I got cuda error with my RTX 5090!!

Nobody has claimed this yet.

I install all correct without errors but when i try to generate i got this errors:

* Running on local URL: http://0.0.0.0:7860

To create a public link, set `share=True` in `launch()`.
Unloaded DynamicSwap_LlamaModel as complete.
Unloaded CLIPTextModel as complete.
Unloaded SiglipVisionModel as complete.
Unloaded AutoencoderKLHunyuanVideo as complete.
Unloaded DynamicSwap_HunyuanVideoTransformer3DModelPacked as complete.
Loaded CLIPTextModel to cuda:0 as complete.
Traceback (most recent call last):
File "C:\AI\FramePack\demo_gradio.py", line 122, in worker
llama_vec, clip_l_pooler = encode_prompt_conds(prompt, text_encoder, text_encoder_2, tokenizer, tokenizer_2)
^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
File "C:\AI\FramePack\venv\Lib\site-packages\torch\utils\_contextlib.py", line 116, in decorate_context
return func(*args, **kwargs)
^^^^^^^^^^^^^^^^^^^^^
File "C:\AI\FramePack\diffusers_helper\hunyuan.py", line 31, in encode_prompt_conds
llama_attention_length = int(llama_attention_mask.sum())
^^^^^^^^^^^^^^^^^^^^^^^^^^
RuntimeError: CUDA error: no kernel image is available for execution on the device
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.

Unloaded CLIPTextModel as complete.
Unloaded DynamicSwap_LlamaModel as complete.
Unloaded CLIPTextModel as complete.
Unloaded SiglipVisionModel as complete.
Unloaded AutoencoderKLHunyuanVideo as complete.
Unloaded DynamicSwap_HunyuanVideoTransformer3DModelPacked as complete.

any idea what is causing it and how i can solve?

Read demo_gradio.py around the worker and diffusers_helper/hunyuan.py at line 31, where the traceback reports the failure during prompt encoding. Reproduce the generation attempt on the RTX 5090 and determine which CUDA or PyTorch compatibility issue causes the missing kernel; done means a verified resolution or documented support limitation.

Written by the indexing model from the issue text.