RuntimeError: 0D or 1D target tensor expected, multi-target not supported

[LukeLIN-web]

I can train Reddit, yelp, but I don't know how to train the amazon dataset.
The problem is the loss calculation.

class SAGE(torch.nn.Module):
    def __init__(self, in_channels, hidden_channels, out_channels,
                 num_layers=2):
        super(SAGE, self).__init__()
        self.num_layers = num_layers
        self.convs = torch.nn.ModuleList()
        self.convs.append(SAGEConv(in_channels, hidden_channels))
        for _ in range(self.num_layers - 2):
            self.convs.append(SAGEConv(hidden_channels, hidden_channels))
        self.convs.append(SAGEConv(hidden_channels, out_channels))

    def forward(self, x, adjs):
        for i, (edge_index, _, size) in enumerate(adjs):
            x_target = x[:size[1]]  # Target nodes are always placed first.
            x = self.convs[i]((x, x_target), edge_index)
            if i != self.num_layers - 1:
                x = F.relu(x)
                x = F.dropout(x, p=0.5, training=self.training)
        return x.log_softmax(dim=-1)

dataset = AmazonProducts('/data/AmazonProducts') 
... 
    model = SAGE(data.num_features, 256, dataset.num_classes)
....
        for seeds in train_loader:
            optimizer.zero_grad()
            n_id, batch_size, adjs = quiver_sampler.sample(seeds)
            target_node = n_id[:len(seeds)]
            out = model(x[n_id], adjs)  # forward
            # loss = F.nll_loss(out, y[target_node][:batch_size]) # reddit loss,  can work.
            # loss = F.cross_entropy(out, y[target_node][:batch_size]) #yelp loss  ,can work
            loss = F.nll_loss(out, y[target_node][:batch_size]) # amazon loss
            loss.backward()
            optimizer.step()

It error is :

    loss = F.nll_loss(out, y[target_node][:batch_size]) # amazon loss
  File "/opt/conda/lib/python3.8/site-packages/torch/nn/functional.py", line 2671, in nll_loss
    return torch._C._nn.nll_loss_nd(input, target, weight, _Reduction.get_enum(reduction), ignore_index)
RuntimeError: 0D or 1D target tensor expected, multi-target not supported

I tried ogbn-product, but also meet this problem.

[LukeLIN-web]

loss = F.cross_entropy(out, y[target_node][:batch_size].squeeze() ) # products loss