CUDAMatException: Incompatible matrix dimensions

[TianCuan]

I am a beginner of deepnet and I met this problem:

Train Step: 0Traceback (most recent call last):
File "/home/deepnet/trainer.py", line 60, in
main()
File "/home/deepnet/trainer.py", line 54, in main
model.Train()
File "/home/deepnet/neuralnet.py", line 649, in Train
losses = self.TrainOneBatch(step)
File "/home/deepnet/dbm.py", line 270, in TrainOneBatch
losses1 = self.PositivePhase(train=True, step=step)
File "/home/deepnet/dbm.py", line 133, in PositivePhase
self.ComputeUp(node, train=train)
File "/home/deepnet/dbm.py", line 80, in ComputeUp
layer.GetData()
File "/home/deepnet/replicated_softmax_layer.py", line 91, in GetData
self.state.assign(self.data)
File "/home/deepnet/cudamat/cudamat.py", line 340, in assign
raise generate_exception(err_code)
cudamat.cudamat.CUDAMatException: Incompatible matrix dimensions.

how can i resolve it ?