functionalize_lstm_fail.log

<eval_with_key>.0:5: UserWarning: RNN module weights are not part of single contiguous chunk of memory. This means they need to be compacted at every call, possibly greatly increasing memory usage. To compact weights again call flatten_parameters(). (Triggered internally at /scratch/dberard/dynamo38/pytorch/aten/src/ATen/native/cudnn/RNN.cpp:982.)
  lstm = torch.lstm(permute, (zeros, zeros_1), [self_model_lstm_lstm_flat_weights_0_, self_model_lstm_lstm_flat_weights_1_, self_model_lstm_lstm_flat_weights_2_, self_model_lstm_lstm_flat_weights_3_, self_model_lstm_lstm_flat_weights_4_, self_model_lstm_lstm_flat_weights_5_, self_model_lstm_lstm_flat_weights_6_, self_model_lstm_lstm_flat_weights_7_, self_model_lstm_lstm_flat_weights_8_, self_model_lstm_lstm_flat_weights_9_, self_model_lstm_lstm_flat_weights_10_, self_model_lstm_lstm_flat_weights_11_, self_model_lstm_lstm_flat_weights_12_, self_model_lstm_lstm_flat_weights_13_, self_model_lstm_lstm_flat_weights_14_, self_model_lstm_lstm_flat_weights_15_], True, 2, 0.0, True, True, False);  permute = zeros = zeros_1 = self_model_lstm_lstm_flat_weights_0_ = self_model_lstm_lstm_flat_weights_1_ = self_model_lstm_lstm_flat_weights_2_ = self_model_lstm_lstm_flat_weights_3_ = self_model_lstm_lstm_flat_weights_4_ = self_model_lstm_lstm_flat_weights_5_ = self_model_lstm_lstm_flat_weights_6_ = self_model_lstm_lstm_flat_weights_7_ = self_model_lstm_lstm_flat_weights_8_ = self_model_lstm_lstm_flat_weights_9_ = self_model_lstm_lstm_flat_weights_10_ = self_model_lstm_lstm_flat_weights_11_ = self_model_lstm_lstm_flat_weights_12_ = self_model_lstm_lstm_flat_weights_13_ = self_model_lstm_lstm_flat_weights_14_ = self_model_lstm_lstm_flat_weights_15_ = None
graph():
    %self_model_lstm_lstm_flat_weights_0_ : torch.Tensor [#users=1] = placeholder[target=self_model_lstm_lstm_flat_weights_0_]
    %self_model_lstm_lstm_flat_weights_1_ : torch.Tensor [#users=1] = placeholder[target=self_model_lstm_lstm_flat_weights_1_]
    %self_model_lstm_lstm_flat_weights_2_ : torch.Tensor [#users=1] = placeholder[target=self_model_lstm_lstm_flat_weights_2_]
    %self_model_lstm_lstm_flat_weights_3_ : torch.Tensor [#users=1] = placeholder[target=self_model_lstm_lstm_flat_weights_3_]
    %self_model_lstm_lstm_flat_weights_4_ : torch.Tensor [#users=1] = placeholder[target=self_model_lstm_lstm_flat_weights_4_]
    %self_model_lstm_lstm_flat_weights_5_ : torch.Tensor [#users=1] = placeholder[target=self_model_lstm_lstm_flat_weights_5_]
    %self_model_lstm_lstm_flat_weights_6_ : torch.Tensor [#users=1] = placeholder[target=self_model_lstm_lstm_flat_weights_6_]
    %self_model_lstm_lstm_flat_weights_7_ : torch.Tensor [#users=1] = placeholder[target=self_model_lstm_lstm_flat_weights_7_]
    %self_model_lstm_lstm_flat_weights_8_ : torch.Tensor [#users=1] = placeholder[target=self_model_lstm_lstm_flat_weights_8_]
    %self_model_lstm_lstm_flat_weights_9_ : torch.Tensor [#users=1] = placeholder[target=self_model_lstm_lstm_flat_weights_9_]
