two modalities are separated by the [SEP] token,the numbers in each attention module represent the Eq. number.
E
h
_h
h? is the mean,
?
\nabla
?A :=
?
y
t
?
A
{?y_t}\over?A
?A?yt??for
y
t
y_t
yt? which is the model’s output.
⊙
\odot
⊙ is the Hadamard product,remove the negative contributions before averaging.
aggregated self-attention matrix R
q
q
^{qq}
qq,previous layers’ mixture of context is embodied by R
q
k
^{qk}
qk.
作者的实验在coco和ImageNet验证集上做的,不好follow