JieN1
Results
2
comments of
JieN1
I also have a question about online inference. I notice that the evaluate() function in the training process randomly selects 100 items and then makes use of "argsort()" to rank...
> 你好，现在还不可以用hf的trainer。想实现的话需要修改一下hf的trainer训练循环中的backward和step，换成adalomo里的fused_backward。 请问adalomo除了collie中的实现，有类似lomo代码中的lomo_trainer的实现吗