    %self_model_lstm_lstm_flat_weights_10_ : torch.Tensor [#users=1] = placeholder[target=self_model_lstm_lstm_flat_weights_10_]
    %self_model_lstm_lstm_flat_weights_11_ : torch.Tensor [#users=1] = placeholder[target=self_model_lstm_lstm_flat_weights_11_]
    %self_model_lstm_lstm_flat_weights_12_ : torch.Tensor [#users=1] = placeholder[target=self_model_lstm_lstm_flat_weights_12_]
    %self_model_lstm_lstm_flat_weights_13_ : torch.Tensor [#users=1] = placeholder[target=self_model_lstm_lstm_flat_weights_13_]
    %self_model_lstm_lstm_flat_weights_14_ : torch.Tensor [#users=1] = placeholder[target=self_model_lstm_lstm_flat_weights_14_]
    %self_model_lstm_lstm_flat_weights_15_ : torch.Tensor [#users=1] = placeholder[target=self_model_lstm_lstm_flat_weights_15_]
    %permute : [#users=1] = placeholder[target=permute]
    %zeros : [#users=1] = placeholder[target=zeros]
    %zeros_1 : [#users=1] = placeholder[target=zeros_1]
    %lstm : [#users=1] = call_function[target=torch.lstm](args = (%permute, (%zeros, %zeros_1), [%self_model_lstm_lstm_flat_weights_0_, %self_model_lstm_lstm_flat_weights_1_, %self_model_lstm_lstm_flat_weights_2_, %self_model_lstm_lstm_flat_weights_3_, %self_model_lstm_lstm_flat_weights_4_, %self_model_lstm_lstm_flat_weights_5_, %self_model_lstm_lstm_flat_weights_6_, %self_model_lstm_lstm_flat_weights_7_, %self_model_lstm_lstm_flat_weights_8_, %self_model_lstm_lstm_flat_weights_9_, %self_model_lstm_lstm_flat_weights_10_, %self_model_lstm_lstm_flat_weights_11_, %self_model_lstm_lstm_flat_weights_12_, %self_model_lstm_lstm_flat_weights_13_, %self_model_lstm_lstm_flat_weights_14_, %self_model_lstm_lstm_flat_weights_15_], True, 2, 0.0, True, True, False), kwargs = {})
    return (lstm,)
~~~~~~~~LSTM sanity check (FakeTensors)
~~~~~~~~LSTM sanity check PASSED
~~~~~~~~LSTM check with functional tensors
Traceback (most recent call last):
  File "repro3.py", line 63, in <module>
    traced_mod(*converted_args)
  File "/scratch/dberard/dynamo38/pytorch/torch/fx/graph_module.py", line 660, in call_wrapped
    return self._wrapped_call(self, *args, **kwargs)
  File "/scratch/dberard/dynamo38/pytorch/torch/fx/graph_module.py", line 279, in __call__
    raise e
  File "/scratch/dberard/dynamo38/pytorch/torch/fx/graph_module.py", line 269, in __call__
    return super(self.cls, obj).__call__(*args, **kwargs)  # type: ignore[misc]
  File "/scratch/dberard/dynamo38/pytorch/torch/nn/modules/module.py", line 1482, in _call_impl
    return forward_call(*args, **kwargs)
  File "<eval_with_key>.0", line 5, in forward
  File "/scratch/dberard/dynamo38/pytorch/torch/_subclasses/fake_tensor.py", line 886, in __torch_dispatch__
    r = func(*args, **kwargs)
  File "/scratch/dberard/dynamo38/pytorch/torch/_ops.py", line 285, in __call__
    return self._op(*args, **kwargs or {})
RuntimeError: Attempted to set the storage of a tensor on device "meta" to a storage on different device "cuda:0".  This is no longer allowed; the devices must match.