lyzltysgithub comments

Results 7 comments of


                                            lyzltysgithub

BGE中的retroMAE阶段，为什么encode阶段不用position embedding呢？

非常感谢大佬！想再请问大佬下： 1. enhanced decoding阶段（论文中C阶段）是对decoding阶段（论文中B阶段）的替代吗？ 2. BGE reranker的模型大小和参数量大概为多少 3. BGE embedding的pretrain阶段是否就是retroMAE的复用，是否有其他改进再次感谢您的解答

BGE-M3 不管是CPU还是GPU运行都很慢，这个怎么解决

作者您好，请问BGE有相关的性能测试报告吗，大概每秒多少token 或样本

BGE性能测试

感谢！

How to get the video of beergame？

I can't open this website, either. Could you please tell us the new website?

Training script takes more than 2 hours to finish

hello , I get the same result as you. It turned out that it took more than 2 hours to finish on flan-t5-base instead of 20 minutes. Could you please...

写到第2个问题时遇到报错“TypeError：列表索引必须是整数或切片，而不是str”

同样的问题，请问你解决了吗~

Support multiple treatments in Causal ML

> I think you could try upliftml, this is based on spark environment, and works for me it only support binary treatment