# Questions tagged [neural-network]

7161 questions

1

votes

1

answer

37

Views

### Which loss function for ratings

I'm have the MovieLens 100k dataset and have join the movie & user datasets together and split the ratings like this:
age 49
gender 1
year 1997
unknown 0
action 0
adventure...

1

votes

1

answer

369

Views

### Overfitting, underfitting or good fit?

So im training an lstm rnn for a binary text classification task and i am having some issues understanding the loss results. The training set is roughly 700 000 examples, and the validation set is around 35000 examples.
The validation set and training set are independent, so i am not training on th...

1

votes

0

answer

118

Views

### I do not know why in my Keras neural network model, the prediction shape is not consistent with the shape of labels while training?

I have built a Keras ConvLSTM neural network, and I want to predict one frame ahead based on a sequence of 10 time steps:
Model:
from keras.models import Sequential
from keras.layers.convolutional import Conv3D
from keras.layers.convolutional_recurrent import ConvLSTM2D
from keras.layers.normalizati...

1

votes

2

answer

87

Views

### Gender Voice with Python coding

I am working on a sample data set from a link below.
https://www.kaggle.com/enirtium/gender-voice/data
I am trying to open .csv file(maybe I am opening it wrongly) and trying to create fully connected neural layers. Then, I am trying to train them but unfortunately, I am getting input shape not fit...

1

votes

1

answer

44

Views

### Append inverse_scale NN output and model.predict_classes output into a csv

I have trained a Neural Network and I want to append the prediction values to inverse_scaled test data so I can check the predictions vs the original feature values. However, when I run the code, the following line:
Xtest['prediciton'] = pred
throws the following error:
IndexError: only integers, sl...

1

votes

0

answer

46

Views

### In Keras,already have fed a proper shape and dtype value for placeholder , why it raised an Invalid ArgumentError error ?

I need to tune parameter for a CNN at the first layer ,So what i did is made a circulation ,and each step changed hyper parameter (the parameter is the number of kernel and the size of kernel ) .then output the result to local file,including the acc and loss ,so on .
Frame is :
for kernel_number ,...

1

votes

2

answer

126

Views

### Neural network works with the cross entropy and does not with an other loss function

I'm using tensorflow with julia to create a neural network.
I can create a network with the cross_entropy loss function and it works:
ENV['CUDA_VISIBLE_DEVICES'] = '0' # It is to use the gpu
using TensorFlow
using Distributions
function weight_variable(shape)
initial = map(Float32, rand(Normal(0, .0...

1

votes

0

answer

31

Views

### Remove input_variable node from cntk model

I want to remove the input_variable from the model in order to make it auto_regressive (i.e. apply UnfoldFrom on it)
I have a simple example below.
def auto_regressive(model):
@C.Function
def unfold(seq_start, dyn_axis: Sequence[Tensor[121]]):
unfoldfrom = UnfoldFrom(lambda x: model(x))
return unfo...

1

votes

1

answer

34

Views

### Permutate over output of CNN to return smallest loss and rearange output

Lets say i am detecting dogs on images.
Output of my CNN is
Dense(24,activation='relu')
Which means i want to detect up to 6 dogs ( each dogs should be represented by xmin,ymin,xmax,ymax = 4 values , 4 * 6 = 24 )
Now lets say i have two dogs on pictures and their positions are ( bounding box )
dog...

1

votes

0

answer

78

Views

### why does my Neural network not work with MNIST dataset?

i have written the attached neural net twice reading each line carefully i have taken reference from tariq rashid's book how to train your neural network
in the book hi claims to have an accuracy of upto 95%.
But my code does not converge, accuracy revolves around 10% which means it is guessing rand...

1

votes

1

answer

52

Views

### How to use scikit-learn metrics in CNTK?

I wish to use classification metrics like matthews_corrcoef as a metric to a neural network built with CNTK. The way I could find as of now was to evaluate the value by passing the predictions and label as shown
matthews_corrcoef(cntk.argmax(y_true, axis=-1).eval(), cntk.argmax(y_pred, axis=-1).eval...

1

votes

0

answer

33

Views

### Find the smallest loss in multiobject detection

I am detecting multiple ( up to 10 ) objects on video image or anything.
input is image and labels for that image, for example
[[box1],[box2],[box3],[00...],....]
Each box is represented as xmin, ymin , xmax, and ymax.
What is causing my trouble is situation, when my output of CNN is something like...

1

votes

0

answer

62

Views

### TensorFlow error: Shape [10] has rank 1 < 2

I am currently working on the facebook comment prediction dataset (http://uksim.info/uksim2015/data/8713a015.pdf). As an evaluation metric, I use top_10 and AUC (as described in the paper). When I try to evaluate top_10 on test set, I get error, 'TensorFlow error: Shape [10] has rank 1 < 2.'.
My dat...

