Questions tagged [neural-networks]
Questions involving defining, training, executing, importing and exporting neural networks.
542
questions
2
votes
0
answers
43
views
Unable to Replicate an ANN Using the Output from Predict Function
I’m working on constructing an ANN. I first used the Predict function with the method “Neural Network” to determine the structure according to my data.
This is the given output of the structure:
...
- 21
18
votes
4
answers
634
views
Improve scanned image with handwritten text for high-contrast printing: how it can be done with Mathematica?
I have a set of scanned documents with handwritten text which by default is printed with very low contrast. Here is an example (a part of a document):
...
- 58.4k
2
votes
0
answers
58
views
GAN using Graph Convolution Network
I am using the Pokemon GAN tutorial in:
https://blog.wolfram.com/2020/08/18/generative-adversarial-networks-gans-in-the-wolfram-language/
Where convolution blocks and deconvolution blocks are given by:...
- 21
4
votes
0
answers
128
views
Diffusion probabilistic model in deep generative modeling
Diffusion Models (http://proceedings.mlr.press/v37/sohl-dickstein15.html and https://arxiv.org/abs/2006.11239) are generative models gaining popularity in the community of deep generative modeling, ...
- 535
4
votes
0
answers
46
views
Training NetGraph with a 2nd order optician
I would like to train NetGraph using a quasi Newton algorithm rather than a first order stochastic gradient method. However, in NetTrain there are only such optimisers. Is there any way I can make ...
- 81
1
vote
0
answers
63
views
Why accuracy of Neural Network does not change as its structure is changed?
I have the following neural network:
...
- 125
4
votes
0
answers
80
views
Derivation from a network
how can I represent the derivative of a network (e.g. NetGraph or NetChain) as a network? I have tried it with FunctionLayer[] and NetPortGradient, but without success.
I would be very pleased to ...
- 81
1
vote
0
answers
46
views
How to measure accuracy of a neural network?
I have this Neural Network for a specific data set:
...
- 125
3
votes
0
answers
72
views
Applying Super-Resolution to differently sized images
I would like to make some images larger and noticed that there is a page in the documentation for doing super-resolution using a network structure. It's called "A Model for Super-Resolution."...
- 65.8k
4
votes
1
answer
90
views
Does RandomArrayLayer implement the reparameterization trick?
I want to backpropagate through random operations (e.g. learn the mean and variance of a Gaussian random variable). I assumed that
...
- 101
2
votes
0
answers
36
views
Create "PredictorFunction" with NeuralNetwork
I tried to construct a predictor for a continuous variable using neural networks, something like
f(a,b,c,d,e) = x,
trying to predict x for a given set of a,b,c,d, and e.
The builtin ...
- 21
0
votes
0
answers
27
views
Reshaping Output of a NetGraph during training
I'm trying to train a (per-pixel) segmentation neural net using the undocumented HDF5 training facilities that are laid out in this question. Here is a slightly modified version of the answer to that ...
- 720
2
votes
1
answer
67
views
How to make the output of NetTrain obey constraints?
Imagine I have this
trained = NetTrain[LinearLayer[], {1 -> {0.1,0.7,0.2}, 2 ->
{0,0,1}, 3 -> {0.4,0.4,0.2}, 4 -> {0.1,0.0,0.9}}]
I want to ...
- 1,410
1
vote
1
answer
44
views
BasicRecurrentLayer with no bias and linear activation function?
Trying to design a small neural network for modeling a LTI system, I'd like to configure a BasicRecurrentLayer with these particular settings:
No bias vector should be used (I've tried the option &...
- 123
6
votes
1
answer
140
views
Unexpected high memory consumption using out-of-core NetTrain
I'm working on training a neural network on an image dataset. There are 14k images and each image contains 3x150x150 pixels. I have built a generator function following the approach in Training on ...
- 386
4
votes
1
answer
168
views
Constructing neural networks from Keras example
I am trying to convert neural networks from a book to Wolfram language. The book has this example with the quote
two layers of 10 neurons and a dropout of 0.2 for each
...
