yolov7的loss计算代码详解

 yolov7最重要的创新就是其网络设计和正负样本分配策略，下面的代码注释解释了yolov7的样本分配策略，帮助理解

class ComputeLossOTA:
    # Compute losses
    def __init__(self, model, autobalance=False):
        super(ComputeLossOTA, self).__init__()
        device = next(model.parameters()).device  # get model device
        h = model.hyp  # hyperparameters

        # Define criteria
        BCEcls = nn.BCEWithLogitsLoss(pos_weight=torch.tensor([h['cls_pw']], device=device))
        BCEobj = nn.BCEWithLogitsLoss(pos_weight=torch.tensor([h['obj_pw']], device=device))

        # Class label smoothing https://arxiv.org/pdf/1902.04103.pdf eqn 3
        self.cp, self.cn = smooth_BCE(eps=h.get('label_smoothing', 0.0))  # positive, negative BCE targets

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

        det = model.module.model[-1] if is_parallel(model) else model.model[-1]  # Detect() module
        self.balance = {3: [4.0, 1.0, 0.4]}.get(det.nl, [4.0, 1.0, 0.25, 0.06, .02])  # P3-P7
        self.ssi = list(det.stride).index(16) if autobalance else 0  # stride 16 index
        self.BCEcls, self.BCEobj, self.gr, self.hyp, self.autobalance = BCEcls, BCEobj, model.gr, h, autobalance
        for k in 'na', 'nc', 'nl', 'anchors', 'stride':
            setattr(self, k, getattr(det, k))

    def __call__(self, p, targets, imgs):  # predictions, targets, model   
        device = targets.device
        lcls, lbox, lobj = torch.zeros(1, device=device), torch.zeros(1, device=device), torch.zeros(1, device=device) #初始化0,0,0
        bs, as_, gjs, gis, targets, anchors = self.build_targets(p, targets, imgs) # 返回匹配到的image index, anchor index, gj, gi, GT, anchor
        pre_gen_gains = [torch.tensor(pp.shape, device=device)[[3, 2, 3, 2]] for pp in p] # [[80,80,80,80], [40,40,40,40], [20,20,20,20]]
    

        # Losses
        for i, pi in enumerate(p):  # layer index, layer predictions
            b, a, gj, gi = bs[i], as_[i], gjs[i], gis[i]  # image, anchor, gridy, gridx
            tobj = torch.zeros_like(pi[..., 0], device=device)  # target obj, pi[..., 0].shape=(1, 3, 80, 80), tobj的shape为(1, 3, 80, 80)值为0

            n = b.shape[0]  # number of targets
            if n:
                ps = pi[b, a, gj, gi]  # prediction subset corresponding to targets, shape为(MP, num class+5),MP为匹配到GT框的正样本

                # Regression
                grid = torch.stack([gi, gj], dim=1)
                pxy = ps[:, :2].sigmoid() * 2. - 0.5
                #pxy = ps[:, :2].sigmoid() * 3. - 1.
                pwh = (ps[:, 2:4].sigmoid() * 2) ** 2 * anchors[i]
                pbox = torch.cat((pxy, pwh), 1)  # predicted box
                selected_tbox = targets[i][:, 2:6] * pre_gen_gains[i] # targets[i].shape=(MP, 6),MP为匹配到GT框的正样本
                selected_tbox[:, :2] -= grid
                iou = bbox_iou(pbox.T, selected_tbox, x1y1x2y2=False, CIoU=True)  # iou(prediction, target)
                lbox += (1.0 - iou).mean()  # iou loss

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

                # Classification
                selected_tcls = targets[i][:, 1].long() # targets[i].shape=(MP, 6),MP为匹配到GT框的正样本
                if self.nc > 1:  # cls loss (only if multiple classes)
                    t = torch.full_like(ps[:, 5:], self.cn, device=device)  # targets, ps[:, 5:].shape=(MP, num class+5), self.cn=0, t.shape=(MP, num class+5)
                    t[range(n), selected_tcls] = self.cp # self.cp=1, t.shape=(MP, num class+5), t即为smooth_BCE的权重
                    lcls += self.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)]

            obji = self.BCEob