Yolov7一代版本与二代版本改LOSS注意点_yolov7改损失函数-优快云博客

本文链接：https://blog.youkuaiyun.com/DylanZS1/article/details/130986107

YOLOv7的损失计算在一代和二代中有不同，主要在于是否使用compute_loss_ota。在hyp文件中，一代会根据loss_ota参数选择损失函数，而二代默认使用compute_loss_ota。在代码中，根据CUDA环境和特定优化选项如quad来调整损失并平均梯度。

摘要生成于 C知道，由 DeepSeek-R1 满血版支持，前往体验 >

YOLOv7的utils文件夹下的loss文件中给了两种损失函数：computeloss和computelossota，默认都使用computelossota。

YOLOv7一代计算LOSS：

 # Forward
            with amp.autocast(enabled=cuda):
                pred = model(imgs)  # forward
                if 'loss_ota' not in hyp or hyp['loss_ota'] == 1:
                    loss, loss_items = compute_loss_ota(pred, targets.to(device), imgs)  # loss scaled by batch_size
                else:
                    loss, loss_items = compute_loss(pred, targets.to(device))  # loss scaled by batch_size
                if rank != -1:
                    loss *= opt.world_size  # gradient averaged between devices in DDP mode
                if opt.quad:
                    loss *= 4.

YOLOv7二代计算LOSS：

   # Forward
            with amp.autocast(enabled=cuda):
                pred = model(imgs)  # forward
                loss, loss_items = compute_loss_ota(pred, targets.to(device), imgs)  # loss scaled by batch_size
                if rank != -1:
                    loss *= opt.world_size  # gradient averaged between devices in DDP mode
                if opt.quad:
                    loss *= 4.

一代版本是根据hyp文件中最后一行，判断使用何种loss，例如默认的hyp.scratch.p5.yaml:

lr0: 0.01  # initial learning rate (SGD=1E-2, Adam=1E-3)
lrf: 0.1  # final OneCycleLR learning rate (lr0 * lrf)
momentum: 0.937  # SGD momentum/Adam beta1
weight_decay: 0.0005  # optimizer weight decay 5e-4
warmup_epochs: 3.0  # warmup epochs (fractions ok)
warmup_momentum: 0.8  # warmup initial momentum
warmup_bias_lr: 0.1  # warmup initial bias lr
box: 0.05  # box loss gain
cls: 0.3  # cls loss gain
cls_pw: 1.0  # cls BCELoss positive_weight
obj: 0.7  # obj loss gain (scale with pixels)
obj_pw: 1.0  # obj BCELoss positive_weight
iou_t: 0.20  # IoU training threshold
anchor_t: 4.0  # anchor-multiple threshold
# anchors: 3  # anchors per output layer (0 to ignore)
fl_gamma: 0.0  # focal loss gamma (efficientDet default gamma=1.5)
hsv_h: 0.015  # image HSV-Hue augmentation (fraction)
hsv_s: 0.7  # image HSV-Saturation augmentation (fraction)
hsv_v: 0.4  # image HSV-Value augmentation (fraction)
degrees: 0.0  # image rotation (+/- deg)
translate: 0.2  # image translation (+/- fraction)
scale: 0.9  # image scale (+/- gain)
shear: 0.0  # image shear (+/- deg)
perspective: 0.0  # image perspective (+/- fraction), range 0-0.001
flipud: 0.0  # image flip up-down (probability)
fliplr: 0.5  # image flip left-right (probability)
mosaic: 1.0  # image mosaic (probability)
mixup: 0.15  # image mixup (probability)
copy_paste: 0.0  # image copy paste (probability)
paste_in: 0.15  # image copy paste (probability), use 0 for faster training
loss_ota: 1 # use ComputeLossOTA, use 0 for faster training

loss_ota: 1 # use ComputeLossOTA, use 0 for faster training，其中的1代表使用ComputeLOSSOTA

二代中hyp文件删掉了最后一行，改为默认使用 computelossota

因此在两代版本改loss函数会有不同，注意改的位置要和使用的loss函数对应，例如使用默认computelossota计算，就要在class computelossota中进行修改