vgg16-397923af.pth download

[shangjianan2]

链接：https://pan.baidu.com/s/1DFxRkVBRjbDBDKKXdND47A 密码：u0xf

[shangjianan2]

Sometimes we can not download vgg16-397923af.pth. Here is vgg16-397923af.pth I have already download. I hope it will useful to you.

[chenyuntc]

Thank you! I'll update it in README and close the issue here.

[Stephenfang51]

Dear Sir

Thanks for the Baidu Pan.

However, the "convert_caffe_pretrain.py" can't convert the "vgg16-397923af" into torch version. I tried and it failed.

The network has different structure between vgg16-397923a and vgg16-00b39a1b I guess

I suggest people can download the vgg16-00b39a1b.pth on Google Colab and covert to torch version using "convert_caffe_pretrain.py" as Chenyun said in README.

Thanks !

[greenyin]

Dear Sir

Thanks for the Baidu Pan.

However, the "convert_caffe_pretrain.py" can't convert the "vgg16-397923af" into torch version. I tried and it failed.

The network has different structure between vgg16-397923a and vgg16-00b39a1b I guess

I suggest people can download the vgg16-00b39a1b.pth on Google Colab and covert to torch version using "convert_caffe_pretrain.py" as Chenyun said in README.

Thanks !

I also encountered the same problem, did you solve it?

[shida666]

Hey , everyone ! The file “vgg16-397923af.pth“ we downloaded is a pre-trained model pretrained in Pytroch environment ! it doesn't be pretrained in the Caffe ! So after we download the vgg16-397923af.pth file into "checkpoints" folder , and we dont need to make the "python misc/convert_caffe_pretrain.py"in our shell , and we should directly to the "python train.py train --env='fasterrcnn-caffe' --plot-every=100 --caffe-pretrain“ ！ Don't forget change the name of this file! via"mv vgg16-397923af.pth vgg16_caffe.pth "

[shida666]

Dear Sir
Thanks for the Baidu Pan.
However, the "convert_caffe_pretrain.py" can't convert the "vgg16-397923af" into torch version. I tried and it failed.
The network has different structure between vgg16-397923a and vgg16-00b39a1b I guess
I suggest people can download the vgg16-00b39a1b.pth on Google Colab and covert to torch version using "convert_caffe_pretrain.py" as Chenyun said in README.
Thanks !

I also encountered the same problem, did you solve it?

directly to insert the "python train.py train --env='fasterrcnn-caffe' --plot-every=100 --caffe-pretrain“ don't make the ‘’python misc/convert_caffe_pretrain.py‘’

[chenaifang]

I want to use vgg16_bn, but directly use vgg16_bn-6c64b313.pth in Pytorch. I can only get 58mAP. I would like to ask if there are other pre-training models of vgg16_bn? @shangjianan2 @chenyuntc @greenyin @ZichengDuan @Stephenfang51

[ZichengDuan]

@chenaifang Hi, Aifang. Unfortunately, I didn't explore other backbone networks for this project.

[chenaifang]

Excuse me, do you know how to get the pre-training model of vgg16_bn in general? @ZichengDuan

[yuanzhenhuan]

thank you对对对

[DanZhang123]

I want to use vgg16_bn, but directly use vgg16_bn-6c64b313.pth in Pytorch. I can only get 58mAP. I would like to ask if there are other pre-training models of vgg16_bn? @shangjianan2 @chenyuntc @greenyin @ZichengDuan @Stephenfang51

Hello, have you solved it?