1

votes

0

answer

170

Views

### Testing Neural Network with Continuous Output in R

Trying to use the neuralnet package in calculating the continuous output and apply it to my testset to calculate error rate.
However, my predicted output seems to be the same.
m1

1

votes

1

answer

109

Views

### Load dataset for DeepLearnig model

I'm new in deepLearning and tensorflow. I'm trying to create a neural network with tensorflow and the places365 standard dataset.
I've seen in the tensorFlow's MNIST example that, at the beggining you have to load all the data. That's fine if you have a small dataset but this is not the case. Places...

1

votes

0

answer

62

Views

### How to train an existing multilayer perceptron in weka

I want to know if there is a way of training an existing weka multilayer perceptron.
What I mean by this is that I have trained a model for 500 epochs and now I want to train it for another 100 to see if that improves its performance, potentially also changing the learning rate and/or momentum. Is...

1

votes

0

answer

93

Views

### Tensorflow model does not minimize error

I am trying to use Tensorflow to implement a non linear regression (with 4 linear terms and 4 nonlinear terms based on Tanh(x).
The sum of squared errors , which is supposed to be minimized, only increases. After a relatively few training steps, the weights and bias become 'inf'
There ought to be...

1

votes

0

answer

24

Views

### HOG feature based recognition

scene
I used haar cascade to detect face and that face ROI was passed to estimate its HoG feature...while calculating HoG feature...I used 8x8 patch of image(assuring all image was a square matrix).
problem
I used to create a 9 bin histogram for each patch.for example for a image ROI of 177x177 dime...

1

votes

1

answer

163

Views

### Fine-tuning a neural network in tensorflow

I've been working on this neural network with the intent to predict TBA (time based availability) of simulated windmill parks based on certain attributes. The neural network runs just fine, and gives me some predictions, however I'm not quite satisfied with the results. It fails to notice some very...

1

votes

1

answer

86

Views

### Very low accuracy on Digit recgonition dataset with images having 4 channels, using Convolutional Neural Networks

I am currently working on a digit recognition challenge by Analytics Vidhya, the link to which is https://datahack.analyticsvidhya.com/contest/practice-problem-identify-the-digits/ .
The images in the dataset pertaining to this challenge are of dimensions 28*28*4 (28 = length = width , 4 = no. of c...

1

votes

1

answer

1.1k

Views

### Error when checking input: expected input_19 to have 4 dimensions, but got array with shape (1190, 200, 200)

I am new to CNN and I am not able to identify how to solve this problem.
In this code I am training a set of images to obtain mask from convolutional network.the images are grayscale with shape (200,200). I am not able to identify where I am making a mistake.Also everytime I run my code there is err...

1

votes

1

answer

24

Views

### Is there any way to add a new layer in caffe that keeps weight which type is unsigned int?

I want to add a new binary convolution layer that implement the xor-net with binary weight and the problem is that I don't know how to save the weight as type of unsigned int. Is there any way to change the blobs_ data type of a layer in caffe?

1

votes

1

answer

20

Views

### How to merge two trained incpton_v3 models with difrent number of class for multiple output?

example:
inceptionv3_model1 with 10 classes (banana, orange, grape...)
inceptionv3_model2 with 3 classes (expired, note expired, normal)
I need to merge this two models to get multiple output result. for example expired banana or normal banana...

1

votes

2

answer

1.3k

Views

### InvalidArgumentError: You must feed a value for placeholder tensor 'Placeholder_1'

I trained a simple neural network with TensorFlow on the MNIST dataset. The training portion of the code works fine. However, when I feed a single image into the network, it gives me the following traceback:
Traceback (most recent call last):
File '/Library/Frameworks/Python.framework/Versions/3.6/l...

1

votes

0

answer

328

Views

### Tensorflow memory allocation error: Dst tensor is not initialized - plenty of memory on GPU's

When I begin the training of a feedforward 2 layer ANN with GPU on a new cluster using tensorflow, I am greeted by a Dst tensor is not initialized error. This error is apparently when there is not enough memory to handle the batch size, however the error states that the allocator ran out of memory w...

1

votes

0

answer

98

Views

### How can I make a “convolutional” kernel with 3x3 px input and 2x2 px output?

I'm working on a toy problem trying to increase the resolution of an image by a factor of 2 with a keras model. A basic operation to achieve this with keras is Conv2DTranspose. Using the functional model API, I split each pixel into four with:
upconv = Conv2DTranspose(kernel_size=(2, 2), strides=2,...

