-
when using the script marcov1_to_dureader.py to convert MARCOv1 to dureader, it failed because ValueError: Trailing data
-
NLP/DuReader-Robust-BASELINE的训练程序,单卡时正常运行,多卡时则会报错,具体信息如下:
--------------------------------------------
C++ Call Stacks (More useful to developers):
--------------------------------------------
0 …
-
I have Dureader2 dataset.
But I don't know what Meaning of test1, test2.
I see Dureader Preprocessed data's train, dev test folder all has two files: search, zhidao.
What test1, test2 means?
T…
-
Hello, the data set of the Chinese competition can not be downloaded. Can you provide a copy? Thank you~
My email: 470294527@qq.com
-
I inspected the json format and prepared with labeled answers.
-
-
请问ERNIE在DuReader实验的相关代码是会放出?
`ERNIE/run_mrc.py`中是针对Squad数据格式的。而DuReader是多文档的阅读理解数据集,和Squad不同。不知道这部分是如何处理的?
-
Traceback (most recent call last):
File "utils/preprocess.py", line 217, in
find_fake_answer(sample)
File "utils/preprocess.py", line 158, in find_fake_answer
for p_idx, para_tokens i…
-
感谢楼主的code和分享。执行sh run.sh --para_extraction报错,这个因为官方是用的python2.7。用py3报错信息为: __init__() got an unexpected keyword argument 'encoding'
,把json.dumps()改成python3版本后,解决这个错误了。
但依然不能执行,这个代码if p_idx < para_id…
-
感谢分享,predict_dureader.py文件是空的;此外prepare_submit.py 提交文件类型之类的都是空的?