I want to know some names in this program.

[ZeKunZhang1998]

❔Question

if n: nt += n # cumulative targets ps = pi[b, a, gj, gi] # prediction subset corresponding to targets

Regression

        t_box = ps[:,:4]

        pxy = ps[:, :2].sigmoid() * 2. - 0.5
        pwh = (ps[:, 2:4].sigmoid() * 2) ** 2 * anchors[i]

        pbox = torch.cat((pxy, pwh), 1).to(device)  # predicted box
        #print(pbox)
        giou = bbox_iou(pbox.T, tbox[i], x1y1x2y2=False, CIoU=True)  # giou(prediction, target)
        #print(tbox[i])
        print(len(giou))
        lbox += (1.0 - giou).mean()  # giou loss

Additional context

Could you tell me the meaning of pi,ps,pbox?

[glenn-jocher]

@ZeKunZhang1998 the info you're looking for is in the comments in the code you pasted.

[github-actions[bot]]

This issue has been automatically marked as stale because it has not had recent activity. It will be closed if no further activity occurs. Thank you for your contributions.

[wandaoyi]

pi: there are 3 anchors in a coord, pi is a predict info under an anchor. ps: is a pre subset info pbox: pre_box info ---> [[center_x, center_y, width, height], ...]

[wandaoyi]

However, I want to know why: pxy = ps[:, :2].sigmoid() 2. - 0.5 pwh = (ps[:, 2:4].sigmoid() 2) * 2 anchors[i]

do 2. - 0.5 and 2) ** 2 these things. I wonder!