I am trying to implement a Character LSTM using Pytorch.But I am getting cudnn_status_bad_params errors.This is the training loop.I getting error on line output = model(input_seq).
for epoch in tqdm(range(epochs)): for i in range(len(seq)//batch_size): sidx = i*batch_size eidx = sidx + batch_size x = seq[sidx:eidx] x = torch.tensor(x).cuda() input_seq =torch.nn.utils.rnn.pack_padded_sequence(x,seq_lengths,batch_first = True) y = out_seq[sidx:eidx] output = model(input_seq) loss = criterion(output,y) loss.backward() optimizer.step() /usr/local/lib/python3.6/dist-packages/torch/nn/modules/module.py in __call__(self, *input, **kwargs) 487 result = self._slow_forward(*input, **kwargs) 488 else: --> 489 result = self.forward(*input, **kwargs) 490 for hook in self._forward_hooks.values(): 491 hook_result = hook(self, input, result) /usr/local/lib/python3.6/dist-packages/torch/nn/modules/rnn.py in forward(self, input, hx) 180 else: 181 result = _impl(input, batch_sizes, hx, self._flat_weights, self.bias, --> 182 self.num_layers, self.dropout, self.training, self.bidirectional) 183 output = result[0] 184 hidden = result[1:] if self.mode == 'LSTM' else result[1] RuntimeError: cuDNN error: CUDNN_STATUS_BAD_PARAM 3 Answers
I got the same error, if you switch to CPU, you'll get a much better description of the error. In my case the problem was in type of input that I was giving to the network. I was sending I guess long, while the model needed float. I made the following changes and the code worked. Basically switching to cpu gives better error descriptions.
input_seq = input_seq.float().cuda() I had the same issue and the problem was with torch==1.6. The solution can be found here git issue. Take a look. It may be your solution as well.
I encounter the same error. Here's the solution.
You should change the type of input from float64 to float32, which means you should type:
input_seq = input_seq.float() 
