CUFFT error

[abcdvzz]

When I train my model on cars dataset, I encountered this error, it seems like fft's error:
Check failed: (cufftPlanMany(&plan_noinv_1 , 1, &n, __null, 0, 0, __null, 0, 0, CUFFT_R2C, hw)) == CUFFT_SUCCESS (2 vs. 0) ...
How could i solve this problem? Does it have any relationship with my cuda or cudnn?

[gy20073]

as shown by the cufft return error: typedef enum cufftResult_t { CUFFT_SUCCESS = 0, // The cuFFT operation was successful CUFFT_INVALID_PLAN = 1, // cuFFT was passed an invalid plan handle CUFFT_ALLOC_FAILED = 2, // cuFFT failed to allocate GPU or CPU memory Read more at: http://docs.nvidia.com/cuda/cufft/index.html#ixzz55JV5SWto <http://docs.nvidia.com/cuda/cufft/index.html#ixzz55JV5SWto> Follow us: @gpucomputing on Twitter <http://ec.tynt.com/b/rw?id=aBENEGgL0r44W6acwqm_6r&u=GPUComputing> | NVIDIA on Facebook <http://ec.tynt.com/b/rf?id=aBENEGgL0r44W6acwqm_6r&u=NVIDIA> You don’t have enough GPU or CPU memory left.

…

On Jan 26, 2018, at 5:19 AM, abcdvzz ***@***.***> wrote: cufftPlanMany