Closed. This question needs to be more focused. It is not currently accepting answers.
Want to improve this question? Update the question so it focuses on one problem only by editing this post.
Closed 2 years ago.
Improve this question
I(x, y, no of channels) is the image, and Fi(x, y, no of filters ) is the feature map at some layer 'i'.
Given the architecture of a Convolutional Neural Network like VGGNet and a feature map after a certain layer Fi, is there an efficient way to find which pixels of the input image I, that contribute to a location in the feature map?
I will want to implement this in python.
Related
Closed. This question needs details or clarity. It is not currently accepting answers.
Want to improve this question? Add details and clarify the problem by editing this post.
Closed 2 years ago.
Improve this question
I am making AI that will recognise price from url. I found best arcitecture for it but still have last question: How many neurons should I use in last output layer? The otput should look like: "1100usd, can anyone help me?
Well, you don't need a dollar sign to be a part of the neural network's prediction, so you are looking for it to output a single value. Therefore, your output layer would be a dense layer with a single neuron.
Closed. This question needs to be more focused. It is not currently accepting answers.
Want to improve this question? Update the question so it focuses on one problem only by editing this post.
Closed 2 years ago.
Improve this question
Is there some sort of software that can do so? Specifically, I would like to visualize Resnet18. Is there no other way other than to just draw it myself? Here is an example of what I want to see:
Sample Architecture Visualization
You can use this one : http://alexlenail.me/NN-SVG/LeNet.html . It lets you visualize neural networks by letting you modify several parameters and finally lets you export the architectures as SVG files. You can also choose between 3 visualization styles, namely FCNN, LeNet & AlexNet.
Closed. This question needs to be more focused. It is not currently accepting answers.
Want to improve this question? Update the question so it focuses on one problem only by editing this post.
Closed 3 years ago.
Improve this question
As the title says I have an image (well a bunch of images) and I want to turn it from a 200x200 image into a 1-D list of 40,000.
Try to flatten the image and convert it to list.
img.ravel().tolist()
A ndarray of shape Nx200x200 can be converted by reshaping
bunch_of_images.reshape(N, 40000)
Closed. This question needs to be more focused. It is not currently accepting answers.
Want to improve this question? Update the question so it focuses on one problem only by editing this post.
Closed 4 years ago.
Improve this question
Can anyone please help me to distinguish between darknet and darkflow. Advantages of one over the other. My understanding about YOLO (You Only Look Once) is an algorithm for fast object detection.
Darknet is the name of the framework YOLO is originally implemented on.
Note DarkNet-XX (XX=19/51) is also the name of the backbone YOLO uses.
Darkflow is a nickname of an implementation of YOLO on TensorFlow.
Closed. This question needs to be more focused. It is not currently accepting answers.
Want to improve this question? Update the question so it focuses on one problem only by editing this post.
Closed 5 years ago.
Improve this question
With boosting and random forests, I used early stopping and plotted the train AND test error while training in order to get a feeling when the model is overfitting.
a) in tensorflow (multilayer perceptron), I can plot the training error while training (cost/epoch graph), but how can I get the test error graph while training?
b) is there a built-in cross-validation function in tensorflow (eg 5-fold cv)? If not, what is the most efficient way to do cv with tensorflow?