search

Jianlong-Fu / Recurrent-Attention-CNN

225 stars 45 forks source link

Implement RA-CNN in tensorflow #9

Open Kongsea opened 6 years ago

Kongsea commented 6 years ago

How to implement formular (4) in the paper? If we have the three tensors, tx, ty and tl with the shape (None, 1) respectively, how do we get the corresponding mask M?

In python, we can use two loops as: for x in tx: for y in ty: mask[x, y] = hx(x, y)

Then how to implement this in TensorFlow? Thank you.

Kongsea commented 6 years ago

Maybe we need tensor operations or broadcasting? I'm not very familiar with that.

Kongsea commented 6 years ago

It seems it's my fault. I have used an image with unequal width and height, so it caused error when calculating Mask.

cocowf commented 6 years ago

@Kongsea Is it convenient for you to leave a contact?for qq or others

felix-duan commented 6 years ago

who has implemented the network by tensorflow?

Ostnie commented 6 years ago

@cocowf @felix-duan Do you have implemented the network by tensorflow? I have some trouble in APN ,and I described my question in Issue area. Could you please help me?

XZNWU commented 5 years ago

Do you have implemented the network by tensorflow? @cocowf @felix-duan @Ostnie