Results 5 comments of Wenxuan Zhang

Hi 我刚临时配置了环境,跑了一轮,结果跟paper里是一致的,你要不再检查下具体的实验设置,特别是各个lib的版本,或者看看其他数据集上的结果是否一致。如果还是有问题,可以给我发邮件,谢谢

Thanks @kingbuff for answering the question. Since I did not use "do_eval" in the final experiment, I forgot to change the code of that part when releasing. You can modify...

你好,由于训练代价比较小,我们没有checkpoint提供。目前我们收到的复现反馈都是正常的,麻烦你再检查下实验设定等是否有问题,谢谢。

Thanks for pointing out this issue! We already conducted a post-processing to detect possible inconsistent placeholders. I will double-check the data to see whether we uploaded the wrong version for...

你好,由于暂时没有多语言多模态大模型,多模态部分我们目前只测试了英文数据(请看对应小节的介绍)。我们正在进行翻译相关的补充实验,后续可能会补充更多分析。