Skip to content
New issue

Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.

By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.

Already on GitHub? Sign in to your account

testing code does not work, the results are not correct, mAP values are easy 0.125, medium 0.134, hard 0.156 #8

Open
hdjsjyl opened this issue Oct 18, 2018 · 1 comment

Comments

@hdjsjyl
Copy link

hdjsjyl commented Oct 18, 2018

Both model I trained by this code or your pretrained model doesnot work for the testing code. And I saw your SSH folder in the insightface project. The testing code of that folder detects faces via original image, I do not think, it can achieve 81.4% of mAP for hard set of validation set of WIDER FACE. So could you provide some advice? Thanks

@hdjsjyl
Copy link
Author

hdjsjyl commented Oct 18, 2018

@nttstar when I change the function 'generate_anchors_fpn' to this:
def generate_anchors_fpn(base_size, scales, ratios):
anchors = []
_ratios = ratios.reshape((len(base_size), -1))
_scales = scales.reshape((len(base_size), -1))
for i, bs in enumerate(base_size):
__ratios = _ratios[i]
__scales = _scales[i]
# print('anchors_fpn', bs, __ratios, __scales, file=sys.stderr)
r = generate_anchors(bs, __ratios, __scales)
# print('anchors_fpn', r.shape, file=sys.stderr)
anchors.append(r)

return  anchors

It works. But The results of your pretrained model with single scale(1200, 1600) testing, the mAP value for hard set of validation set of WIDER FACE only achieves 78.2. I do not know the reason, do you have some suggestions? thank you.

Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment
Labels
None yet
Projects
None yet
Development

No branches or pull requests

1 participant