- 545
0
votes
0
answers
35
views
Transfer Learning Model in two ways
I have tried a lot. But no method can't work. How to correct it?
The link is the code below.
Dataset Preparation -> blancedTrain, blancedTest
...
- 1
3
votes
1
answer
48
views
Reuse same layer of Neural Network
I hope to wrap up ConvolutionLayer in a way so that it takes two inputs, does the same operation on those, and outputs their result respectively. For example, ...
- 1,971
2
votes
1
answer
75
views
CrossValidate for image classification with NetTrain
I am trying to set up transfer learning using instructions from this repository page on EfficientNet. I am able to perform a stand-alone training and testing using the instructions on that page. The ...
- 545
2
votes
0
answers
40
views
Neural Nets: Can the back propagation/deep dream step be computed within the Net itself
I am trying see if I can do a deep dreaming algorithm without exiting a neural net.
This Generalized Backpropagation for Neural Networks (e.g. DeepDream) has a nice discussion of using the ...
- 2,647
1
vote
1
answer
87
views
Is there a difference between ElementwiseLayer@Ramp and Ramp in NetGraph?
This is about neural net! I put what I want to ask in the picture! This is the code!
...
- 41
0
votes
0
answers
38
views
Ranges of ANN predict values too narrow
I plan to train a neural network on a large dataset. Each example contain 16 parameters (inputs) and 1 output. A snapshot of the dataset looks like:
net = ParallelTable[Parasets[[i]] -> {degraeffr[...
0
votes
0
answers
45
views
How to manipulate neural network output data?
I have a large list data for neural network training in the form
{x1 -> y1, … , xn -> yn}
Where the outputs $y_i$ is either 1 or 0 and the inputs $x_i$ are ...
- 285
2
votes
0
answers
34
views
EmbeddingLayer with variable input length is incompatible with LinearLayer?
EmbeddingLayer\[size, n\] trains a representation of a list of integers between 1 and n into a real vector of dimension ...
- 2,360
7
votes
1
answer
197
views
Exactly recreating image NetEncoder output
I need to run an exported NetChain in ONNX, surprisingly the model exports and runs successfully, but there's a discrepancy in the image loading stage (the NetEncoder'ing that must be written manually)...
- 3,523
0
votes
0
answers
52
views
Spliting training dataset in ANN training using NetTrain
I am training a neural network on large training dataset containing 25,525,200 examples. each example contain 15 parameters (inputs) and 1 output, such as
Inputs 0.015,83.6,0.663,0.104,0.097,0.167,0....
3
votes
1
answer
77
views
Automated mapping between input and output columns of FeatureExtractorFunctions
This problem is about FeatureExtraction. It arose in the course of training a neural network where the data contains a mixture of nominal and numeric variables. Basically, my code looks like the ...
- 2,982
1
vote
0
answers
30
views
Repeated calculation with neural network measurement slows down
I have a list of 200 neural nets nets0. I have to measure the output of several layers of each of them on a list of 1000 associations ...
- 215
1
vote
0
answers
93
views
Increase the speed of Machine learning based on 'Predict'
I am training a ANN model with a large dataset (~million level) using Predict or NetTrain function. I do not have a GPU machine but have a server equipped with 16 CPU cores. How can I maximize the ...
1
vote
1
answer
69
views
Issue with Transfer Learning using VGG16 NetModel
I am trying to use the VGG16 architecture available in the Wolfram Neural Network Repository to classify the Fashion MNIST dataset. The VGG 16 by default accepts images of size 3x224x224. However the ...
- 89
1
vote
0
answers
81
views
Unknown Error in NetGraph with RTX 3090 GPU
I am running NetGraph on my RTX 3090 GPU. The Net itself loads properly in my GPU, having an approximate size of 9GB (the 3090 has 24GB of memory). However, once the net actually starts NetGraph ...
7
votes
1
answer
91
views
How to calculate a PDF of a neural network output?
I have to find the probability density function of the output of an internal layer in a neural network.
I wanted to do it using NetMeasurements adding a layer like
<...
