Huiqiang Jiang

The ace04 log is shown in that: I forgot to save the log of ACE05, I'll update when I rerun the experiment. ``` Better speed can be achieved with apex...

> 1、Do you use BERT_base instead of BERT_large? > we used BERT_base > > 2、Do you have some other trick not in the above processing? > No. I think there...

No hurry for replying due to the DDL and only to record this experiment's situation. --- After evaluation, I also find some problems. 1. It seems like your code has...

> @iofu728 > hello， I experimented on the ACE2004 data set provided by the author. The training effect is not as good as yours, only 50% F1. I suspect it...

Hi @wjczf123, yeah. If you remove the [#384-385](https://github.com/microsoft/vert-papers/blob/master/papers/DecomposedMetaNER/learner.py#L384-L385) and [#447](https://github.com/microsoft/vert-papers/blob/master/papers/DecomposedMetaNER/learner.py#L447) of learner.py, the code'll skip fine-tuning on meta-test support set.

> Thanks for your reply. I ran it once under inter 5-way 1-shot setting and the results looked very bad. > > 2022-09-20 22:57:35 INFO: - span_f1 = 0.7218073781712385 2022-09-20...

Hi @wjczf123, this may be reasonable, although we have not done the corresponding ablation experiments on 5shot. First of all the 5shot and 1shot datasets cannot be compared in parallel,...

Hi @dongguanting, We don't have the copyright of Few-NERD dataset. Please contact the owner of this dataset. We already clear the Few-NERD version in our paper footnote 5. And we...

Hi @dongguanting, not really, in the Cross-Domain dataset, you only need to train once on the training set (Span+Type) and then evaluate it directly. In the training phase, the model...

Hi @liyongqi2002, thanks for the reminder. We have some problems with the presentation of the Few-NERD dataset version. I will fix it as soon as possible. In fact, the first...