beholder

Hi, @shuxjweb. Did you trained on SVHN dataset? I have trained on SVHN dataset with same experiment setting as the paper, but there is always 1% point accuracy lower than...

@zhanglaplace turn up minSize and thresholds to accelerate detection but at the cost of sacrificing recall.

@allenxcp the half-parameter version performs a bit worse than full-parameter one. you can adjust thresholds and minSize.

@gobigrassland Hi,I met the same question . Have you solved the problem ? Although my landmarks loss converge at last it descend so slowly and it does badly on my...