Lam1360 YOLOv3-model-pruning issues

Lam1360 / YOLOv3-model-pruning

在 oxford hand 数据集上对 YOLOv3 做模型剪枝（network slimming）

MIT License

1.67k stars 431 forks source link

issues

Newest

Newest Most commented Recently updated Oldest Least commented Least recently updated

Calling the following command after pruning fails to complete the test and reports an error

#105 tricoffee opened 2 months ago
0
请问如果训练多个类别，该如何配置？看标签文件没有写类别，是只能训练一类吗？

#104 WestbrookZero opened 1 year ago
0
Update LICENSE

#103 victor30608 opened 1 year ago
0
推理速度自测是22ms左右，不是7ms

#102 steady-pace opened 2 years ago
0
能用于别的模型么？比如bisenet

#101 WEIZHIHONG720 opened 2 years ago
0
精度过低

#100 TZWX-0 opened 2 years ago
0
请问大家检测速度的代码在哪里啊？没有看到

#99 906096481 opened 2 years ago
0
为什么剪枝后的 sample_metrics 为空

#98 1-snail opened 2 years ago
2
为什么会出现No such file or directory: 'data\\images\\train\\Buffy_1.jpg'

#97 1-snail closed 3 years ago
0
为什么跑的这个数据集的结果只有map0.68，作者写的0.75

#96 Xingxu1996 opened 3 years ago
1
为什么上一层丢掉的权重，下一层的BN层runing_meaning要减去offset

#95 cqray1990 opened 3 years ago
0
训练时候cls_acc一直是100%

#94 Pong-97 opened 3 years ago
1
这个模型适用于其他crnn结构嘛？

#93 GUOhm230 opened 3 years ago
0
模型裁剪以后重新训练可以生成新的pth, 但是同时生成.weights文件报错。

#92 alisonzhang88 opened 3 years ago
0
Cooperation Proposal: Lam1360/YOLOv3-model-pruning & PaddlePaddle

#91 saxon-zh opened 3 years ago
0
prune_model_keep_size 只是将model中冗余的通道通过mask置零，并没有真正的剪掉这些通道？

#90 tangchen2 closed 3 years ago
1
减枝完，检测不到结果

#89 ReOneK opened 4 years ago
0
求FLOPS代码位置

#88 chenaifang opened 4 years ago
1
"Channels would be all pruned!"

#87 liuyuxun12 opened 4 years ago
0
请问这里用于稀疏化训练的预训练权重可以是之前训练好的模型吗？

#86 DongLZY opened 4 years ago
1
train.py训练完为啥没有选出一个最优的yolov3_ckpt.pth额，只是生成了一系列不同epoch的pth

#85 zeta7337 opened 4 years ago
0
剪枝后FPS速度问题

#84 pangkun248 opened 4 years ago
2
执行test_prune.py出错,为了节省时间，我只用了100个数据进行测试，结果出问题

#83 zhousaYolo opened 4 years ago
3
conf_thres=0.01 关于阈值设定问题，为什么会这么小

#82 Ming-Y-ANG opened 4 years ago
0
prune_model_keep_size()这个函数的主要作用是？

#81 ThomasVon2021 opened 4 years ago
0
不理解添加L1规范怎样使得BN层的某些gama值变为0的？

#80 ZhangAdjy opened 4 years ago
5
filters的设置问题

#79 blackantywj opened 4 years ago
0
剪枝后运行模型出错

#78 dengfenglai321 closed 4 years ago
1
Unable to download

#77 7thstorm opened 4 years ago
0
请问一下，如果加入负样本，该如何训练

#76 lehuiwu opened 4 years ago
0
保存的.weights模型为啥ｄａｒｋｎｅｔ调用无法检测

#75 lilei1128 opened 4 years ago
5
为什么每一帧的检测速度都不一样

#74 zhysee opened 4 years ago
0
怎么提高精度，没有达到你说的０．７７那么高啊

#73 lilei1128 opened 4 years ago
0
mAP很低

#72 jenn0727 opened 4 years ago
1
稀疏化训练bbox问题

#71 sunserforever opened 4 years ago
3
ap为0是为什么。没有使用预训练模型。

#70 121786404 opened 4 years ago
0
train error

#69 zhanghongruiupup closed 4 years ago
0
多尺度训练问题

#68 shuncyu opened 4 years ago
1
用 yolo3.weights 运行test_prune 文件，模型加载出错，pickle wrong

#67 muyouyang opened 4 years ago
0
自己设计的网络可以稀疏训练吗？

#66 May-forever opened 4 years ago
4
convert.py代码里为何把train和valid的数据合并放在train里了？

#65 maosiang opened 4 years ago
0
The usage of gpu memory is too high!!!

#64 yxy2529023446 opened 4 years ago
0
使用 yolo3.weights 训练时报错

#63 gooseable closed 4 years ago
1
未剪枝的模型权重参数有吗

#62 maosiang closed 4 years ago
0
为啥我测试之后发现模型前向推断速度并没有加快呢？

#61 texasmotorcycle opened 4 years ago
7
剪枝后的时间变化没有作者那么明显，是什么原因

#60 texasmotorcycle opened 4 years ago
2
关于prune_utils.py中prune_model_keep_size()函数的理解

#59 coldlarry opened 4 years ago
5
段错误

#58 ZhangJiangYong-ccc closed 4 years ago
0
yolov3.weights是在什么数据集上预训练的权重

#57 alanMachineLeraning opened 4 years ago
0
我注意到shortcut部分没有进行剪枝?

#56 alanMachineLeraning opened 4 years ago
0