【GIOU】目标检测中的Loss改进:GIOU

{"train_lr": 9.999999999999769e-06, "train_loss": 21.28127575720381, "train_loss_vfl": 0.7926896240101421, "train_loss_bbox": 0.1730125365169637, "train_loss_giou": 0.7405169542792958, "train_loss_vfl_aux_0": 1.0139826408722117, "train_loss_bbox_aux_0": 0.19708112190893423, "train_loss_giou_aux_0": 0.7683152278389321, "train_loss_vfl_aux_1": 0.9705591896857355, "train_loss_bbox_aux_1": 0.18077695704762292, "train_loss_giou_aux_1": 0.7398171850033026, "train_loss_vfl_aux_2": 0.9128599568607922, "train_loss_bbox_aux_2": 0.17436226969291954, "train_loss_giou_aux_2": 0.7381583543755973, "train_loss_vfl_aux_3": 0.8479711368086453, "train_loss_bbox_aux_3": 0.1736977368308157, "train_loss_giou_aux_3": 0.737467385875354, "train_loss_vfl_aux_4": 0.8121209355457774, "train_loss_bbox_aux_4": 0.1722905259302171, "train_loss_giou_aux_4": 0.7385181724025518, "train_loss_vfl_aux_5": 0.9396382454837393, "train_loss_bbox_aux_5": 0.26576847316740193, "train_loss_giou_aux_5": 0.9272661880127289, "train_loss_vfl_dn_0": 0.4974277570874936, "train_loss_bbox_dn_0": 0.3083435350119193, "train_loss_giou_dn_0": 0.788808505865013, "train_loss_vfl_dn_1": 0.4677735700797235, "train_loss_bbox_dn_1": 0.2513014289161944, "train_loss_giou_dn_1": 0.6824771829995969, "train_loss_vfl_dn_2": 0.4499487520451316, "train_loss_bbox_dn_2": 0.23452719723692542, "train_loss_giou_dn_2": 0.6482247728346778, "train_loss_vfl_dn_3": 0.44208292053930415, "train_loss_bbox_dn_3": 0.22947435307295005, "train_loss_giou_dn_3": 0.6422228983738137, "train_loss_vfl_dn_4": 0.44009303934169264, "train_loss_bbox_dn_4": 0.2277972794350713, "train_loss_giou_dn_4": 0.6420102769536292, "train_loss_vfl_dn_5": 0.4405380960717891, "train_loss_bbox_dn_5": 0.22784265998235476, "train_loss_giou_dn_5": 0.64351078227506, "test_coco_eval_bbox": [0.2850231398773143, 0.5405510341454601, 0.23959214836924067, 0.16947794653844553, 0.21123704386978878, 0.2765617147926568, 0.35497390251294986, 0.4976055955128448, 0.5902341811264838, 0.3709281425679972, 0.4686535422726888, 0.665727636061482], "epoch": 63, "n_parameters": 62820328}这些指标都是社么意思,哪些指标是有用的,这些指标的变化能显示出什么
03-24
评论
添加红包

请填写红包祝福语或标题

红包个数最小为10个

红包金额最低5元

当前余额3.43前往充值 >
需支付:10.00
成就一亿技术人!
领取后你会自动成为博主和红包主的粉丝 规则
hope_wisdom
发出的红包
实付
使用余额支付
点击重新获取
扫码支付
钱包余额 0

抵扣说明:

1.余额是钱包充值的虚拟货币,按照1:1的比例进行支付金额的抵扣。
2.余额无法直接购买下载,可以购买VIP、付费专栏及课程。

余额充值