Arch Linux User Repository

This package will update iff there is a change in the code between releases: https://github.com/immich-app/immich/commits/main/machine-learning

no idea why, working now. perhaps cache held an older build with same version.

I did update the package for that. Is it not working?

python-insightface is now at version 1.0.1, not meeting the <1.0 requirement

The issue below has now been solved; python-albumentationsx has been updated.

As @CountMurphy has identified, the python-albumentationsx AUR package has not been updated for a long time and needs a bump to be compatible with the python-albucore update. Either locally bump python-albumentationsx or downgrade python-albucore.

FYI, the latest version of python-albucore doesn't work with python-numpy 2.4.4 and causes immich-machine-learning to crash. DOwngrading albucore to 0.0.36-1 fixed my crash.

I cannot upgrade my system because the latest version of pillow is 12.2.0-1 and break the dep python-pillow<12.2 required by immich-machine-learning

Hmm, it appears GPUs that have a compute capability below 5.2 are not supported (see https://developer.nvidia.com/cuda/gpus/legacy). I'm guessing you're right.

There was more but I hit the character limit

@aliu here is what I was seeing:

2026-03-06 18:02:18.085617596 [W:onnxruntime:, transformer_memcpy.cc:111 ApplyImpl] 2 Memcpy nodes are added to the graph mai
n_graph for CUDAExecutionProvider. It might have negative impact on performance (including unable to run CUDA graph). Set session_options.log_severity_level=1 to see the detail logs bef
ore this message. 
Mar 06 18:02:18 HOSTNAME_REDACTED immich-machine-learning[1579739]: 2026-03-06 18:02:18.089427915 [E:onnxruntime:, inference_session.cc:2579 operator()] Exception during initialization: /usr/sr
c/debug/onnxruntime/onnxruntime-cuda/onnxruntime/core/providers/cuda/cuda_call.cc:129 std::conditional_t<THRW, void, onnxruntime::common::Status> onnxruntime::CudaCall(ERRTYPE, const ch
ar*, const char*, SUCCTYPE, const char*, const char*, int) [with ERRTYPE = cublasStatus_t; bool THRW = true; SUCCTYPE = cublasStatus_t; std::conditional_t<THRW, void, common::Status> = 
void] /usr/src/debug/onnxruntime/onnxruntime-cuda/onnxruntime/core/providers/cuda/cuda_call.cc:121 std::conditional_t<THRW, void, onnxruntime::common::Status> onnxruntime::CudaCall(ERRT
YPE, const char*, const char*, SUCCTYPE, const char*, const char*, int) [with ERRTYPE = cublasStatus_t; bool THRW = true; SUCCTYPE = cublasStatus_t; std::conditional_t<THRW, void, commo
n::Status> = void] CUBLAS failure 8: CUBLAS_STATUS_ARCH_MISMATCH ; GPU=0 ; hostname=HOSTNAME_REDACTED ; file=/usr/src/debug/onnxruntime/onnxruntime-cuda/onnxruntime/core/providers/cuda/cuda_exe
cution_provider.cc ; line=228 ; expr=cublasCreate(&cublas_handle_); 
Mar 06 18:02:18 HOSTNAME_REDACTED immich-machine-learning[1579739]: Stacktrace: 
Mar 06 18:02:18 HOSTNAME_REDACTED immich-machine-learning[1579739]: /usr/lib/python3.14/site-packages/onnxruntime/capi/libonnxruntime_providers_cuda.so(+0x44d34d) [0x7fdab044d34d] 
Mar 06 18:02:18 HOSTNAME_REDACTED immich-machine-learning[1579739]: /usr/lib/python3.14/site-packages/onnxruntime/capi/libonnxruntime_providers_cuda.so(+0x173f4f) [0x7fdab0173f4f] 
Mar 06 18:02:18 HOSTNAME_REDACTED immich-machine-learning[1579739]: /usr/lib/python3.14/site-packages/onnxruntime/capi/libonnxruntime_providers_cuda.so(+0x178b3a) [0x7fdab0178b3a] 
Mar 06 18:02:18 HOSTNAME_REDACTED immich-machine-learning[1579739]: /usr/lib/python3.14/site-packages/onnxruntime/capi/libonnxruntime_providers_cuda.so(+0x180151) [0x7fdab0180151] 
Mar 06 18:02:18 HOSTNAME_REDACTED immich-machine-learning[1579739]: /usr/lib/python3.14/site-packages/onnxruntime/capi/onnxruntime_pybind11_state.so(+0xe4d72e) [0x7fdb3964d72e] 
Mar 06 18:02:18 HOSTNAME_REDACTED immich-machine-learning[1579739]: /usr/lib/python3.14/site-packages/onnxruntime/capi/onnxruntime_pybind11_state.so(+0xe50361) [0x7fdb39650361] 
Mar 06 18:02:18 HOSTNAME_REDACTED immich-machine-learning[1579739]: /usr/lib/python3.14/site-packages/onnxruntime/capi/onnxruntime_pybind11_state.so(+0x5aa128) [0x7fdb38daa128] 
Mar 06 18:02:18 HOSTNAME_REDACTED immich-machine-learning[1579739]: /usr/lib/python3.14/site-packages/onnxruntime/capi/onnxruntime_pybind11_state.so(+0x41ff07) [0x7fdb38c1ff07] 
Mar 06 18:02:18 HOSTNAME_REDACTED immich-machine-learning[1579739]: /usr/lib/python3.14/site-packages/onnxruntime/capi/onnxruntime_pybind11_state.so(+0x478d8c) [0x7fdb38c78d8c] 
Mar 06 18:02:18 HOSTNAME_REDACTED immich-machine-learning[1579739]: /usr/lib/python3.14/site-packages/onnxruntime/capi/onnxruntime_pybind11_state.so(+0x3b277b) [0x7fdb38bb277b] 
Mar 06 18:02:18 HOSTNAME_REDACTED immich-machine-learning[1579739]: /usr/lib/libpython3.14.so.1.0(PyObject_Vectorcall+0x5d) [0x7fdb3d16fa5d] 
Mar 06 18:02:18 HOSTNAME_REDACTED immich-machine-learning[1579739]: /usr/lib/libpython3.14.so.1.0(_PyEval_EvalFrameDefault+0xd05) [0x7fdb3d1844c5] 
Mar 06 18:02:18 HOSTNAME_REDACTED immich-machine-learning[1579739]: /usr/lib/libpython3.14.so.1.0(+0x17eae5) [0x7fdb3d17eae5] 
Mar 06 18:02:18 HOSTNAME_REDACTED immich-machine-learning[1579739]: /usr/lib/libpython3.14.so.1.0(+0x1b9997) [0x7fdb3d1b9997] 
Mar 06 18:02:18 HOSTNAME_REDACTED immich-machine-learning[1579739]: /usr/lib/libpython3.14.so.1.0(_PyObject_MakeTpCall+0x48b) [0x7fdb3d16cd6b]