Egqawkq
Egqawkq
> I also want to ask whether you have known how to deal the question? Sorry, this problem was ignored by me and the whole project can still run well....
 See fpn_resnet.py function `apply_kfpn`
@Laihu08 Have you solved this problem using the author's script? I want to convert KITTI object dataset into depth too.
@JUGGHM Thanks for your reply! So, If I still using your pretrained model for test, I need to resize all my dataset into the same size? Currently, I just want...
@EmanuelaBoros I have met the same problem, do you solved this problem later? I found conll03 datatset have some problems.
@EeyoreLee 本仓库的代码精度能对齐论文精度吗，我尝试了6个数据集，一般会相差1个点
@gjy-code Have you solved problem? Do you use `robetra_wwm_ext_large ` as pre-trained model? I training BERT-Tagger on zh_MSRA and the dev result is 0.94+ which seems fine to me.
@Josson 我也是cuda11，似乎没有什么问题训练。 顺便问一下您之后如果跑通之后精度能复现对齐吗，多个数据集我都训过，精度总是差一点点，大概一个点
@Josson 挺奇怪，我其实是直接用的提供的脚本训练的。 wx：17714924649
@zhanghaok 我个人习惯是只使用pt1.7（17我从来没遇到什么bug），没记错的话就是1.7版本，你试试看