dialogue_general_understanding linux python3 报错

[zhengya01]

train udc start.......... Traceback (most recent call last): File "main.py", line 23, in from train import do_train File "/ssd2/zhengya01/devlop_conv/test/models/PaddleNLP/PaddleDialogue/dialogue_general_understanding/train.py", line 29, in import dgu.reader as reader File "/ssd2/zhengya01/devlop_conv/test/models/PaddleNLP/PaddleDialogue/dialogue_general_understanding/dgu/reader.py", line 26, in reload(sys) NameError: name 'reload' is not defined train udc finish..........

[xyzhou-puck]

Thanks for reporting this errors, we will fix this problem very soon.

[zhengya01]

Thanks for reporting this errors, we will fix this problem very soon.

这个模型在paddle-develop版本上没有适配，报错信息如下：

---

Python Call Stacks (More useful to users):

File "/opt/_internal/cpython-3.7.0/lib/python3.7/site-packages/paddle/fluid/framework.py", line 2500, in append_op attrs=kwargs.get("attrs", None)) File "/opt/_internal/cpython-3.7.0/lib/python3.7/site-packages/paddle/fluid/layers/math_op_patch.py", line 231, in impl attrs={'axis': -1}) File "/ssd2/zhengya01/devlop_conv/test/models/PaddleNLP/PaddleDialogue/dialogue_general_understanding/dgu/transformer_encoder.py", line 119, in scaled_dot_product_attention product += attn_bias File "/ssd2/zhengya01/devlop_conv/test/models/PaddleNLP/PaddleDialogue/dialogue_general_understanding/dgu/transformer_encoder.py", line 148, in multi_head_attention dropout_rate) File "/ssd2/zhengya01/devlop_conv/test/models/PaddleNLP/PaddleDialogue/dialogue_general_understanding/dgu/transformer_encoder.py", line 274, in encoder_layer name=name + '_multi_head_att') File "/ssd2/zhengya01/devlop_conv/test/models/PaddleNLP/PaddleDialogue/dialogue_general_understanding/dgu/transformer_encoder.py", line 337, in encoder name=name + 'layer' + str(i)) File "/ssd2/zhengya01/devlop_conv/test/models/PaddleNLP/PaddleDialogue/dialogue_general_understanding/dgu/bert.py", line 144, in _build_model name='encoder') File "/ssd2/zhengya01/devlop_conv/test/models/PaddleNLP/PaddleDialogue/dialogue_general_understanding/dgu/bert.py", line 86, in init self._build_model(src_ids, position_ids, sentence_ids, input_mask) File "/ssd2/zhengya01/devlop_conv/test/models/PaddleNLP/PaddleDialogue/dialogue_general_understanding/dgu_net.py", line 49, in create_net use_fp16=False) File "/ssd2/zhengya01/devlop_conv/test/models/PaddleNLP/PaddleDialogue/dialogue_general_understanding/train.py", line 97, in do_train args=args) File "main.py", line 37, in do_train(args)

---

Error Message Summary:

Error: ShapeError: broadcast dimension mismatch. Operands could not be broadcast together with the shape of X = [-1, 12, 210, 210] and the shape of Y = [-1, 12, -1]. Received [210] in X is not equal to [12] in Y at (/paddle/paddle/fluid/operators/elementwise/elementwise_op_function.h:145) [operator < elementwise_add > error]