Impossible to make it work with Blackwell GPU

[aserrallerios]

I've built the blackwell image locally using the Dockerfile.cuda.12.9 Dockerfile and then used the compose instructions from docker-compose.blackwell.yml. When I transcribe I get this error:

Scriberr-1  | time=08:56:46 level=ERROR msg="Job processing failed" worker_id=0 job_id=0737c58f-f995-40e2-bcb9-19aa37eb3625 error="single-track processing failed: transcription failed: WhisperX execution failed: exit status 1\nLogs:\n/app/whisperx-env/WhisperX/.venv/lib/python3.10/site-packages/torch/cuda/__init__.py:182: UserWarning: CUDA initialization: Unexpected error from cudaGetDeviceCount(). Did you run some cuda functions before calling NumCudaDevices() that might have already set an error? Error 804: forward compatibility was attempted on non supported HW (Triggered internally at /pytorch/c10/cuda/CUDAFunctions.cpp:109.)\n  return torch._C._cuda_getDeviceCount() > 0\nTraceback (most recent call last):\n  File \"/home/appuser/.local/share/uv/python/cpython-3.10-linux-x86_64-gnu/lib/python3.10/runpy.py\", line 196, in _run_module_as_main\n    return _run_code(code, main_globals, None,\n  File \"/home/appuser/.local/share/uv/python/cpython-3.10-linux-x86_64-gnu/lib/python3.10/runpy.py\", line 86, in _run_code\n    exec(code, run_globals)\n  File \"/app/whisperx-env/WhisperX/whisperx/__main__.py\", line 102, in <module>\n    cli()\n  File \"/app/whisperx-env/WhisperX/whisperx/__main__.py\", line 98, in cli\n    transcribe_task(args, parser)\n  File \"/app/whisperx-env/WhisperX/whisperx/transcribe.py\", line 127, in transcribe_task\n    model = load_model(\n  File \"/app/whisperx-env/WhisperX/whisperx/asr.py\", line 357, in load_model\n    model = model or WhisperModel(whisper_arch,\n  File \"/app/whisperx-env/WhisperX/.venv/lib/python3.10/site-packages/faster_whisper/transcribe.py\", line 663, in __init__\n    self.model = ctranslate2.models.Whisper(\nRuntimeError: CUDA failed with error forward compatibility was attempted on non supported HW\n"

This is the nvidia-smi command from inside the container:

# nvidia-smi
Mon May  4 09:39:57 2026
+-----------------------------------------------------------------------------------------+
| NVIDIA-SMI 570.172.08             Driver Version: 570.172.08     CUDA Version: 12.9     |
|-----------------------------------------+------------------------+----------------------+
| GPU  Name                 Persistence-M | Bus-Id          Disp.A | Volatile Uncorr. ECC |
| Fan  Temp   Perf          Pwr:Usage/Cap |           Memory-Usage | GPU-Util  Compute M. |
|                                         |                        |               MIG M. |
|=========================================+========================+======================|
|   0  NVIDIA GeForce RTX 5060 Ti     Off |   00000000:01:00.0 Off |                  N/A |
| 46%   32C    P0             16W /  180W |       0MiB /  16311MiB |      0%      Default |
|                                         |                        |                  N/A |
+-----------------------------------------+------------------------+----------------------+

+-----------------------------------------------------------------------------------------+
| Processes:                                                                              |
|  GPU   GI   CI              PID   Type   Process name                        GPU Memory |
|        ID   ID                                                               Usage      |
|=========================================================================================|
|  No running processes found                                                             |
+-----------------------------------------------------------------------------------------+

[lukyjay]

Same here, with the same GPU. Seems Blackwell support isn't here yet due to out of date dependencies. You can use Whisper with CPU, or see if you can update those dependencies to support our GPU