Created by: Karbit

环境和模型信息： PaddlePaddle 1.6, 使用 pip 安装 Paddle Models, 1.6, https://github.com/PaddlePaddle/models/blob/release/1.6/PaddleNLP/PaddleLARK/BERT/ 复现： 将 train.sh 中的 use_cuda 改为 false, 然后执行 ./train.sh -local y. 报错信息：

------------------------------------------------
Traceback (most recent call last):
  File "./train.py", line 439, in <module>
    train(args)
  File "./train.py", line 232, in train
    bert_config=bert_config)
  File "./train.py", line 124, in create_model
    use_fp16=args.use_fp16)
  File "/home/liujiaqiang/workdir/paddle-models/models-release-1.6/PaddleNLP/PaddleLARK/BERT/model/bert.py", line 81, in __init__
    self._build_model(src_ids, position_ids, sentence_ids, input_mask)
  File "/home/liujiaqiang/workdir/paddle-models/models-release-1.6/PaddleNLP/PaddleLARK/BERT/model/bert.py", line 139, in _build_m
    name='encoder')
  File "/home/liujiaqiang/workdir/paddle-models/models-release-1.6/PaddleNLP/PaddleLARK/BERT/model/transformer_encoder.py", line 3
    name=name + '_layer_' + str(i))
  File "/home/liujiaqiang/workdir/paddle-models/models-release-1.6/PaddleNLP/PaddleLARK/BERT/model/transformer_encoder.py", line 2
    name=name + '_multi_head_att')
  File "/home/liujiaqiang/workdir/paddle-models/models-release-1.6/PaddleNLP/PaddleLARK/BERT/model/transformer_encoder.py", line 1
    dropout_rate)
  File "/home/liujiaqiang/workdir/paddle-models/models-release-1.6/PaddleNLP/PaddleLARK/BERT/model/transformer_encoder.py", line 1
    out = layers.matmul(weights, v)
  File "/home/liujiaqiang/.jumbo/lib/python3.6/site-packages/paddle/fluid/layers/nn.py", line 6974, in matmul
    'alpha': float(alpha),
  File "/home/liujiaqiang/.jumbo/lib/python3.6/site-packages/paddle/fluid/layer_helper.py", line 43, in append_op
    return self.main_program.current_block().append_op(*args, **kwargs)
  File "/home/liujiaqiang/.jumbo/lib/python3.6/site-packages/paddle/fluid/framework.py", line 2426, in append_op
    attrs=kwargs.get("attrs", None))
  File "/home/liujiaqiang/.jumbo/lib/python3.6/site-packages/paddle/fluid/framework.py", line 1809, in __init__
    self.desc.infer_shape(self.block.desc)
paddle.fluid.core_avx.EnforceNotMet:

--------------------------------------------
C++ Call Stacks (More useful to developers):
--------------------------------------------
0   std::string paddle::platform::GetTraceBackString<std::string const&>(std::string const&&&, char const*, int)
1   paddle::platform::EnforceNotMet::EnforceNotMet(std::string const&, char const*, int)
2   paddle::operators::MatMulOp::InferShape(paddle::framework::InferShapeContext*) const
3   paddle::framework::OpDesc::InferShape(paddle::framework::BlockDesc const&) const

------------------------------------------
Python Call Stacks (More useful to users):
------------------------------------------
  File "/home/liujiaqiang/.jumbo/lib/python3.6/site-packages/paddle/fluid/framework.py", line 2426, in append_op
    attrs=kwargs.get("attrs", None))
  File "/home/liujiaqiang/.jumbo/lib/python3.6/site-packages/paddle/fluid/layer_helper.py", line 43, in append_op
    return self.main_program.current_block().append_op(*args, **kwargs)
  File "/home/liujiaqiang/.jumbo/lib/python3.6/site-packages/paddle/fluid/layers/nn.py", line 6974, in matmul
    'alpha': float(alpha),
  File "/home/liujiaqiang/workdir/paddle-models/models-release-1.6/PaddleNLP/PaddleLARK/BERT/model/transformer_encoder.py", line 1
    out = layers.matmul(weights, v)
  File "/home/liujiaqiang/workdir/paddle-models/models-release-1.6/PaddleNLP/PaddleLARK/BERT/model/transformer_encoder.py", line 1
    dropout_rate)
  File "/home/liujiaqiang/workdir/paddle-models/models-release-1.6/PaddleNLP/PaddleLARK/BERT/model/transformer_encoder.py", line 2
    name=name + '_multi_head_att')
  File "/home/liujiaqiang/workdir/paddle-models/models-release-1.6/PaddleNLP/PaddleLARK/BERT/model/transformer_encoder.py", line 3
    name=name + '_layer_' + str(i))
  File "/home/liujiaqiang/workdir/paddle-models/models-release-1.6/PaddleNLP/PaddleLARK/BERT/model/bert.py", line 139, in _build_m
    name='encoder')
  File "/home/liujiaqiang/workdir/paddle-models/models-release-1.6/PaddleNLP/PaddleLARK/BERT/model/bert.py", line 81, in __init__
    self._build_model(src_ids, position_ids, sentence_ids, input_mask)
  File "./train.py", line 124, in create_model
    use_fp16=args.use_fp16)
  File "./train.py", line 232, in train
    bert_config=bert_config)