-
Snippet from my training outputs
```bash
v3 (iou loss, Normalizer: (iou: 0.07, cls: 1.00) Region 139 Avg (IOU: 0.221490, GIOU: 0.156080), Class: 0.459105, Obj: 0.000986, No Obj: 0.002985, .5R: 0.125…
-
Dear zunzhumu,
Hi!
I have this problem, why the loss is zero
INFO:root:Train Epoch: 160 [0/6 (0%)] cls_Loss: 0.000000 shape_loss:0.000000 offset_loss:0.000000 giou_loss:0.0000…
-
Thank you very much for sharing!
Is the loss function you use in the network GIoU or CIoU?
-
您好,我在运行您的代码的时候,出现了gpu_mem在每一次Epoch后都上涨的问题,最终导致显存爆了,请问这个是什么原因?
Epoch gpu_mem GIoU obj cls total targets img_size
0/99 1.84G 0.07251 0.03948 0.04516 …
-
Hello, can you share the training log? I want to confirm if the ratio of focal loss and giou loss is appropriate.
-
保存时giou=1.4,map=0.47,再训练一个epoch变2,map变0.364
-
I got this result :
v3 (mse loss, Normalizer: (iou: 0.750000, cls: 1.000000) Region 130 Avg (IOU: -nan, GIOU: -nan), Class: -nan, Obj: -nan, No Obj: 0.504471, .5R: -nan, .75R: -nan, count: 0, loss = …
-
C:\Users\bxf\anaconda3\envs\transt\python.exe C:/PyCharmProjects/TransT-main/ltr/run_training.py
Training: transt transt
WARNING: You are using tensorboardX instead sis you have a too old pytorch …
-
v3 (iou loss, Normalizer: (iou: 0.07, cls: 1.00) Region 161 Avg (IOU: 0.270259, GIOU: 0.165639), Class: 0.359618, Obj: 0.541298, No Obj: 0.522445, .5R: 0.000000, .75R: 0.000000, count: 4, class_loss =…
-
Why do you use loss (include GIOU_loss, conf_loss, prob_loss ) to mutiply label_mixw? From the paper, label_mixw is just responsible for the prob_loss.