Stinky-Tofu/Stronger-yolo

Issues

• 请教下为什么pytorch version版本的mAP只有76.5，你这里可以达到88？

  #110 opened 5 years ago by clw5180
  0

• error when rename.py is run

  #107 opened 5 years ago by jwnirvana
  0

• YOLO-Lite哪里用到KL Loss

  #76 opened 5 years ago by gfjiangly
  8

• 中断训练过程后再继续训练

  #83 opened 5 years ago by nobody-cheng
  3

• ModuleNotFoundError: No module named 'dataAug'

  #103 opened 5 years ago by kingmacth
  1

• error exits when rename.py is run

  #108 opened 5 years ago by jwnirvana
  0

• type error

  #106 opened 5 years ago by jwnirvana
  1

• 是否加载预训练模型的精度差异

  #105 opened 5 years ago by litingsjj
  3

• VOC 过拟合问题

  #104 opened 5 years ago by CaoWGG
  3

• 关于focal loss 中alpha=1是相当于没有对正负样本数量进行加权是吗？

  #79 opened 5 years ago by 123lifei
  2

• self.__steps_per_period 错误， numpy.float64' object cannot be interpreted as an integer

  #58 opened 5 years ago by tianyexing
  3

• V3/NameError: name 'file' is not defined

  #97 opened 5 years ago by MichaelCong
  2

• v3 在gtx1080ti训练了多久?

  #96 opened 5 years ago by qiqzhang
  7

• 关于 数据读取 的问题

  #100 opened 5 years ago by ghoshaw
  2

• 方便提供yolov3和yolov3 lite的预训练权重吗？

  #99 opened 5 years ago by sijun-zhou
  1

• COCO训练需要多久呢?

  #98 opened 5 years ago by qiqzhang
  2

• readme上的params参数的大小

  #92 opened 5 years ago by nobody-cheng
  1

• 关于giou

  #95 opened 5 years ago by 791136190
  0

• data.py文件处理anchor的wh缩放问题

  #94 opened 5 years ago by nobody-cheng
  0

• About GIOU loss

  #93 opened 5 years ago by zhangyufei1995
  1

• 关于GIoU loss的疑问

  #72 opened 5 years ago by lyx190
  8

• 您好，请问你是如何使用mixup的，有增加epoch么？

  #91 opened 5 years ago by 711e
  0

• v1版本训练梯度爆炸

  #80 opened 5 years ago by nobody-cheng
  2

• 多尺度训练问题

  #85 opened 5 years ago by zhangyufei1995
  2

• 预训练权重

  #89 opened 5 years ago by zhangyufei1995
  1

• GPU利用率好低，只有百分之几 请问是为啥呢

  #90 opened 5 years ago by zhangyufei1995
  1

• An error occurred when I ran test.py

  #65 opened 5 years ago by caoabc
  2

• v1的有关问题请教

  #88 opened 5 years ago by zhangyufei1995
  6

• v3

  #87 opened 5 years ago by zhangyufei1995
  4

• 您好， 我想问下您的cuda是什么版本的啊，我这个一直报错

  #86 opened 5 years ago by lukaisheng1203
  4

• 有试过加入 SE 模块吗？

  #62 opened 5 years ago by Lam1360
  5

• mobile-yolo模型图片预处理

  #82 opened 5 years ago by lightfate
  3

• 生成了CKPT 文件怎么导出PB文件啊 ？

  #67 opened 5 years ago by basser2
  2

• v1 版本梯度爆是什么原因啊

  #81 opened 5 years ago by 123lifei
  8

• cosine learning rate decay

  #78 opened 5 years ago by Sun-Fan
  0

• 关于样本图片标注框转换成预测数据的问题

  #77 opened 5 years ago by 123lifei
  2

• 你好，请问加载三个权重文件的时候报错怎么解决

  #71 opened 5 years ago by noobgrow
  1

• failed to run cuBLAS in RTX2080

  #75 opened 5 years ago by HochCC
  2

• test with mixup

  #73 opened 5 years ago by pengfeidip
  2

• 如何使用focal loss 、soft nms等这些改进呢？

  #74 opened 5 years ago by zychen2016
  4

• 关于训练的疑问

  #68 opened 5 years ago by annaloua
  0

• Why not use weight decay or other regularization methods

  #69 opened 5 years ago by chenbohua3
  1

• .

  #70 opened 5 years ago by gzz1529657064
  0

• data_aug.py第41 42行max改为min是否更合适些？

  #63 opened 5 years ago by fan0210
  1

• About GIOU code

  #64 opened 5 years ago by Apostatee
  1

• missed anchor parameters?

  #61 opened 5 years ago by LightToYang
  2

• about train log

  #66 opened 5 years ago by 791136190
  2

• darknet cfg format

  #57 opened 5 years ago by DrVijayK
  0

• 关于标签的疑问

  #59 opened 5 years ago by StrongBirds
  3

• About v3 anchor-free methods

  #60 opened 5 years ago by Apostatee
  2