12.6.0 Dec 18, 2025

pavelToman mentioned this pull request Dec 18, 2025

whisperx vscentrum/vsc-software-stack#613

Open

2 tasks

Update torchaudio-2.7.1-foss-2024a-CUDA-12.6.0.eb - pretest export TF…

18359bf

…32-override

pavelToman mentioned this pull request Jan 8, 2026

{ai,bio}[foss/2024a] scvi-tools v1.4.1, captum v0.8.0 w/ CUDA 12.6.0 #25022

Open

Update bitsandbytes-0.46.1-foss-2024a-CUDA-12.6.0.eb

c53dda3

boegel added this to the 5.x milestone Jan 24, 2026

boegel and others added 5 commits January 24, 2026 15:06

Merge branch 'develop' into 20251218161110_new_pr_WhisperX374

b6b0d8b

Update bitsandbytes-0.46.1-foss-2024a-CUDA-12.6.0.eb

f448ac3

Update CTranslate2-4.5.0-foss-2024a-CUDA-12.6.0.eb

8716884

Update bitsandbytes-0.46.1-foss-2024a-CUDA-12.6.0.eb

64a0bef

Update CTranslate2-4.5.0-foss-2024a-CUDA-12.6.0.eb

1c18e8f

JarneRenders mentioned this pull request Jan 26, 2026

{ai}[foss/2024a] PyTorch-Forecasting v1.5.0 w/ CUDA 12.6.0 #24779

Open

2 tasks

migueldiascosta mentioned this pull request Feb 6, 2026

{bio}[foss/2024a] Boltz-2 v2.2.1 w/ CUDA 12.6.0 #25228

Draft

1 task

Update torchaudio-2.7.1-foss-2024a-CUDA-12.6.0.eb

0da6206

pavelToman added 2 commits February 11, 2026 19:06

update bitsandbytes.eb

334c3bb

separate PR for Lightning

f2ce17d

pavelToman mentioned this pull request Feb 12, 2026

{ai}[foss/2024a] Lightning v2.5.6, PyTorch-Lightning v2.5.6, bitsandbytes v0.46.1 w/ CUDA 12.6.0 #25262

Merged

Merge branch 'develop' of https://github.com/easybuilders/easybuild-e…

db9fbdb

…asyconfigs into 20251218161110_new_pr_WhisperX374

{ai}[foss/2024a] WhisperX v3.7.4, ONNX-Runtime v1.23.2, ONNX v1.20.0, pyannote.audio v3.4.0, torchaudio v2.7.1, ... w/ CUDA 12.6.0 #24922

Are you sure you want to change the base?

{ai}[foss/2024a] WhisperX v3.7.4, ONNX-Runtime v1.23.2, ONNX v1.20.0, pyannote.audio v3.4.0, torchaudio v2.7.1, ... w/ CUDA 12.6.0 #24922

Conversation

pavelToman commented Dec 18, 2025 • edited Loading Uh oh! There was an error while loading. Please reload this page.

Uh oh!

Uh oh!

github-actions bot commented Dec 18, 2025 • edited Loading Uh oh! There was an error while loading. Please reload this page.

Uh oh!

Updated software CTranslate2-4.5.0-foss-2024a-CUDA-12.6.0.eb

Updated software cpu_features-0.10.1-GCCcore-13.3.0.eb

Updated software ONNX-Runtime-1.23.2-foss-2024a-CUDA-12.6.0.eb

Updated software pyannote.audio-3.4.0-foss-2024a-CUDA-12.6.0.eb

Updated software torchaudio-2.7.1-foss-2024a-CUDA-12.6.0.eb

Updated software WhisperX-3.7.4-foss-2024a-CUDA-12.6.0.eb

Uh oh!

pavelToman commented Dec 18, 2025 • edited Loading Uh oh! There was an error while loading. Please reload this page.

Uh oh!

Uh oh!

pavelToman commented Dec 18, 2025

Uh oh!

pavelToman commented Jan 6, 2026

Uh oh!

pavelToman commented Jan 6, 2026

Uh oh!

pavelToman commented Jan 7, 2026

Uh oh!

WilleBell commented Jan 26, 2026

Uh oh!

boegelbot commented Jan 26, 2026

Uh oh!

pavelToman commented Jan 26, 2026

Uh oh!

boegelbot commented Jan 28, 2026

Uh oh!

pavelToman commented Jan 28, 2026

Uh oh!

boegel commented Feb 7, 2026

Uh oh!

boegelbot commented Feb 7, 2026

Uh oh!

boegelbot commented Feb 9, 2026

Uh oh!

pavelToman commented Feb 10, 2026

Uh oh!

boegel commented Feb 10, 2026

Uh oh!

Reviewers

Assignees

Labels

Projects

Milestone

Development

Uh oh!

4 participants

pavelToman commented Dec 18, 2025 •

edited

Loading

github-actions bot commented Dec 18, 2025 •

edited

Loading

Updated software `CTranslate2-4.5.0-foss-2024a-CUDA-12.6.0.eb`

Updated software `cpu_features-0.10.1-GCCcore-13.3.0.eb`

Updated software `ONNX-Runtime-1.23.2-foss-2024a-CUDA-12.6.0.eb`

Updated software `pyannote.audio-3.4.0-foss-2024a-CUDA-12.6.0.eb`

Updated software `torchaudio-2.7.1-foss-2024a-CUDA-12.6.0.eb`

Updated software `WhisperX-3.7.4-foss-2024a-CUDA-12.6.0.eb`

pavelToman commented Dec 18, 2025 •

edited

Loading