search

jiangycTarheel-zz / EPAr

MIT License
24 stars 2 forks source link

CudnnRNNParamsSize Error #1

Open slobLX opened 4 years ago

slobLX commented 4 years ago

Hello, when I use './scripts/qangaroo/full-epar-train.sh' I encountered this problem:

tensorflow.python.framework.errors_impl.InvalidArgumentError: No OpKernel was registered to support Op 'CudnnRNNParamsSize' used by node model_0/prepro/u1/forward/CudnnRNNParamsSize (defined at /QApaper/EPA/EPAr/my/tensorflow/cudnn_recurrent_layers.py:83) with these attrs: [seed=87654321, dropout=0, S=DT_INT32, input_mode="linear_input", T=DT_FLOAT, direction="unidirectional", rnn_mode="lstm", seed2=0] Registered devices: [CPU, XLA_CPU, XLA_GPU] Registered kernels: device='GPU'; T in [DT_DOUBLE]; S in [DT_INT32] device='GPU'; T in [DT_FLOAT]; S in [DT_INT32] device='GPU'; T in [DT_HALF]; S in [DT_INT32]

 [[model_0/prepro/u1/forward/CudnnRNNParamsSize]]

So could you tell me how can I solve this problem. Thanks!

slobLX commented 4 years ago

I'm using ubuntu 16.04 tensorflow==1.14.0 cuda-10.1 cudnn=7.6.3

shuxiaobo commented 3 years ago

get the same error, did you solve it? @slobLX