RuntimeError: mat1 dim 1 must match mat2 dim 0

[weida wang]

Hi, @bentrevett, I met another issue in running the colab of Alexnet.
My amount of classe is 4, so i modified the OUTPUT_DIM = 4 only without anything else changed, then when colab run to lrs, losses = lr_finder.range_test(train_iterator, END_LR, NUM_ITER), it happened following error:

RuntimeError                              Traceback (most recent call last)
<ipython-input-39-937508221c7d> in <module>()
      3 
      4 lr_finder = LRFinder(model, optimizer, criterion, device)
----> 5 lrs, losses = lr_finder.range_test(train_iterator, END_LR, NUM_ITER)

8 frames
<ipython-input-37-f3b559f5a5f9> in range_test(self, iterator, end_lr, num_iter, smooth_f, diverge_th)
     22         for iteration in range(num_iter):
     23 
---> 24             loss = self._train_batch(iterator)
     25 
     26             lrs.append(lr_scheduler.get_last_lr()[0])

<ipython-input-37-f3b559f5a5f9> in _train_batch(self, iterator)
     58         y = y.to(self.device)
     59 
---> 60         y_pred, _ = self.model(x)
     61 
     62         loss = self.criterion(y_pred, y)

/usr/local/lib/python3.7/dist-packages/torch/nn/modules/module.py in _call_impl(self, *input, **kwargs)
    725             result = self._slow_forward(*input, **kwargs)
    726         else:
--> 727             result = self.forward(*input, **kwargs)
    728         for hook in itertools.chain(
    729                 _global_forward_hooks.values(),

<ipython-input-32-1ea0008653c6> in forward(self, x)
     32         x = self.features(x)
     33         h = x.view(x.shape[0], -1)
---> 34         x = self.classifier(h)
     35         return x, h

/usr/local/lib/python3.7/dist-packages/torch/nn/modules/module.py in _call_impl(self, *input, **kwargs)
    725             result = self._slow_forward(*input, **kwargs)
    726         else:
--> 727             result = self.forward(*input, **kwargs)
    728         for hook in itertools.chain(
    729                 _global_forward_hooks.values(),

/usr/local/lib/python3.7/dist-packages/torch/nn/modules/container.py in forward(self, input)
    115     def forward(self, input):
    116         for module in self:
--> 117             input = module(input)
    118         return input
    119 

/usr/local/lib/python3.7/dist-packages/torch/nn/modules/module.py in _call_impl(self, *input, **kwargs)
    725             result = self._slow_forward(*input, **kwargs)
    726         else:
--> 727             result = self.forward(*input, **kwargs)
    728         for hook in itertools.chain(
    729                 _global_forward_hooks.values(),

/usr/local/lib/python3.7/dist-packages/torch/nn/modules/linear.py in forward(self, input)
     91 
     92     def forward(self, input: Tensor) -> Tensor:
---> 93         return F.linear(input, self.weight, self.bias)
     94 
     95     def extra_repr(self) -> str:

/usr/local/lib/python3.7/dist-packages/torch/nn/functional.py in linear(input, weight, bias)
   1688     if input.dim() == 2 and bias is not None:
   1689         # fused op is marginally faster
-> 1690         ret = torch.addmm(bias, input, weight.t())
   1691     else:
   1692         output = input.matmul(weight.t())

RuntimeError: mat1 dim 1 must match mat2 dim 0

Would you please give me some tips to debug it ? Thanks !

[Ben Trevett]

Hi @wwdok, sorry it took a while to get back to you. I'm unable to re-create your issue. What sizes images are used in your dataset?

[weida wang]

I also guess it maybe related with dataset, the image sizes are veried, the dataset is this.
To reproduce this issue, this is the colab notebook