Bharat Singh

I think if you put a random seed, you can control it. I have not tested it though.

you can use ps aux | grep caffe, get the pids and use kill -9 pid. that should work.

happy to know!

NMS should work fine if there is no overlap. Although Soft-NMS would also give the same results.

you can use the python code def rescore(overlap, scores, thresh, type='gaussian'): assert overlap.shape[0] == scores.shape[0] if type == 'linear': inds = np.where(overlap >= thresh)[0] scores[inds] = scores[inds] * (1 -...

Should be length, check the msra repo, I pasted from there On Mon, May 21, 2018 at 9:28 AM niuniu111 wrote: > Thank you for your reply. I have another...

i'll put it in 2-3 weeks.

we mention in the the paper why it is not good for ssd/yolo. you can still get some improvement at 0.75 iou though

If you don't do NMS for 2 stage detectors like faster-rcnn and 1 stage detectors like yolo/ssd, and compare recall, you would observe a gap in recall (2 stage >...

We have not released the GPU version yet. It will be released in a month or so.