pony
pony
how to infer one image？ Still need 4 or 8 GPU？I just wanna infer.
HI, Thanks for your job and Congrats. I know you will private more useful code, waiting~ I have saw "aug_model" on tensorborad, it's very complex. So, it would be great...
tools/utils.py文件: ```python def homograpthy_g2im(cam_pitch, cam_height, K): # transform top-view region to original image region R_g2c = np.array([[1, 0, 0], [0, np.cos(np.pi / 2 + cam_pitch), -np.sin(np.pi / 2 + cam_pitch)],...
https://github.com/yuliangguo/Pytorch_Generalized_3D_Lane_Detection/blob/5bb190ed9f0dc211cd6265ab30c151d3699d3ed4/dataloader/Load_Data_3DLane_ext.py#L104 should be `self.anchor_y_steps - self.y_ref`
https://github.com/hhaAndroid/mmdetection-mini/blob/3858c8c2f071c064fe78ce24088a1c9815ae1a21/mmdet/det_core/bbox/assigners/grid_assigner.py#L140 这里为什么要+1呢? 我理解argmax_overlaps里存储了max_overlaps所对应的gt_box的idx, 直接赋值不就好了吗, +1是出于什么考虑呢? 而且在同文件的161~162行进行分配label的时候又去-1已获得gt_box idx, 感觉有点多此一举啊, 但感觉肯定不是没有原因的,所以想问下 ```python if pos_inds.numel() > 0: assigned_labels[pos_inds] = gt_labels[ assigned_gt_inds[pos_inds] - 1] # 正样本anchor处的label值 ```