MessagePassing.propagate issue #26
Description
Traceback (most recent call last):
File "/data1/supersmpl/dy/ReCurrence/HOOD/demo.py", line 114, in
trajectories_dict = runner.valid_rollout(sequence, bare=True) "amax" 15:09 23-9月-24
File "/data1/ReCurrence/HOOD/runners/postcvpr.py", line 114, in valid_rollout
trajectory, obstacle_trajectory, metrics_dict = self._rollout(sequence, n_samples,
File "/data1/ReCurrence/HOOD/runners/postcvpr.py", line 149, in _rollout
state = self.model(state, is_training=False)
File "/data1/.conda/envs/hood/lib/python3.9/site-packages/torch/nn/modules/module.py", line 1102, in _call_impl
return forward_call(*input, **kwargs)
File "/data1/ReCurrence/HOOD/models/postcvpr.py", line 488, in forward
sample = self._learned_model(sample)
File "/data1.conda/envs/hood/lib/python3.9/site-packages/torch/nn/modules/module.py", line 1102, in _call_impl
return forward_call(*input, **kwargs)
File "/data1/ReCurrence/HOOD/models/core/postcvpr.py", line 410, in forward
sample = self._process(sample)
File "/data1/ReCurrence/HOOD/models/core/postcvpr.py", line 400, in _process
sample = processing_block(sample)
File "/data1/.conda/envs/hood/lib/python3.9/site-packages/torch/nn/modules/module.py", line 1102, in _call_impl
return forward_call(*input, **kwargs)
File "/data1/ReCurrence/HOOD/models/core/base.py", line 26, in forward
sample = self.propagate(sample)
File "/data1/.cache/pyg/message_passing/models.core.postcvpr_GraphNetBlock_propagate.py", line 142, in propagate
mutable_size = self._check_input(edge_index, size)
File "/data1/.conda/envs/hood/lib/python3.9/site-packages/torch_geometric/nn/conv/message_passing.py", line 277, in _check_input
raise ValueError(
ValueError: MessagePassing.propagate only supports integer tensors of shape [2, num_messages], torch_sparse.SparseTensor or torch.sparse.Tensor for argument edge_index.