- 215
4
votes
1
answer
56
views
`NetNestOperator` with custom input / output ports
I'd like to use NetNestOperator on a NetGraph that has custom input / output port names, such as
...
- 1,995
16
votes
0
answers
367
views
Neural Network Automatic Differentiation ("Autograd")
Automatic differentiation methods in general have been the topic of past questions, but this is specifically about automatic differentiation of neural networks.
The motivation is an old but useful ...
- 2,360
4
votes
1
answer
106
views
Prevent overfitting in Predict function
I know that you can prevent overtraining automatically in a network if you use NetTrain with NetChain in Mathematica version 12, ...
- 499
13
votes
1
answer
416
views
Wolfram Neural Net Repository Questions
I am a research scientist in the machine learning team, and mostly involved with the neural net repository. One of the discussions with my group raised these question in my mind, and these might be ...
- 131
1
vote
0
answers
34
views
Problems extracting a layer from a deep network
I am trying to reproduce this deep net for pose estimation. (v. 11.3, MacOS)
I download the VGG network, as specified:
...
- 36.5k
0
votes
0
answers
60
views
Mathematica accuracy errors when creating data
I created training data for a neural network and below are two errors I got
The only code that uses "FindRoot" command is the following
...
- 285
0
votes
0
answers
50
views
0
votes
1
answer
47
views
Error Using NetMeasurements (assessing neural network performance)
I am trying to assess how well a neural network regression model performs on test data. To do this I have used NetMeasurements command.
...
- 285
1
vote
1
answer
155
views
Building multilayer perceptron Neural Network
I’m new to machine learning and am a bit confused about how to create a neural network (multilayer perceptron) with a specific design.
I would like to create a neural network with 40 features/neurones ...
- 285
2
votes
1
answer
129
views
Anyone has any succes on GPU NetTrain with RemoteBatchSubmit on AWS EC2?
Since 12.3, the RemoteBatchSubmit to EC2 does not work anymore.
First, it features the same bug as Version 12.3.1 Problems with NetTrain, where the MXNetResources paclet is 12.4 instead of 12.3.
...
9
votes
3
answers
538
views
Version 12.3.1 Problems with NetTrain
Until now I am using MMA 12.2 because NetTrain istn't running stable with 12.3.
Yesterday I have downloaded and installed 12.3.1 to check if the problems using NetTrain are gone.
I have loaded my ...
- 183
4
votes
1
answer
183
views
How can I free GPU memory allocated with NetTrain
When I train a neural network with NetTrain[..., TargetDevice->"GPU"], wolfram's neural network code automatically allocates memory on my GPU for my ...
- 938
0
votes
0
answers
76
views
Generating a sequence from a number using RNN
Can someone please provide me with a neat example about generating a sequence from a number using RNN? (Possibly without using NetGraph[].)
Edit:
I'm trying to ...
2
votes
0
answers
29
views
Does NetChain have a "Distribution" property like PredictorFunction?
PredictorFunction has an incredibly useful "Distribution" property. In the past, this has helped me filter out bad predictions, like this:
...
2
votes
1
answer
130
views
Upsampling layer for convolutional autoencoder
I've been trying to build a convolutional autoencoder with Mathematica 12.3. For this, I need to reverse the action of PoolingLayer[]. Is there any layer similar to ...
- 2,994
2
votes
0
answers
117
views
Simulated Cartpole Example in Mathematica
I am interested in modifying the code underlying the reinforcement learning example SimulatedCartpole in Mathematica 12.
This example is still listed as Experimental. The help function indicates that ...
- 131
3
votes
1
answer
283
views
GAN LSTM Time Series
Does anyone know if it is possible to use LSTM or another RNN in GAN architecture as the generator? Here a reference: 1
Thx!
To clarify, what I am asking is it possible in MMA 12.3 (my current version)...
- 141
1
vote
0
answers
49
views
Inputing pairs of objects to neural networks (eg. pairs of images and sounds)
I'm trying to make a net graph where the input (a pair of objects of distinct but fixed types) will be passed simultaneously trough two net encoders. One will get an array from the image (the first ...
- 862