leepleased / negsampling-ner Goto Github PK

How can I get the file conlleval.pl？
Thank you!

你好，论文里 Figure 4: The misguidance rates of BERT Tagging and our model. 的图好像有点问题，看起来 Our Model 的 misguidance rate 随着 Masking Prob 的提高，降低了

I have arranged the folders as stated in the repo.
But i hav received this error.

1649, in gzopen
raise ReadError("not a gzip file")
tarfile.ReadError: not a gzip file

非常棒的工作！但我在复现您论文结果的时候遇到了一些困难，我想这可能是因为环境问题。
请问您还保留着您当时实验的环境文件吗？
同时祝您一切顺利。

按照Readme操作后出现此问题，请问应该怎么解决

train、dev、test数据用的都是示例提供的数据

[ 
 {
  "sentence": "['Somerset', '83', 'and', '174', '(', 'P.', 'Simmons', '4-38', ')', ',', 'Leicestershire', '296', '.']",
  "labeled entities": "[(0, 0, 'ORG'), (5, 6, 'PER'), (10, 10, 'ORG')]",
 },
 {
  "sentence": "['Leicestershire', '22', 'points', ',', 'Somerset', '4', '.']",
  "labeled entities": "[(0, 0, 'ORG'), (4, 4, 'ORG')]",
 }
]

Hi,

Interesting work! I am trying to reproduce your experiments was hoping you could answer a few questions:

• Could you please list any hyperparameter settings/commands you used for the conll2003 and ontonotes5 evaluations?
• I've found that default batch size is too large and I am getting CUDA OOM issues. What batch size did you use? (My GPU has 16gb ram, so I was also wondering what the gpu capacity was for the machines used to run the published experiments?)
• For ontonotes, some of the sentences are too long to fit into bert, did you just truncate these?

Thanks!

is there any code for inference??

hi where can i find the model.pt file?

您好，我使用您的代码在EC数据集上进行了测试，结果只有63点几，请问一下这是什么原因？

代码里的预训练模型为什么是pt格式，huggingface里的预训练语言模型都是bin格式的