coldlarry/YOLOv3-complete-pruning

Issues

• 怎么观察稀疏化是否到位？

  #23 opened 5 years ago by mozpp
  3

• prune_tiny_yolo.py错误

  #24 opened 5 years ago by Mayzy
  4

• 使用官方的yolov3.weights进行图片测试，什么框都没画出来。

  #43 opened 5 years ago by monoloxo
  10

• 训练作者的数据集，mAP是0？请问什么原因？

  #15 opened 5 years ago by 121786404
  5

• 保存weights在detect.py中有结果，但是用原版的darknet中没有结果

  #42 opened 5 years ago by ZHEQIUSHUI
  5

• 模型可以保存为.weights格式吗

  #37 opened 5 years ago by lilei1128
  14

• 量化后的文件保存格式、路径及test问题

  #56 opened 5 years ago by Blue-Eagle-10
  1

• Pruning problem

  #31 opened 5 years ago by ZhaoMonica
  3

• one question

  #52 opened 5 years ago by zhaojc001
  2

• 遇到维度不匹配问题，希望作者帮忙解决

  #62 opened 5 years ago by yuanlei-2018
  1

• 模型剪枝报错了，请问下什么原因

  #32 opened 5 years ago by baidu88vip
  2

• yolo-tiny剪枝错误

  #28 opened 5 years ago by Whatsetsthisend
  5

• 需要注意的是，这里需要在.py文件内，将opt内的cfg和weights变量指向第2步稀疏化后生成的cfg文件和weights文件。 此外，可通过增大代码中percent的值来获得更大的压缩率。（若稀疏化不到位，且percent值过大，程序会报错。）

  #63 opened 5 years ago by duducode
  1

• cfg文件问题

  #38 opened 5 years ago by zhaodongxi
  1

• win10多显卡

  #61 opened 5 years ago by AadeIT
  4

• 使用训练的.pt文件遇到tensor不匹配的问题

  #64 opened 5 years ago by suojiashun
  0

• curl: (35) OpenSSL SSL_connect: 连接被对方重设 in connection to drive.google.com:443

  #60 opened 5 years ago by RosabanksiaeAit
  2

• examples.ipynb中数据集问题

  #59 opened 5 years ago by RosabanksiaeAit
  0

• 剪枝出错

  #57 opened 5 years ago by lyc6749
  0

• prune tiny size and percent

  #58 opened 5 years ago by PiseyYou
  0

• 训练开始时遇到Segmentation fault (core dumped)

  #54 opened 5 years ago by daiwc
  0

• 关于稀疏因子s

  #53 opened 5 years ago by huoguangdiandian
  0

• 运行代码时出现问题

  #51 opened 5 years ago by AadeIT
  1

• 微调时出现的错误

  #50 opened 5 years ago by AadeIT
  0

• 正常模型剪枝是否需要微调？

  #49 opened 5 years ago by Chase2816
  0

• 量化论文

  #48 opened 5 years ago by JensenHJS
  0

• 量化速度

  #47 opened 5 years ago by lxtong01
  0

• 不同训练epoch后before的mAP值不同

  #46 opened 5 years ago by ZhouwenfengTyrantasteroid
  0

• 大家好，我自己开了个分支，欢迎讨论

  #45 opened 5 years ago by mozpp
  2

• 剪枝时bn层的bias不用mask置零吗？

  #44 opened 5 years ago by zihaozhang9
  0

• 无法使用GPU训练

  #41 opened 5 years ago by lehuiwu
  1

• 如何使用gpu训练

  #30 opened 5 years ago by Damon2019
  1

• 训练自己的数据集报错

  #40 opened 5 years ago by lehuiwu
  0

• non-finite loss

  #39 opened 5 years ago by XinCode
  0

• 剪枝问题

  #36 opened 5 years ago by zhaodongxi
  0

• 我使用和您相同的数据集以及参数训练，channel prun结果相仿，layer prun后的map微调前只有20，可能是什么问题呢？？

  #35 opened 5 years ago by px2017
  2

• 训练自己数据集到50epoch，mAP突然变为0,

  #14 opened 5 years ago by noobliang
  3

• shortcut prune bug

  #34 opened 5 years ago by Gaondong
  0

• 剪枝后map=0

  #21 opened 5 years ago by yangdeqieyingli
  5

• 稀疏率和剪枝率对map影响

  #33 opened 5 years ago by huoguangdiandian
  0

• 数据集下载

  #29 opened 5 years ago by Damon2019
  0

• 训练自己的数据集，减枝后Map 77.4 run detect.py 结果却很差，这是为什么啊？

  #19 opened 5 years ago by 121786404
  1

• yolov3网络置换

  #27 opened 5 years ago by huoguangdiandian
  1

• 恢复训练

  #26 opened 5 years ago by zhanghongruiupup
  0

• 关于稀疏化的问题

  #25 opened 5 years ago by zouxinhua
  0

• 训练参数问题

  #22 opened 5 years ago by mozpp
  0

• 作者您好，感谢您能开源，我想问下如果我更换backbone的话也可以做剪枝吗？

  #20 opened 5 years ago by lukaisheng1203
  2

• detect

  #18 opened 5 years ago by 121786404
  0

• 第一步正常训练属于必须嘛？剪枝的话只需要第二、三步吧？

  #17 opened 5 years ago by xncaffe
  1

• model

  #16 opened 5 years ago by xncaffe
  0