成功解決torch.cuda.CudaError: CUDA driver version is insufficient for CUDA runtime version (35) [ WARN:0

解決問題

[INFO] Initializing front camera...

Traceback (most recent call last):

 File ".\image_animation.py", line 76, in <module>

   kp_driving_initial = kp_detector(source1)

 File "F:\File_Anaconda\CV2020_RealTimeImageAnimation\envs\lib\site-packages\torch\nn\modules\module.py", line 489, in __call__

   result = self.forward(*input, **kwargs)

 File "F:\File_Anaconda\CV2020_RealTimeImageAnimation\envs\lib\site-packages\torch\nn\parallel\data_parallel.py", line 139, in forward

   inputs, kwargs = self.scatter(inputs, kwargs, self.device_ids)

 File "F:\File_Anaconda\CV2020_RealTimeImageAnimation\envs\lib\site-packages\torch\nn\parallel\data_parallel.py", line 150, in scatter

   return scatter_kwargs(inputs, kwargs, device_ids, dim=self.dim)

 File "F:\File_Anaconda\CV2020_RealTimeImageAnimation\envs\lib\site-packages\torch\nn\parallel\scatter_gather.py", line 35, in scatter_kwargs

   inputs = scatter(inputs, target_gpus, dim) if inputs else []

 File "F:\File_Anaconda\CV2020_RealTimeImageAnimation\envs\lib\site-packages\torch\nn\parallel\scatter_gather.py", line 28, in scatter

   return scatter_map(inputs)

 File "F:\File_Anaconda\CV2020_RealTimeImageAnimation\envs\lib\site-packages\torch\nn\parallel\scatter_gather.py", line 15, in scatter_map

   return list(zip(*map(scatter_map, obj)))

 File "F:\File_Anaconda\CV2020_RealTimeImageAnimation\envs\lib\site-packages\torch\nn\parallel\scatter_gather.py", line 13, in scatter_map

   return Scatter.apply(target_gpus, None, dim, obj)

 File "F:\File_Anaconda\CV2020_RealTimeImageAnimation\envs\lib\site-packages\torch\nn\parallel\_functions.py", line 95, in forward

   main_stream.wait_stream(streams[i])

 File "F:\File_Anaconda\CV2020_RealTimeImageAnimation\envs\lib\site-packages\torch\cuda\streams.py", line 55, in wait_stream

   self.wait_event(stream.record_event())

 File "F:\File_Anaconda\CV2020_RealTimeImageAnimation\envs\lib\site-packages\torch\cuda\streams.py", line 68, in record_event

   event = Event()

 File "F:\File_Anaconda\CV2020_RealTimeImageAnimation\envs\lib\site-packages\torch\cuda\streams.py", line 162, in __init__

   check_error(self._cudart.cudaEventCreateWithFlags(ctypes.byref(ptr), ctypes.c_uint(flags)))

 File "F:\File_Anaconda\CV2020_RealTimeImageAnimation\envs\lib\site-packages\torch\cuda\__init__.py", line 208, in check_error

   raise CudaError(res)

torch.cuda.CudaError: CUDA driver version is insufficient for CUDA runtime version (35)

[ WARN:0] global C:\projects\opencv-python\opencv\modules\videoio\src\cap_msmf.cpp (674) SourceReaderCB::~SourceReaderCB terminating async callback

解決思路

torch.cuda.Cuda錯誤:Cuda驅動版本是不夠的Cuda運作時版本SourceReaderCB::~SourceReaderCB終止異步回調

解決方法

相關文章：Pytorch：深度學習中pytorch/torchvision版本和CUDA版本最正确版本比對、對應版本安裝之詳細攻略

檢視torch，以及torch所适合的cuda版本，此問題，大多數都是版本不比對造成的！