🐛 Describe the bug
Errors in detecting ffmpeg is already reported. I modified code to get the origin error, the version 7 of ffmpeg being installed, which is:
Traceback (most recent call last):
File "/data/whisper/.venv/lib64/python3.13/site-packages/torch/_ops.py", line 1488, in load_library
ctypes.CDLL(path)
~~~~~~~~~~~^^^^^^
File "/usr/lib64/python3.13/ctypes/__init__.py", line 361, in __init__
self._handle = self._load_library(name, mode, handle, winmode)
~~~~~~~~~~~~~~~~~~^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
File "/usr/lib64/python3.13/ctypes/__init__.py", line 401, in _load_library
return _dlopen(name, mode)OSError: /lib64/libtiff.so.6: undefined symbol: jpeg12_write_raw_data, version LIBJPEG_8.0`
The above exception was the direct cause of the following exception:
Traceback (most recent call last):
File "/data/whisper/whisper.py", line 33, in <module>
result = pipe("/home/yves/20260428_163726_10min.mp3")
File "/data/whisper/.venv/lib64/python3.13/site-packages/transformers/pipelines/automatic_speech_recognition.py", line 244, in __call__
return super().__call__(inputs, **kwargs)
~~~~~~~~~~~~~~~~^^^^^^^^^^^^^^^^^^
File "/data/whisper/.venv/lib64/python3.13/site-packages/transformers/pipelines/base.py", line 1256, in __call__
return next(
iter(
...<3 lines>...
)
)
File "/data/whisper/.venv/lib64/python3.13/site-packages/transformers/pipelines/pt_utils.py", line 126, in __next__
item = next(self.iterator)
File "/data/whisper/.venv/lib64/python3.13/site-packages/transformers/pipelines/pt_utils.py", line 271, in __next__
processed = self.infer(next(self.iterator), **self.params)
~~~~^^^^^^^^^^^^^^^
File "/data/whisper/.venv/lib64/python3.13/site-packages/torch/utils/data/dataloader.py", line 732, in __next__
data = self._next_data()
File "/data/whisper/.venv/lib64/python3.13/site-packages/torch/utils/data/dataloader.py", line 788, in _next_data
data = self._dataset_fetcher.fetch(index) # may raise StopIteration
File "/data/whisper/.venv/lib64/python3.13/site-packages/torch/utils/data/_utils/fetch.py", line 33, in fetch
data.append(next(self.dataset_iter))
~~~~^^^^^^^^^^^^^^^^^^^
File "/data/whisper/.venv/lib64/python3.13/site-packages/transformers/pipelines/pt_utils.py", line 188, in __next__
processed = next(self.subiterator)
File "/data/whisper/.venv/lib64/python3.13/site-packages/transformers/pipelines/automatic_speech_recognition.py", line 365, in preprocess
import torchcodec
File "/data/whisper/.venv/lib64/python3.13/site-packages/torchcodec/__init__.py", line 12, in <module>
from . import decoders, encoders, samplers # noqa
^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
File "/data/whisper/.venv/lib64/python3.13/site-packages/torchcodec/decoders/__init__.py", line 7, in <module>
from .._core import AudioStreamMetadata, VideoStreamMetadata
File "/data/whisper/.venv/lib64/python3.13/site-packages/torchcodec/_core/__init__.py", line 8, in <module>
from ._metadata import (
...<5 lines>...
)
File "/data/whisper/.venv/lib64/python3.13/site-packages/torchcodec/_core/_metadata.py", line 16, in <module>
from torchcodec._core.ops import (
...<3 lines>...
)
File "/data/whisper/.venv/lib64/python3.13/site-packages/torchcodec/_core/ops.py", line 104, in <module>
ffmpeg_major_version, core_library_path = load_torchcodec_shared_libraries()
~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~^^
File "/data/whisper/.venv/lib64/python3.13/site-packages/torchcodec/_core/ops.py", line 56, in load_torchcodec_shared_libraries
torch.ops.load_library(core_library_path)
~~~~~~~~~~~~~~~~~~~~~~^^^^^^^^^^^^^^^^^^^
File "/data/whisper/.venv/lib64/python3.13/site-packages/torch/_ops.py", line 1491, in load_library
raise OSError(f"Could not load this library: {path}") from e
OSError: Could not load this library: /data/whisper/.venv/lib/python3.13/site-packages/torchcodec/libtorchcodec_core7.so
I identified that /lib64/libtiff.so.6 is the lib from my system, which loads well outside the virtualenv. In fact, I can reproduce the problem when I load libjpeg from the virtualenv before loading libtiff library:
python
Python 3.13.12 (main, Feb 6 2026, 00:42:45) [GCC 15.2.0] on linux
Type "help", "copyright", "credits" or "license" for more information.
