some questions about the structure of cornerNet-saccade

princeton-vl / CornerNet-Lite

BSD 3-Clause "New" or "Revised" License

1.78k stars 431 forks source link

some questions about the structure of cornerNet-saccade #152

Open liushuai839 opened 3 years ago

liushuai839 commented 3 years ago

Hi,Guys: Thanks for your code, I have some questions about the structure of cornerNet-saccade. In the paper, it says that the you will use the attention map to get the possible location of object and correspoinding scales. But i don't find it in the code. I hope you can point that where it is!!! Thanks

aksenventwo commented 3 years ago

Hi, see https://github.com/princeton-vl/CornerNet-Lite/blob/6a54505d830a9d6afe26e99f0864b5d06d0bbbaf/core/test/cornernet_saccade.py#L320

From the inferred code, attention map is used by default when using image inference for the first time. Secondly, attention map is no longer used. https://github.com/princeton-vl/CornerNet-Lite/blob/6a54505d830a9d6afe26e99f0864b5d06d0bbbaf/core/models/py_utils/modules.py#L272