I have no idea why I ran into this error, here is the code snippet:
def message(self, x_j: Tensor, alpha_j: Tensor, alpha_i: OptTensor, alpha_e,index: Tensor, ptr: OptTensor, size_i: Optional[int]) -> Tensor:
alpha = alpha_j if alpha_i is None else alpha_j + alpha_i
alpha += alpha_e # add edge features...
alpha = F.leaky_relu(alpha, self.negative_slope)
print('here')
alpha = softmax(alpha, index, ptr, size_i)
print('pass')
self._alpha = alpha
alpha = F.dropout(alpha, p=self.dropout, training=self.training)
The softmax function came from from torch_geometric.utils import softmax. The error pops at the softmax line, interesting it can always execute one time before there is an error. How can I fix this?
Here is the softmax code in torch_geometric
@torch.jit.script
def softmax(src: Tensor, index: Optional[Tensor] = None,
ptr: Optional[Tensor] = None, num_nodes: Optional[int] = None,
dim: int = 0) -> Tensor:
r""
Computes a sparsely evaluated softmax.
Given a value tensor :attr:`src`, this function first groups the values
along the first dimension based on the indices specified in :attr:`index`,
and then proceeds to compute the softmax individually for each group.
Args:
src (Tensor): The source tensor.
index (LongTensor, optional): The indices of elements for applying the
softmax. (default: :obj:`None`)
ptr (LongTensor, optional): If given, computes the softmax based on
sorted inputs in CSR representation. (default: :obj:`None`)
num_nodes (int, optional): The number of nodes, *i.e.*
:obj:`max_val + 1` of :attr:`index`. (default: :obj:`None`)
dim (int, optional): The dimension in which to normalize.
(default: :obj:`0`)
:rtype: :class:`Tensor`
""
if ptr is not None:
dim = dim + src.dim() if dim < 0 else dim
size = ([1] * dim) + [-1]
ptr = ptr.view(size)
src_max = gather_csr(segment_csr(src, ptr, reduce='max'), ptr)
out = (src - src_max).exp()
out_sum = gather_csr(segment_csr(out, ptr, reduce='sum'), ptr)
elif index is not None:
N = maybe_num_nodes(index, num_nodes)
src_max = scatter(src, index, dim, dim_size=N, reduce='max')
src_max = src_max.index_select(dim, index)
out = (src - src_max).exp()
out_sum = scatter(out, index, dim, dim_size=N, reduce='sum')
out_sum = out_sum.index_select(dim, index)
else:
raise NotImplementedError
return out / (out_sum + 1e-16)