F1124 15:33:06.386255 42445 smooth_l1_loss_layer.cpp:14] Check failed: bottom.size() == 4 (3 vs. 4) If weights are used, must specify both inside and outside weights

YuwenXiong / py-R-FCN

R-FCN with joint training and python support

MIT License

1.05k stars 471 forks source link

F1124 15:33:06.386255 42445 smooth_l1_loss_layer.cpp:14] Check failed: bottom.size() == 4 (3 vs. 4) If weights are used, must specify both inside and outside weights #29

Open rootlessweed opened 8 years ago

rootlessweed commented 8 years ago

when i use pre_trained_models/ResNet-50L. I got this erro. who knows why?

liyi14 commented 7 years ago

Hi, you'd better use SmoothL1LossOHEM rather than SmoothL1Loss in your .pt file, or define a bbox_outside_weights in input blobs. Unlike that in py-faster-rcnn, our SmoothL1LossOHEM's interface inherits from SmoothL1Loss used in fast-rcnn. Ours only has bbox_inside_weights and doesn't have bbox_outside_weights.