Unable to use EdgeDataLoader with GraphConvolution

[TudorAndrei]

I get the following error:

Traceback (most recent call last):
  File "/home/tudor/projects/xtailab/task2_graph_attention/link-prediction-sampler.py", line 64, in <module>
    pos_score, neg_score = model(positive_graph, negative_graph, blocks, node_feats, node_type)
  File "/home/tudor/.miniconda3/envs/xtailab/lib/python3.9/site-packages/torch/nn/modules/module.py", line 889, in _call_impl
    result = self.forward(*input, **kwargs)
  File "/home/tudor/projects/xtailab/task2_graph_attention/model_multi.py", line 249, in forward
    h = self.rgcn(blocks, x)
  File "/home/tudor/.miniconda3/envs/xtailab/lib/python3.9/site-packages/torch/nn/modules/module.py", line 889, in _call_impl
    result = self.forward(*input, **kwargs)
  File "/home/tudor/projects/xtailab/task2_graph_attention/model_multi.py", line 212, in forward
    x = self.conv1(blocks[0], x)
  File "/home/tudor/.miniconda3/envs/xtailab/lib/python3.9/site-packages/torch/nn/modules/module.py", line 889, in _call_impl
    result = self.forward(*input, **kwargs)
  File "/home/tudor/.miniconda3/envs/xtailab/lib/python3.9/site-packages/dgl/nn/pytorch/hetero.py", line 176, in forward
    dstdata = self.mods[etype](
  File "/home/tudor/.miniconda3/envs/xtailab/lib/python3.9/site-packages/torch/nn/modules/module.py", line 889, in _call_impl
    result = self.forward(*input, **kwargs)
  File "/home/tudor/.miniconda3/envs/xtailab/lib/python3.9/site-packages/dgl/nn/pytorch/conv/graphconv.py", line 417, in forward
    graph.srcdata['h'] = feat_src
  File "/home/tudor/.miniconda3/envs/xtailab/lib/python3.9/site-packages/dgl/view.py", line 81, in __setitem__
    self._graph._set_n_repr(self._ntid, self._nodes, {key : val})
  File "/home/tudor/.miniconda3/envs/xtailab/lib/python3.9/site-packages/dgl/heterograph.py", line 3992, in _set_n_repr
    raise DGLError('Expect number of features to match number of nodes (len(u)).'
dgl._ffi.base.DGLError: Expect number of features to match number of nodes (len(u)). Got 1893 and 349 instead.

Process finished with exit code 1

This happends because in the GraphConv Layer the following update is being made

                graph.srcdata['h'] = feat_src

Here graph is a block.

[TudorAndrei]

I think I know what my issue could be, here

    def forward(self, blocks, x):
        x = self.conv1(blocks[0], x)
        x = self.conv2(blocks[1], x)
        return 

I send for the first conv1 all the node features, while blocks[0] doesnt’ have all of them. so the feat_src ends up being(len(nodes), hid_feat), so 1893, while it’s supposed to have 349.

Edit: the mistake was that the input wasnt take from the first block, but from the whole graph.