pytroch==1.6.0 需要注释extended_attention_mask = extended_attention_mask.to(dtype=next(self.parameters()).dtype)才能运行 about cluener2020

[zhangyifei1]

大佬，注释源码是不是太暴力了？有木有其他的解决方案？我也遇到这个问题了

[1a2cjitenfei]

大哥，按照你的说法注释完后出现
warnings.warn('Was asked to gather along dimension 0, but all '
Traceback (most recent call last):
File "run_ner_span.py", line 541, in
main()
File "run_ner_span.py", line 483, in main
global_step, tr_loss = train(args, train_dataset, model, tokenizer)
File "run_ner_span.py", line 128, in train
loss.backward()
File "/data/jtf/Minicoda3/envs/python3_8/lib/python3.8/site-packages/torch/_tensor.py", line 255, in backward
torch.autograd.backward(self, gradient, retain_graph, create_graph, inputs=inputs)
File "/data/jtf/Minicoda3/envs/python3_8/lib/python3.8/site-packages/torch/autograd/init.py", line 147, in backward
Variable._execution_engine.run_backward(
File "/data/jtf/Minicoda3/envs/python3_8/lib/python3.8/site-packages/torch/autograd/function.py", line 87, in apply
return self._forward_cls.backward(self, *args) # type: ignore[attr-defined]
File "/data/jtf/Minicoda3/envs/python3_8/lib/python3.8/site-packages/torch/nn/parallel/_functions.py", line 34, in backward
return (None,) + ReduceAddCoalesced.apply(ctx.input_device, ctx.num_inputs, *grad_outputs)
File "/data/jtf/Minicoda3/envs/python3_8/lib/python3.8/site-packages/torch/nn/parallel/functions.py", line 45, in forward
return comm.reduce_add_coalesced(grads, destination)
File "/data/jtf/Minicoda3/envs/python3_8/lib/python3.8/site-packages/torch/nn/parallel/comm.py", line 143, in reduce_add_coalesced
flat_result = reduce_add(flat_tensors, destination)
File "/data/jtf/Minicoda3/envs/python3_8/lib/python3.8/site-packages/torch/nn/parallel/comm.py", line 96, in reduce_add
nccl.reduce(inputs, output=result, root=root_index)
File "/data/jtf/Minicoda3/envs/python3_8/lib/python3.8/site-packages/torch/cuda/nccl.py", line 89, in reduce
torch._C._nccl_reduce(inputs, _output, root, op, streams, comms)
RuntimeError: NCCL Error 2: unhandled system error