Deep Knowledge

I applied on another domain, was fine, can you increase the resolution: make your image size: 160 x160

Do you think that could be improved by **multiprocessing** or **joblib** packages?

@xiahaifeng1995 , @okokchoi , you could also move the following into the training and save it with mean. `conv_inv = np.linalg.inv(train_outputs[1][:, :, i])` So, in the training part: ``` train_outputs...

@okokchoi , Did you compute the conv_inv and save it? see, in the training part, and replace it with : ``` for i in range(H * W): cov[:, :, i]...

I think so, please try it and share your findings

> Hi @GreatScherzo , thanks for your improvement, it is faster but the score is different , the scores for the normal images are higher than the defective images, do...

> > @okokchoi , Did you compute the conv_inv and save it? > > see, in the training part, and replace it with : > > ``` > > for...

By the way, i fixed that, So now, the distance is vectorized, works if you have one or many images The inference time was improved a lot

you do not need to save cov as well //save only train_outputs = [mean, conv_inv]

@xfby2016, I hope you can get the idea, You can also see this [blog](https://medium.com/the-owl/extracting-features-from-an-intermediate-layer-of-a-pretrained-model-in-pytorch-c00589bda32b) `def CutModel(): pretrained_model = resnet18(pretrained=True, progress=True) NetFeatureSize = OrderedDict([('layer1', [64]), ('layer2', [128]), ('layer3', [256]), ('layer4', [512])])...