conan7882 / googlenet-inception Goto Github PK

https://www.dropbox.com/sh/axnbpd1oe92aoyd/AADpmuFIJTtxS7zkL_LZrROLa?dl=0 不能下载

Thanks for you great job, it really help me.
But the net to download the pretrained GoogleNet npy file can not open. I don't know if it only happen on me. Anyway could you provide it for me?
Thanks so much.

When I run the python inception_pretrained.py,the result is Empty test kit.
How can I deal with the issue?

Sorry to interrupt，I have a question.
In the inception_cifar. py file, there is a four-path setting, where IM_PATH is set according to what.It would be nice to have a basis for the other three path Settings

I trianed the same dataset in caffe-googlenet with the batch size of 32 ,but can only set batch size of 4. That is strange.
My device is 1080Ti and has a memory of 12GB.

I take the liberty to bother you. I want to ask you a question about image classification, but images are not images in the usual sense. The experiment is to detect different objects with the collected WiFi signals. The above figure shows the rssi and phase change curves of the WiFi signal when a bottle is placed indoors. Now I want to classify objects according to these curves. Currently, Resnet50 DenseNet201 is used but the accuracy is not high and only 80。So I would like to ask what kind of network structure is better to use deep learning to classify this kind of image?

Hi there,

I wanted to try out the googlenet pretrained model, but it seems there is no googlenet.py in the lib/models directory. Also the googlenet.py from deeplearningmodel.net seems to be different.

Greetings

LRN (Local Response Normalization) layers are used in the original GoogLeNet a.k.a. original Inception architecture.

Going deeper with convolutions.
Szegedy, Christian, et al. 
Proceedings of the IEEE Conference on Computer Vision and Pattern Recognition. 2015.

Not sure where these layers have been implemented in the source code.

Thanks in advance. Any help would be appreciated!

Thanks for sharing such a nice code :-)

but when I tried to run demo you described, It occurs an importing error.

so it needs to be added that __init__.py in each nets, model, utils, dataflows
after I added it as a null file, It runs well :-)

Is it really necessary to scale the 32x32 images up to 224x224 to adapt to the original network? 10 classes classification is rather a light-weighted task. Maybe simply changing the pre-inception convolution layers to two 3x3 convolutions would still be deep enough while also keeping Inception-3a's input feature size 28x28.

Hello, Thanks for your job very much.
I have download the code and datafile 'cifar-10', After training, I predicted the file of ‘inception-cifar-epoch-99.meta’ with some different kind images, But the predicted results are the same 'cat'.
So, could you please give me some ideas?

And another question is the connect to download the pretrained GoogleNet.npy file still can not open.