>>> import ctypes
>>> ctypes.CDLL("./.venv/lib/python3.13/site-packages/torchvision.libs/libjpeg.d246b9ea.so.8")
<CDLL './.venv/lib/python3.13/site-packages/torchvision.libs/libjpeg.d246b9ea.so.8', handle 562681ea2bd0 at 0x7f13f71c27b0>
>>> path = "/lib64/libtiff.so.6"
>>> ctypes.CDLL(path)
Traceback (most recent call last):
File "<python-input-3>", line 1, in <module>
ctypes.CDLL(path)
~~~~~~~~~~~^^^^^^
File "/usr/lib64/python3.13/ctypes/__init__.py", line 361, in __init__
self._handle = self._load_library(name, mode, handle, winmode)
~~~~~~~~~~~~~~~~~~^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
File "/usr/lib64/python3.13/ctypes/__init__.py", line 401, in _load_library
return _dlopen(name, mode)
OSError: /lib64/libtiff.so.6: undefined symbol: jpeg12_write_raw_data, version LIBJPEG_8.0
>>> exit()
$ python
Python 3.13.12 (main, Feb 6 2026, 00:42:45) [GCC 15.2.0] on linux
Type "help", "copyright", "credits" or "license" for more information.
>>> import ctypes
>>> path = "/lib64/libtiff.so.6"
>>> ctypes.CDLL(path)
<CDLL '/lib64/libtiff.so.6', handle 5577ac5ca270 at 0x7f5b99e927b0>
Our libjpeg is the one from https://libjpeg-turbo.org/ thus this is why the function jpeg12_write_raw_data is used. However, the libjpeg from torchvision.libs interferes and causes the crash.
I installed pytorch through:
pip install torch torchvision torchaudio --index-url https://download.pytorch.org/whl/rocm6.4
Versions
python collect_env.py
Collecting environment information...
/opt/amdgpu/share/libdrm/amdgpu.ids: No such file or directory
/opt/amdgpu/share/libdrm/amdgpu.ids: No such file or directory
PyTorch version: 2.9.1+rocm6.4
Is debug build: False
CUDA used to build PyTorch: N/A
ROCM used to build PyTorch: 6.4.43484-123eb5128
OS: Mageia 10 (x86_64)
GCC version: (Mageia 15.2.0-1.mga10) 15.2.0
Clang version: Could not collect
CMake version: version 4.1.3
Libc version: glibc-2.42
Python version: 3.13.12 (main, Feb 6 2026, 00:42:45) [GCC 15.2.0] (64-bit runtime)
Python platform: Linux-6.18.22-desktop-1.mga10-x86_64-with-glibc2.42
Is CUDA available: True
CUDA runtime version: Could not collect
CUDA_MODULE_LOADING set to:
GPU models and configuration: AMD Radeon Graphics (gfx1100)
Nvidia driver version: Could not collect
cuDNN version: Could not collect
Is XPU available: False
HIP runtime version: 6.4.43484
MIOpen runtime version: 3.4.0
Is XNNPACK available: True
Caching allocator config: N/A
CPU:
Architecture : x86_64
Mode(s) opératoire(s) des processeurs : 32-bit, 64-bit
Tailles des adresses: 48 bits physical, 48 bits virtual
Boutisme : Little Endian
Processeur(s) : 12
Liste de processeur(s) en ligne : 0-11
Identifiant constructeur : AuthenticAMD
Nom de modèle : AMD Ryzen 5 7600X 6-Core Processor
Famille de processeur : 25
Modèle : 97
Thread(s) par cœur : 2
Cœur(s) par socket : 6
Socket(s) : 1
Révision : 2
Accroissement de fréquence : activé
multiplication des MHz du/des CPU(s) : 65%
Vitesse maximale du processeur en MHz : 5457,1050
Vitesse minimale du processeur en MHz : 427,3640
BogoMIPS : 9381,88
