InternalError (see above for traceback): cuDNN launch failure : input shape ([64,1024,1,1])

InternalError (see above for traceback): cuDNN launch failure : input shape ([64,1024,1,1])
[[node generator/g_bn0/FusedBatchNorm (defined at E:\zb\DCGAN-tensorflow-master\venv\lib\site-packages\tensorflow\contrib\layers\python\layers\layers.py:368) = FusedBatchNorm[T=DT_FLOAT, data_format="NCHW", epsilon=1.001e-05, is_training=true, _device="/job:localhost/replica:0/task:0/device:GPU:0"](generator/g_bn0/Fused
BatchNorm-0-TransposeNHWCToNCHW-LayoutOptimizer, generator/g_bn0/gamma/read, generator/g_bn0/beta/read, discriminator/d_bn1/Const, discriminator/d_bn1/Const)]]
[[{{node discriminator/Sigmoid/_7}} = _Recv[client_terminated=false, recv_device="/job:localhost/replica:0/task:0/device:CPU:0", send_device="/job:localhost/replica:0/task:0/device:GPU:0", send_device_incarnation=1, tensor_name="edge_272_discriminator/Sigmoid", tensor_type=DT_FLOAT, _device="/job:localhost/replica:0/task
:0/device:CPU:0"]()]]

Start with the reported cuDNN launch failure and the traceback entry at tensorflow/contrib/layers/python/layers/layers.py:368. Reproduce the failure using the generator/g_bn0/FusedBatchNorm and discriminator/Sigmoid nodes shown in the report, then identify the relevant environment or input-shape conditions. Done means the reported GPU failure is explained and a verified correction or reproduction guidance is documented.

Written by the indexing model from the issue text.