提交 a10caf7c 编写于 作者: Y Yibing Liu

Fix typos in label_smooth_op

上级 5ed07ef1
......@@ -79,8 +79,7 @@ it is not guaranteed to generalize. Second, it encourages the differences
between the largest logit and all others to become large, reducing the ability
of the model to adapt. Label smoothing is proposed to encourage the model to
be less confident, which replaces the ground-truth label $y$ with the weighted
sum of itselft and some fixed distribution $\mu$,
i.e.
sum of itself and some fixed distribution $\mu$, i.e.
$$
\tilde{y} = (1 - \epsilon) * y + \epsilon * \mu,
......
Markdown is supported
0% .
You are about to add 0 people to the discussion. Proceed with caution.
先完成此消息的编辑!
想要评论请 注册