Drapeaux : fpu vme de pse tsc msr pae mce cx8 apic sep mtrr pge mca cmov pat pse36 clflush mmx fxsr sse sse2 ht syscall nx mmxext fxsr_opt pdpe1gb rdtscp lm constant_tsc rep_good amd_lbr_v2 nopl xtopology nonstop_tsc cpuid extd_apicid aperfmperf rapl pni pclmulqdq monitor ssse3 fma cx16 sse4_1 sse4_2 movbe popcnt aes xsave avx f16c rdrand lahf_lm cmp_legacy svm extapic cr8_legacy abm sse4a misalignsse 3dnowprefetch osvw ibs skinit wdt tce topoext perfctr_core perfctr_nb bpext perfctr_llc mwaitx cpuid_fault cpb cat_l3 cdp_l3 hw_pstate ssbd mba perfmon_v2 ibrs ibpb stibp ibrs_enhanced vmmcall fsgsbase bmi1 avx2 smep bmi2 erms invpcid cqm rdt_a avx512f avx512dq rdseed adx smap avx512ifma clflushopt clwb avx512cd sha_ni avx512bw avx512vl xsaveopt xsavec xgetbv1 xsaves cqm_llc cqm_occup_llc cqm_mbm_total cqm_mbm_local avx512_bf16 clzero irperf xsaveerptr rdpru wbnoinvd cppc arat npt lbrv svm_lock nrip_save tsc_scale vmcb_clean flushbyasid decodeassists pausefilter pfthreshold avic vgif x2avic v_spec_ctrl vnmi avx512vbmi umip pku ospke avx512_vbmi2 gfni vaes vpclmulqdq avx512_vnni avx512_bitalg avx512_vpopcntdq rdpid overflow_recov succor smca fsrm flush_l1d amd_lbr_pmc_freeze
Virtualisation : AMD-V
Cache L1d : 192 KiB (6 instances)
Cache L1i : 192 KiB (6 instances)
Cache L2 : 6 MiB (6 instances)
Cache L3 : 32 MiB (1 instance)
Nœud(s) NUMA : 1
Nœud NUMA 0 de processeur(s) : 0-11
Vulnérabilité Gather data sampling : Not affected
Vulnérabilité Ghostwrite : Not affected
Vulnérabilité Indirect target selection : Not affected
Vulnérabilité Itlb multihit : Not affected
Vulnérabilité L1tf : Not affected
Vulnérabilité Mds : Not affected
Vulnérabilité Meltdown : Not affected
Vulnérabilité Mmio stale data : Not affected
Vulnérabilité Old microcode : Not affected
Vulnérabilité Reg file data sampling : Not affected
Vulnérabilité Retbleed : Not affected
Vulnérabilité Spec rstack overflow : Mitigation; Safe RET
Vulnérabilité Spec store bypass : Mitigation; Speculative Store Bypass disabled via prctl
Vulnérabilité Spectre v1 : Mitigation; usercopy/swapgs barriers and __user pointer sanitization
Vulnérabilité Spectre v2 : Mitigation; Enhanced / Automatic IBRS; IBPB conditional; STIBP always-on; PBRSB-eIBRS Not affected; BHI Not affected
Vulnérabilité Srbds : Not affected
Vulnérabilité Tsa : Mitigation; Clear CPU buffers
Vulnérabilité Tsx async abort : Not affected
Vulnérabilité Vmscape : Mitigation; IBPB before exit to userspace
Versions of relevant libraries:
[pip3] numpy==2.4.4
[pip3] nvidia-cublas==13.1.0.3
[pip3] nvidia-cuda-cupti==13.0.85
[pip3] nvidia-cuda-nvrtc==13.0.88
[pip3] nvidia-cuda-runtime==13.0.96
[pip3] nvidia-cudnn-cu13==9.19.0.56
[pip3] nvidia-cufft==12.0.0.61
[pip3] nvidia-curand==10.4.0.35
[pip3] nvidia-cusolver==12.0.4.66
[pip3] nvidia-cusparse==12.6.3.3
[pip3] nvidia-cusparselt-cu13==0.8.0
[pip3] nvidia-nccl-cu13==2.28.9
[pip3] nvidia-nvjitlink==13.0.88
[pip3] nvidia-nvtx==13.0.85
[pip3] pytorch-triton-rocm==3.5.1
[pip3] torch==2.9.1+rocm6.4
[pip3] torchaudio==2.9.1+rocm6.4
[pip3] torchcodec==0.9.1
[pip3] torchvision==0.24.1+rocm6.4
[pip3] triton==3.6.0
[conda] Could not collect
🐛 Describe the bug
Errors in detecting ffmpeg is already reported. I modified code to get the origin error, the version 7 of ffmpeg being installed, which is:
I identified that /lib64/libtiff.so.6 is the lib from my system, which loads well outside the virtualenv. In fact, I can reproduce the problem when I load libjpeg from the virtualenv before loading libtiff library:
Our libjpeg is the one from https://libjpeg-turbo.org/ thus this is why the function
jpeg12_write_raw_datais used. However, the libjpeg from torchvision.libs interferes and causes the crash.I installed pytorch through:
pip install torch torchvision torchaudio --index-url https://download.pytorch.org/whl/rocm6.4Versions