CUDA backend fails on RTX 5080 with "no kernel image is available for execution on the device" #400

Activity

1. mvanhorn

   added a commit that references this issue on Apr 13, 2026

   feat: add Blackwell GPU (sm_120) CUDA support

   ...

   f9835d2
2. cod3blox commented on Apr 13, 2026

   

   cod3blox

   on Apr 13, 2026

   Last edited by cod3blox

   More actions

   I’m experiencing the same issue. Yesterday somehow, after the initial installation, Voicebox was working without any problems. Today, out of nowhere, I’m getting CUDA errors.

   

   INFO:     127.0.0.1:61403 - "POST /generate/332ad2b6-0fcc-4827-a698-4ba8624ba950/retry HTTP/1.1" 200 OK
   INFO:     127.0.0.1:61403 - "GET /history?limit=20 HTTP/1.1" 200 OK
   INFO:     127.0.0.1:62786 - "GET /generate/332ad2b6-0fcc-4827-a698-4ba8624ba950/status HTTP/1.1" 200 OK
   Traceback (most recent call last):
     File "backend\services\generation.py", line 65, in run_generation
     File "backend\services\profiles.py", line 435, in create_voice_prompt_for_profile
     File "backend\backends\pytorch_backend.py", line 172, in create_voice_prompt
     File "asyncio\threads.py", line 25, in to_thread
     File "concurrent\futures\thread.py", line 59, in run
     File "backend\backends\pytorch_backend.py", line 165, in _create_prompt_sync
     File "torch\utils\_contextlib.py", line 124, in decorate_context
       return func(*args, **kwargs)
              ^^^^^^^^^^^^^^^^^^^^^
     File "qwen_tts\inference\qwen3_tts_model.py", line 427, in create_voice_clone_prompt
     File "qwen_tts\inference\qwen3_tts_tokenizer.py", line 248, in encode
     File "transformers\feature_extraction_utils.py", line 246, in to
       self.data = {k: maybe_to(v) for k, v in self.items()}
                       ^^^^^^^^^^^
     File "transformers\feature_extraction_utils.py", line 240, in maybe_to
       return v.to(*args, **kwargs)
              ^^^^^^^^^^^^^^^^^^^^^
   torch.AcceleratorError: CUDA error: no kernel image is available for execution on the device
   Search for `cudaErrorNoKernelImageForDevice' in https://docs.nvidia.com/cuda/cuda-runtime-api/group__CUDART__TYPES.html for more information.
   CUDA kernel errors might be asynchronously reported at some other API call, so the stacktrace below might be incorrect.
   For debugging consider passing CUDA_LAUNCH_BLOCKING=1
   Compile with `TORCH_USE_CUDA_DSA` to enable device-side assertions.
   INFO:     127.0.0.1:62786 - "GET /history?limit=20 HTTP/1.1" 200 OK
   INFO:     127.0.0.1:62786 - "GET /models/status HTTP/1.1" 200 OK
   INFO:     127.0.0.1:61403 - "GET /models/cache-dir HTTP/1.1" 200 OK
   INFO:     127.0.0.1:53319 - "GET /tasks/active HTTP/1.1" 200 OK
   

   Operating System: Windows 11 Pro 64-bit (10.0, Build 26200) (26100.ge_release.240331-1435)
   Processor: AMD Ryzen 9 7950X 16-Core Processor
   Available OS Memory: 65174MB RAM
   Card name: NVIDIA GeForce RTX 5080
   Nvidia Driver Version: 595.97
3. jamiepine

   added a commit that references this issue on Apr 16, 2026

   feat: add Blackwell GPU (sm_120) CUDA support (#401)

   ...

   c9d8142
4. jamiepine

   added

   bugSomething isn't workingSomething isn't working

   duplicateThis issue or pull request already existsThis issue or pull request already exists

   on Apr 19, 2026
5. capy-ai-staging commented on Oct 7, 2026

   

   capy-ai-stagingBot

   on Oct 7, 2026 – with capy-ai-staging

   Contributor

   More actions

   This was fixed by #316 and #401 (thanks @mvanhorn): the Windows CUDA backend moved from cu126 to cu128 wheels with sm_120 in the build, so RTX 50-series (Blackwell) cards have worked since Voicebox 0.4.0. If you are still on 0.3.x, please update and re-download the CUDA backend from Settings. Closing as fixed; if you still hit "no kernel image is available" on 0.4.0 or later, please reopen with the log.
6. capy-ai-staging

   closed this as completedon Oct 7, 2026