YOLOv5 compute loss代码注释

def compute_loss(p, targets, model):  # predictions, targets, model
    device = targets.device
    lcls, lbox, lobj = torch.zeros(1, device=device), torch.zeros(1, device=device), torch.zeros(1, device=device)
    tcls, tbox, indices, anchors = build_targets(p, targets, model)  # targets
    h = model.hyp  # hyperparameters

    # Define criteria，BCEWithLogitsLoss = sigmoid + BCELoss
    BCEcls = nn.BCEWithLogitsLoss(pos_weight=torch.Tensor([h['cls_pw']])).to(device)
    BCEobj = nn.BCEWithLogitsLoss(pos_weight=torch.Tensor([h['obj_pw']])).to(device)

    # Class label smoothing https://arxiv.org/pdf/1902.04103.pdf eqn 3
    cp, cn = smooth_BCE(eps=0.0)

    # Focal loss
    g = h['fl_gamma']  # focal loss gamma
    if g > 0:
        BCEcls, BCEobj = FocalLoss(BCEcls, g), FocalLoss(BCEobj, g)

    # Losses
    nt = 0  # number of targets
    no = len(p)  # number of outputs
    balance = [4.0, 1.0, 0.4] if no == 3 else [4.0, 1.0, 0.4, 0.1]  # P3-5 or P3-6
    for i, pi in enumerate(p):  # layer index, layer predictions
        b, a, gj, gi = indices[i]  # image, anchor, gridy, gridx
        tobj = torch.zeros_like(pi[..., 0], device=device)  # target obj

        n = b.shape[0]  # number of targets
        if n:
            nt += n  # cumulative targets
            ps = pi[b, a, gj, gi]  # prediction subset corresponding to targets，找到对应目标的xywh、类别、置信度

            # Regression
            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
            iou = bbox_iou(pbox.T, tbox[i], x1y1x2y2=False, CIoU=True)  # iou(prediction, target)，ciou computation，
            lbox += (1.0 - iou).mean()  # iou loss，lbox的loss值，由ciou决定

            # Objectness
            tobj[b, a, gj, gi] = (1.0 - model.gr) + model.gr * iou.detach().clamp(0).type(tobj.dtype)  # iou ratio

            # Classification
            if model.nc > 1:  # cls loss (only if multiple classes)
                t = torch.full_like(ps[:, 5:], cn, device=device)  # targets
                t[range(n), tcls[i]] = cp
                lcls += BCEcls(ps[:, 5:], t)  # BCE

            # Append targets to text file
            # with open('targets.txt', 'a') as file:
            #     [file.write('%11.5g ' * 4 % tuple(x) + '\n') for x in torch.cat((txy[i], twh[i]), 1)]

        lobj += BCEobj(pi[..., 4], tobj) * balance[i]  # obj loss

    s = 3 / no  # output count scaling
    lbox *= h['box'] * s
    lobj *= h['obj'] * s * (1.4 if no == 4 else 1.)
    lcls *= h['cls'] * s
    bs = tobj.shape[0]  # batch size

    loss = lbox + lobj + lcls
    return loss * bs, torch.cat((lbox, lobj, lcls, loss)).detach()

def build_targets(p, targets, model):
    # Build targets for compute_loss(), input targets(image,class,x,y,w,h)
    det = model.module.model[-1] if is_parallel(model) else model.model[-1]  # Detect() module
    na, nt = det.na, targets.shape[0]  # number of anchors, targets
    tcls, tbox, indices, anch = [], [], [], []
    gain = torch.ones(7, device=targets.device)  # normalized to gridspace gain
    ai = torch.arange(na, device=targets.device).float().view(na, 1).repeat(1, nt)  # same as .repeat_interleave(nt)
    targets = torch.cat((targets.repeat(na, 1, 1), ai[:, :, None]), 2)#将targets重复三遍，方便和3个anchor单独匹配  # append anchor indices

    g = 0.5  # bias，网格中心偏移
    off = torch.tensor([[0, 0],
                        [1, 0], [0, 1], [-1, 0], [0, -1],  # j,k,l,m
                        # [1, 1], [1, -1], [-1, 1], [-1, -1],  # jk,jm,lk,lm
                        ], device=targets.device).float() * g  # offsets，附近的4个网格

    for i in range(det.nl):#三个输出分支
        anchors = det.anchors[i]#当前分支的三个anchor
        gain[2:6] = torch.tensor(p[i].shape)[[3, 2, 3, 2]]  # xyxy gain，特征图大小

        # Match targets to anchors
        t = targets * gain#targets的xywh本身是归一化尺度，故需要变成特征图尺度
        if nt:
            # Matches
            r = t[:, :, 4:6] / anchors[:, None]  # wh ratio，计算targets和anchors的宽高比
            j = torch.max(r, 1. / r).max(2)[0] < model.hyp['anchor_t']  # compare
            # j = wh_iou(anchors, t[:, 4:6]) > model.hyp['iou_t']  # iou(3,n)=wh_iou(anchors(3,2), gwh(n,2))
            t = t[j]  # filter，仅留下满足上式即宽高比小于anchor_t的targets

            # Offsets，选择最近的3个网格
            gxy = t[:, 2:4]  # grid xy，targets索引2、3为targets的中心x、y坐标
            gxi = gain[[2, 3]] - gxy  # inverse
            j, k = ((gxy % 1. < g) & (gxy > 1.)).T
            l, m = ((gxi % 1. < g) & (gxi > 1.)).T
            j = torch.stack((torch.ones_like(j), j, k, l, m))
            t = t.repeat((5, 1, 1))[j]
            offsets = (torch.zeros_like(gxy)[None] + off[:, None])[j]
        else:
            t = targets[0]
            offsets = 0

        # Define
        b, c = t[:, :2].long().T  # image, class
        gxy = t[:, 2:4]  # grid xy
        gwh = t[:, 4:6]  # grid wh
        gij = (gxy - offsets).long()#当前label落在哪个网格上，通过long将其转化为整型。
        gi, gj = gij.T  # grid xy indices，代表格子的左上角坐标
        # Append
        a = t[:, 6].long()  # anchor indices
        indices.append((b, a, gj.clamp_(0, gain[3] - 1), gi.clamp_(0, gain[2] - 1)))  # image, anchor, grid indices
        tbox.append(torch.cat((gxy - gij, gwh), 1))  # box
        anch.append(anchors[a])  # anchors
        tcls.append(c)  # class

    return tcls, tbox, indices, anch