報錯indexerror: tensors used as indices must be long, byte or bool tensors

報錯

Traceback (most recent call last):
  File "main.py", line 306, in <module>
    args_config=args_config,
  File "main.py", line 224, in train
    avg_reward,
  File "main.py", line 56, in train_one_epoch
    selected_neg_items_list, _ = sampler(batch_data, adj_matrix, edge_matrix)
  File "/home/zzy/anaconda3/lib/python3.7/site-packages/torch/nn/modules/module.py", line 493, in __call__
    result = self.forward(*input, **kwargs)
  File "/home/zzy/KG-Policy/kgpolicy/modules/sampler/kgpolicy.py", line 91, in forward
    one_hop, one_hop_logits = self.kg_step(pos, users, adj_matrix, step=1)
  File "/home/zzy/KG-Policy/kgpolicy/modules/sampler/kgpolicy.py", line 139, in kg_step
    i_e = gcn_embedding[one_hop]
IndexError: tensors used as indices must be long, byte or bool tensors

定位到kgpolicy檔案的第139行

    one_hop = adj_matrix[pos]
    i_e = gcn_embedding[one_hop] 

感覺索引indices是one_hop ,檢查下one_hop:

print(one_hop)

結果為:

tensor([[ 1295.,  6801.,  6590.,  ..., 43463., 63427., 23697.],
        [ 2940., 23298.,  4720.,  ..., 54026., 45077., 68521.],
        [  136.,   137.,  1033.,  ..., 47605., 59133., 51274.],
        ...,
        [ 1629.,  6126.,   118.,  ..., 56004., 23626., 41774.],
        [12689.,  6415., 21290.,  ..., 28091., 24405., 37709.],
        [  739.,   818.,  1202.,  ..., 54062., 29978., 25789.]])

因為預設的是float型別,所以不對,就直接型別轉換為long就可以了.
關於究竟轉化為哪種型別,參考連結:
https://blog.csdn.net/junqing_wu/article/details/99692296
https://blog.csdn.net/jacke121/article/details/82703640
https://blog.csdn.net/weixin_38314865/article/details/105949825

        one_hop = adj_matrix[pos].type(torch.long)#已修改
        i_e = gcn_embedding[one_hop] 

ok,這樣就可以啦…這麼個錯誤弄了一上午,崩潰…希望能幫到小可愛們~