nn.Sequential giving error while nn.Linear is not

I am kind of newbie to pytorch. I am trying to change the head of transformer which was:

lm_head = nn.Linear(768, 32128)

For this part, the model runs without error, However when I tried to add another fully connected layer using the Sequential function like:

lm_head = nn.Sequential(
nn.Linear(768, 32128),
nn.ReLU,
nn.Linear(768, 32128)
)

I am using a batch size of 32, and I am getting the following error:

RuntimeError: mat1 and mat2 shapes cannot be multiplied (32x32128 and 768x32128)

shouldnt both of these format give me the same output? I am confused!

As the error message explains, the out_features of the first linear layer in the nn.Sequential block do not match the in_features of the second linear layer.
A fix would be:

lm_head = nn.Sequential(
    nn.Linear(768, 32128),
    nn.ReLU,
    nn.Linear(32128, 32128)
)