bubbliiiing/yolov4-keras

Issues

• 导，你的heatmap生成原理是什么？

  #99 opened a year ago by lty1234-deep
  0

• 版本问题

  #98 opened a year ago by Iamlaoli
  1

• 请问模型训练完成怎么看IoU和Average time的值

  #96 opened 2 years ago by kola515
  7

• 请问map里面detection-results 文件夹内容是什么意思？

  #97 opened 2 years ago by kola515
  0

• 训练自己的数据集shape不匹配问题

  #95 opened 2 years ago by Seibelgys
  1

• 训练自己的数据集shape不匹配问题

  #94 opened 2 years ago by Seibelgys
  2

• 标签名称可以是中文吗？

  #90 opened 2 years ago by smallfang2008
  2

• 如果是自定义数据集训练模型，需要重新生成先验框还是使用默认的？

  #89 opened 2 years ago by smallfang2008
  1

• 自定义训练数据集图片大小有要求吗？

  #88 opened 2 years ago by smallfang2008
  1

• TP FP問題

  #84 opened 3 years ago by bro278911
  9

• detect.py的問題

  #86 opened 3 years ago by larry3425527
  2

• 【同一目标出现多个标签】

  #81 opened 3 years ago by chizao89757
  2

• 使用预训练模型出现问题

  #61 opened 4 years ago by beathunder
  5

• val_loss在有些epoch突变特别大

  #76 opened 3 years ago by PuckBlink
  1

• 这个代码有问题，训练自己的数据集检测不到，不如yolov3

  #75 opened 3 years ago by luogantt
  1

• 请问一下，loss在第一个epoch的时候出现nan

  #56 opened 3 years ago by mengmeng0320
  8

• 你好我想请教下， 如何将.h5的权重换成.onnx的呢

  #74 opened 3 years ago by tzl1
  1

• 使用预训练权重训练voc模型，损失减到8，但是测试时检测不到物体，请问对预测图片的处理是否跟训练是一致？

  #72 opened 3 years ago by diegolitianfu
  5

• 在使用predict.py時報錯

  #69 opened 3 years ago by hoho84310
  2

• GPU训练问题

  #68 opened 3 years ago by RyanCCC
  1

• 为什么yolov4的map还没有SSD的高啊，相同的数据集

  #53 opened 4 years ago by Cui7
  4

• 边框回归问题

  #66 opened 4 years ago by YEYEAKEI
  0

• 关于维度不匹配的问题

  #58 opened 4 years ago by raylzhou
  4

• 请问train.py里的classes_path变量在哪呀？没有搜索到这段代码

  #65 opened 4 years ago by Aegsteh
  0

• 循环读取文件夹图片

  #64 opened 4 years ago by Lun-Wang
  1

• 修改model结构后重新训练报错

  #63 opened 4 years ago by chunxuli90
  1

• 请问不能用COCO格式训练吗？

  #62 opened 4 years ago by dexter2406
  1

• 单个检测与detection-result结果不一样

  #59 opened 4 years ago by geyu1998
  1

• 为什么yolov4的检测效果不如yolov3

  #60 opened 4 years ago by youquderen97
  1

• 训练结束后识别不到物体

  #57 opened 4 years ago by Aurora-LiZhi
  27

• voc07_test mAP

  #55 opened 4 years ago by MiniLisaBear
  5

• 请问你这里面用上了Dropblock了吗

  #54 opened 4 years ago by Cui7
  1

• 请教一下学习数据大小的问题

  #50 opened 4 years ago by 43998213
  5

• 有没有提高GPU利用率的方法呢

  #51 opened 4 years ago by moonsonK
  2

• 如果从头开始训练，不用预训练模型model_path怎么改呢？

  #47 opened 4 years ago by tommyMessi
  3

• 请教一下，跑原项目的时候video.py y4只识别出2个目标，y3识别出5个目标

  #49 opened 4 years ago by 43998213
  1

• 可否转为TensorRT

  #48 opened 4 years ago by simplew2011
  1

• PR Curve 問題

  #46 opened 4 years ago by green27698
  1

• FPS 問題

  #45 opened 4 years ago by green27698
  3

• epoch问题

  #44 opened 4 years ago by Xu-Fengnian
  1

• 内存不足问题

  #43 opened 4 years ago by Xuyinsong
  0

• 您好，怎么样计算模型的flops值呢。

  #42 opened 4 years ago by kz258852
  2

• is there any way to reduce the size of the model?

  #41 opened 4 years ago by RoozbehSanaei
  0

• 请问我可不可以参考一下您把.weights文件转化为.h5文件的代码convert.py

  #40 opened 4 years ago by Marc-lu
  1

• 训练自己的数据问题

  #39 opened 4 years ago by yanlin1995
  1

• FPS 和yolo3

  #38 opened 4 years ago by dexter2406
  1

• 训练时loss值为nan

  #36 opened 4 years ago by myleosu
  4

• 检测速度与AB大神的有点差异

  #37 opened 4 years ago by luliang-eng
  1

• DropBlock Issue

  #35 opened 4 years ago by rongrong9005
  5

• .h5转saved在tf serving部署问题

  #34 opened 4 years ago by chenhui0518
  2