RuntimeError: shape mismatch: value tensor of shape [24, 0] cannot be broadcast to indexing result of shape [24, 10]

[Linawdjn]

Traceback (most recent call last):
File "/home/nudt/linyawei/AStarNet/script/visualize.py", line 175, in
visualize(solver, solver.test_set[i], entity_vocab, relation_vocab, output_file="/home/nudt/linyawei/AStarNet/results.txt")
File "/home/nudt/linyawei/AStarNet/script/visualize.py", line 112, in visualize
paths, weights, num_steps = solver.model.visualize(vis_batch)
File "/home/nudt/anaconda3/envs/astar/lib/python3.10/site-packages/torchdrug-0.2.1-py3.10.egg/torchdrug/tasks/reasoning.py", line 222, in visualize
return self.model.visualize(self.fact_graph, h_index, t_index, r_index)
File "/home/nudt/linyawei/AStarNet/reasoning/model.py", line 404, in visualize
lengths, source_indexes = self.beam_search_length(graph, subgraphs, graph.num_node, h_index, t_index)
File "/home/nudt/linyawei/AStarNet/reasoning/model.py", line 449, in beam_search_length
length[node_out] = sub_length[out_mask]
File "/home/nudt/linyawei/AStarNet/reasoning/data.py", line 81, in setitem
new_values[inverse[len(self.keys):]] = values
RuntimeError: shape mismatch: value tensor of shape [24, 0] cannot be broadcast to indexing result of shape [24, 10]
I have encountered this problem on visualize.py, how do I solve it?