1

votes

2

answer

82

Views

### Training CNN for NIST digits using tiny-dnn

I have been trying to train a CNN using tiny-dnn library for digit recognition. The database used was NIST 19. The number of samples per class is 1000 for training and 30 for testing. So total number of samples for training is
1000*10=10000. OpenCV is used for image processing.
The maximum accuracy...

1

votes

1

answer

180

Views

### Fail to train large data in pytorch

I tried to construct two fully-connected layers in pytorch to embed features like [x1,x2,...,xn] into multiple targets [y1,y2,y3,y4,y5]. I post my code below:
class FullConnect(nn.Module):
def __init__(self):
super(FullConnect, self).__init__()
self.fc = nn.Sequential(
nn.Linear(195, 100),
n...

1

votes

0

answer

39

Views

### Is it possible to forecast a variable number of points using LSTM neural nets?

I'm currently working on time series forecasting using a LSTM network. The univariate dataset looks like this:
As shown in the graph, each cycle's period (ranging from a peak to peak) grows over time. Is it possible to train a model to predict a full cycle from this dataset? I've used a Keras LSTM n...

1

votes

1

answer

26

Views

### TF Graph does not correspond to the code

I am trying to create a very simple neural network reading in information with the shape 1x2048 and to create a classification for two categories (object or not object). The graph structure however, deviates from what I believe to have coded. The dense layers should be included in the scope of 'inne...

1

votes

1

answer

37

Views

### Normal distribution of input data [closed]

Should I normalize input data to normal distribution before fit it into RNN? If yes, why? At the moment almost all the columns are right shifted, so it's not a normal distribution at all.

1

votes

0

answer

39

Views

### Understanding distributing deep learning with tensorflow

I have written an RNN with tensorflow that trains on some time series data. I am training the model on moving window taking in 2 time slices and predicting for the third.
I batch my inputs and labels and run the entire set of batches over 30 epochs. The model runs fine as a single TF application....

1

votes

1

answer

358

Views

### Feed stacked RNN output into fully connected layer

I'm trying to solve a regression problem with a stacked RNN in tensorflow. The RNN output should be fed into a fully connected layer for the final prediction. Currently I'm struggeling on how to feed the RNN output into the final fully_connected layer.
My input is of shape [batch_size, max_sequence...

1

votes

0

answer

172

Views

### CNN Overfitting

I have a siamese CNN that is performing very well (96% accuracy, 0.08 loss) on training data but poorly (70% accuracy, 0.1 loss) on testing data.
The architecture is below:
input_main = Input(shape=input_shape, dtype='float32')
x = Conv2D(32, (3, 3), padding='same', activation='relu',
kernel_regula...

1

votes

0

answer

51

Views

### Back propagation with batch for neural network

I'm learning about back propagation from a coursera course.
In it they say that back propagation is only done on a single training example at a time.
This description can be found at this link
https://www.coursera.org/learn/machine-learning/lecture/1z9WW/backpropagation-algorithm(9 minutes in).
H...

1

votes

0

answer

82

Views

### Calculating error in neural network back propagation

This is a very simple question as I am new to the concepts.
I have a 4-4-1 neural network that I am running on 16x4 binary data to predict a 16x1 column of outputs.
I have utilized random weights and biases to generate a rough predicted output vector. I then calculate a vector of errors (actual-outp...

1

votes

1

answer

200

Views

### Keras generator

I wrote my generator, but I get an error at the output
def batch_generator(X_data, y_data, batch_size):
samples_per_epoch = X_data.shape[0]
number_of_batches = samples_per_epoch/batch_size
counter=0
index = np.arange(np.shape(y_data)[0])
while 1:
index_batch = index[batch_size*counter:batch_size*(co...

1

votes

0

answer

379

Views

### How do we import a .pkl file in Matlab?

I trained a neural model in python and saved it in my directory in pkl format.
I want to import this pkl file in Matlab. Thanks in advance.

1

votes

0

answer

257

Views

### Keras accuracy calculation and meaning

I am detecting objects using CNN and keras
When i test/train model it outputs acc and loss.
I am using MSE loss functions so i understand what loss mean, however what is accuracy and how is it calculated? I have 4000 loss and accuracy 80% which is stupid. It does not detect object 80% correctly. Wha...

1

votes

2

answer

174

Views

### Is it possible to load weight with definite amount or with an initalizer in NN in keras in model.add?

Is it possible to load weight in NN in keras in model.add?
I want to load the weight based on Xavier or another initializers. How I can do this in keras?
For instance, weight=[w1,w2,w3,w4] how we could do this in keras?
For instance, in TF we have: initializer=tf.contrib.layers.xavier_initializer()