m 1 = [ a × b ] ⇒ [ b a t c h s i z e × i n f e a t u r e ] m_1 = [a \times b] \Rightarrow [batch size \times in feature] m1=[a×b]⇒[batchsize×infeature]
m 2 = [ c × d ] ⇒ [ b a t c h s i z e × i n f e a t u r e ] m_2 = [c \times d] \Rightarrow [batch size \times in feature] m2=[c×d]⇒[batchsize×infeature]
需要保持 b = c b=c b=c，避免报错
