george
george
I think you should compare it with yolox-L. In addition, it seems that your yolox/v6 has not been completely implementated.
> > > > you are right, in my YOLOX, the depth ratio also control the number of decoupled conv in the head part, I just want to check if...
> 请问您的yolox相比torch原版的yolox有比较具体的改进吗，因为训练效果比原版的要好的多。谢谢回复 请问你训的COCO还是自定义数据集？具体指标怎么样，比原版好多少？
@bobo0810 学习率改成0.001，lr_decay改成0.95.跑100个epoch ，测试集每张图概率仍然都是0.5左右啊？是过拟合了？还是梯度消失？
@bobo0810 谢谢，你的是 max_epoch = 100吧？。。我是阿里云主机训练过程还没可视化，你这是之前跑成功的吧，最近不知道作者有改动什么地方没。我参数基本就是按照作者的默认设置的，lr_decay =0.95是跟你一样的，作者默认lr_decay=0.5也试过了，但都是训练后在测试集上表现不太正常啊，每张图测试都是0.49左右，跟随机猜测没区别，不知道为什么？
@adi1999 Just use the data format of this repo, the MOT results of MOT16-test in this repo can be submitted to the https://motchallenge.net , got MOTA 64.4 (1088x608). But for...
@Zhongdao same question 73.1 on MOTA instead of 74.8
mask_rcnn_r4_ct_c3-c5_r50_sbn_fpn_1x.py -- | Context | box AP | mask AP -- | -- | -- | -- BN | - | 37.3 | 34.2 BN | GC(c3-c5, r4) | 37.1...
@RainHxj normalize should be added after line "style='pytorch'"
Also need, [email protected], thanks !!!