Charles Shang
Charles Shang
@1292765944 1. anchors outside the image: There is a param: allow_border controlls how many pixels an anchor can spread outside @jasonhsiao97 1. RoI Cropping w.r.t. roi height and width. I...
@so-as hi, Have you already done `detection + landmark` pipeline? If so, would you plz share your work? It would be very helpful! Thanks!!
Was it solved? There must be a typo in the test net. I'll run some debug in my spare time
I'm sorry about it. I must commit a wrong config last month. And I'm also trying to locate the problem. Suggestions: 1. check the recall rate for rpn outputs: write...
Hi, @Esugi8 Did you figure out this problem? ResourceExhaustedError sounds like your GPU does not have enough memories. You can try to set `TRAIN.SCALES` to a small number, like 400.
1. train + val 2. I found out with or without hard samples the mAPs are roughly the same.. Maybe more efforts should be invested
in eval_ssd_network.py, calling `slim.evaluation.evaluation_loop` with ``` eval_op=flatten(list(names_to_updates.values())), ``` don't forget to import flatten with `from compiler.ast import flatten`