Peterisfar/YOLOV3

Issues

• 训练问题

  #53 opened 2 years ago by mawei-north
  1

• 换自己的数据集为什么跑一点就报错了

  #48 opened 3 years ago by mx985211
  2

• loss 降不下来

  #52 opened 2 years ago by BaseJiang
  0

• Anchors setting when training on the custom dataset

  #43 opened 4 years ago by ShaofengZou
  1

• 聚类anchors

  #47 opened 4 years ago by gll-sketch
  0

• Is it possible to convert weight from pt to darknet

  #46 opened 4 years ago by mhasnat
  0

• index 56 is out of bound for axis 1 with size 56?

  #32 opened 4 years ago
  1

• The trick about `Label Smoothing` in `Bag of Freebies for Training Object Detection Neural Networks`

  #45 opened 4 years ago by crownz-sec
  1

• After validating for once,the programe can not train any more.

  #28 opened 5 years ago by momo666666
  9

• validate的速度特别慢

  #44 opened 4 years ago by crownz-sec
  4

• multi_scale在单个epoch里面并不起作用？

  #42 opened 4 years ago by crownz-sec
  9

• 作者您好，请问yolo_loss中，怎么计算的loss confidence的呀？

  #38 opened 4 years ago by ZhongyuanW
  1

• 每个epoch训练完之后validate的速度特别慢？

  #19 opened 5 years ago by Shank2358
  10

• How many epochs have you trained the best.pt file?

  #41 opened 4 years ago by crownz-sec
  1

• What's the difference between darknet53-448.weights and darknet53.conv.74?

  #40 opened 4 years ago by crownz-sec
  2

• VOC训练问题

  #36 opened 4 years ago by Yzichen
  1

• mixup have some problems when pics‘s shape different

  #39 opened 4 years ago by iwanggp
  0

• Change backbone

  #35 opened 4 years ago by andyJJz
  4

• 您好，想请问一个您的label处理的问题

  #37 opened 4 years ago by bigKoki
  1

• 【Question】Batch size?

  #34 opened 4 years ago by neesetifa
  2

• Anchor

  #33 opened 4 years ago by whyisma
  1

• About training

  #27 opened 4 years ago by yehaizi1995
  1

• About utils/datasets.py

  #30 opened 4 years ago by GesilaA
  1

• Detect too many extra bounding boxes with the test threshold of 0.01.

  #29 opened 4 years ago by wangren97
  2

• IDE

  #31 opened 4 years ago by ThanhPham1987
  1

• How can the fps be when input image is 416*416?

  #24 opened 5 years ago by BucolicWind
  1

• 感谢您的代码。关于训练loss变为NAN

  #26 opened 5 years ago by ww-zwj
  4

• 作者您好，可否告知一下训练至80%mAP时的loss?

  #25 opened 5 years ago by ww-zwj
  1

• Why do you set epoch >= 20 in train.py ?

  #23 opened 5 years ago by XiangqianMa
  0

• About training

  #9 opened 5 years ago by yehaizi1995
  4

• IndexError: Caught IndexError in DataLoader worker process 1.

  #22 opened 5 years ago by jiangzeshuo
  2

• Dear author, why do you choose alpha =1.0 in the focal loss?

  #21 opened 5 years ago by jumoran
  1

• some questions about the 'datasets' and 'yolo_loss' code

  #20 opened 5 years ago by wangren97
  2

• 作者您好：感谢您的付出，我有个问题看了一下您的loss函数在计算conf 和cls loss时用的是p 而不是p_d。这一块是不是代码写错了？

  #18 opened 5 years ago by jumoran
  1

• where is the params.py?please

  #17 opened 5 years ago by BucolicWind
  1

• 关于voc.py中的逻辑错误

  #14 opened 5 years ago by yuanliangxie
  19

• Can you share the AP of each class in VOC for the baseline version?

  #16 opened 5 years ago by AlexofNTHU
  1

• 关于模型的map的疑问？

  #15 opened 5 years ago by yuanliangxie
  27

• baseline

  #1 opened 5 years ago by yyx1107
  1

• Dear author, thank you for your code. How many epoches have you trained for pascal voc to get map 85%? Is the batches set by 8?

  #13 opened 5 years ago by jumoran
  4

• augment

  #12 opened 5 years ago by gjfzjj1995
  4

• GPU training

  #11 opened 5 years ago by MichaelCong
  1

• 关于yololoss的疑问

  #10 opened 5 years ago by yuanliangxie
  1

• 关于测试

  #8 opened 5 years ago by yehaizi1995
  2

• 关于训练

  #7 opened 5 years ago by yehaizi1995
  1

• Training on custom dataset guide and Image, video detection demo

  #6 opened 5 years ago by smalldroid
  1

• 关于结构

  #3 opened 5 years ago by yehaizi1995
  4

• LOSS

  #4 opened 5 years ago by Gaondong
  3

• best.pt链接

  #2 opened 5 years ago by tanluren
  2