pytorch - 💡(How to fix) Fix CUDNN version issues in 2.13.0.dev20260530+cu132 nightly wheel

🐛 Describe the bug

Torch version: 2.13.0.dev20260530+cu132 GPU: RTX 5090 (SM 12.0) OS: Windows 11 Error: CUDNN_STATUS_SUBLIBRARY_VERSION_MISMATCH at F.conv2d() during VAEDecode Root cause found: Mismatched cuDNN sublibraries in the nightly wheel — five DLLs report 9.20.0.48 (cuDNN 9.2.0), but cudnn_graph64_9.dll and cudnn_heuristic64_9.dll have no embedded version info, causing cuDNN's internal version check to fail at cudnnFinalize Reproduction: Any F.conv2d() call triggers it; VAEDecode is the most reliable reproducer

Versions

Collecting environment information... PyTorch version: 2.13.0.dev20260530+cu132 Is debug build: False CUDA used to build PyTorch: 13.2 ROCM used to build PyTorch: N/A

OS: Microsoft Windows 11 Home (10.0.26200 64-bit) GCC version: Could not collect Clang version: 20.1.8 CMake version: version 4.2.1 Libc version: N/A

Python version: 3.13.11 (tags/v3.13.11:6278944, Dec 5 2025, 16:26:58) [MSC v.1944 64 bit (AMD64)] (64-bit runtime) Python platform: Windows-11-10.0.26200-SP0 Is CUDA available: True CUDA runtime version: 13.3.33 CUDA_MODULE_LOADING set to: GPU models and configuration: GPU 0: NVIDIA GeForce RTX 5090 Nvidia driver version: Could not collect cuDNN version: Could not collect Is XPU available: False HIP runtime version: N/A MIOpen runtime version: N/A Is XNNPACK available: True Caching allocator config: N/A

CPU: Name: 13th Gen Intel(R) Core(TM) i7-13700KF Manufacturer: GenuineIntel Family: 198 Architecture: 9 ProcessorType: 3 DeviceID: CPU0 CurrentClockSpeed: 3400 MaxClockSpeed: 3400 L2CacheSize: 24576 L2CacheSpeed: None Revision: None

Versions of relevant libraries: [pip3] numpy==2.4.4 [pip3] nvidia-cublas==13.4.0.1 [pip3] nvidia-cuda-nvrtc==13.2.78 [pip3] nvidia-cuda-runtime==13.0.96 [pip3] nvidia-cudnn-cu13==9.21.1.3 [pip3] onnx==1.19.1 [pip3] onnx_graphsurgeon==0.6.1 [pip3] onnx-ir==0.2.1 [pip3] onnxconverter-common==1.16.0 [pip3] onnxruntime==1.23.2 [pip3] onnxruntime-gpu==1.22.0 [pip3] onnxscript==0.7.0 [pip3] onnxslim==0.1.91 [pip3] open_clip_torch==3.2.0 [pip3] pytorch-wpe==0.0.1 [pip3] rotary-embedding-torch==0.8.9 [pip3] sageattention==2.2.0+cu130torch2.9.0andhigher.post4 [pip3] segmentation_models_pytorch==0.5.0 [pip3] torch==2.13.0.dev20260530+cu132 [pip3] torch-complex==0.4.4 [pip3] torchaudio==2.11.0.dev20260530+cu132 [pip3] torchcodec==0.11.0 [pip3] torchsde==0.2.6 [pip3] torchvision==0.28.0.dev20260530+cu132 [pip3] triton-windows==3.6.0.post26 [conda] Could not collect

cc @malfet @atalman @tinglvv @nWEIdia @peterjc123 @mszhanyi @skyline75489 @nbcsm @iremyux @Blackhex @nkhasbag-nv @csarofeen @ptrblck @eqy