Questions tagged [deep-learning]

a new area of Machine Learning research concerned with the technologies used for learning hierarchical representations of data, mainly done with deep neural networks (i.e. networks with two or more hidden layers), but also with some sort of Probabilistic Graphical Models.

Filter by
Sorted by
Tagged with
0
votes
0answers
8 views

The channel dimension of the inputs should be defined. Found `None`

Hello I'm trying to use SegNet in my project with tensorflow, for educational purpose. And I'm surely following someone else's code on GitHub: ...
1
vote
1answer
10 views

CNN's Understanding Max Pooling Better

In learning about CNN's I understand the max pooling can help decrease the computational load due to down sampling. Another thing mentioned is that max pooling can help provide a sort of "spatial ...
1
vote
0answers
8 views

How Can I Convert Dataset Annaotations To Fixed(YoloV5) Format Without Hand Encoding

So I Am Working On This Awesome Project On Object Detection,Where The Prior Task Is To Identify Brand Logos, So after Doing some research i found this dateset available for the brand logo For More ...
1
vote
1answer
21 views

How to decrease LSTM RMSE?

I am using an LSTM model to predict the next measurement of a sensor. The dataset looks as follows: There are approximately 13000 measurements. My code for the LSTM looks as follows: ...
0
votes
0answers
29 views

Should rescaling be used on test images in keras?

I am kind of confused regarding the topic. I have built a CNN architecture for the cat-dog image classification around 6000 images of cat and 6000 images of dog and I am predicting on test images. I ...
1
vote
0answers
11 views

Is there a list of neural style transfer architectures?

I'm trying to choose what neural style transfer architecture to use but I can't find a centralized list of all possible architectures I could choose from. Is there a place I could find this or could ...
0
votes
0answers
20 views

Using Object Detection or Image Segmentation without labelled input data to build a dataset to then be manually labelled?

I'm looking to build an object detection model or image segmentation model, ideally the latter, which will identify and label objects from satellite imagery but I don't currently have any labelled ...
0
votes
0answers
25 views

Is my choice of input space for a DQN problematic?

I am working on a reinforcement learning algorithm that attempts to solve a (discrete, finite) MDP with a non-Markovian reward function. It uses a DDQN with a large input space, which might be a bad ...
3
votes
0answers
34 views

Why margin loss is used in Capsule Network instead of Cross Entropy loss?

I'm reading the Capsule Network paper proposed by Hinton. I'm not sure why the margin loss is used instead of the cross entropy loss. Any intuitive explaination for this?
0
votes
0answers
6 views

Fusion (concatenation or elementwise) of coarse (deep) and high-res (shallow) features in ResNet, FPN and UNet

I understand this functionality, but I've neither the intuition nor reasons why this works. I'm looking at three cases of this fusion: ResNet, UNet and Feature Pyramid Net (FPN). In ResNet, in each ...
1
vote
0answers
10 views

Deep learning / computer vision technique: aggregating many input images to a single representation of the features within

I have a few thousand grayscale images, and I would like to generate a universal representation of the patterns within - a semantic/ordered composition of all features, so to speak. For instance, take ...
0
votes
1answer
27 views

Improving the performance of neural networks

I have 3 questions in mind about the neural network For the best model performance, is it better to train a model only on high resolution images or does it not matter whether the training data ...
0
votes
0answers
9 views

validation loss early increase (during warm-up)

Several questions have been asked about validation loss behavior during training of a DNN. It's clear to me that validation loss and accuracy are somehow correlated, but their curves can differ from ...
0
votes
0answers
14 views

MNIST trained network tested with my own samples

I trained a Dense Neural Network with MNIST dataset in order to classify 28x28 images of numbers. Now I was trying to make it work with my own samples (I draw the image of a "7" in paint and ...
0
votes
1answer
16 views

How does TensorFlow handle multiple samples?

Say the mini-batch has $N$ samples $(x, y)$, how will tensorflow utilize this $N$ samples to train the network. Will it do $N$ forward loop for each sample independently? Will it do $N$ backward ...
2
votes
1answer
36 views

Why do Transformers need positional encodings?

At least in the first self-attention layer in the encoder, inputs have a correspondence with outputs, I have the following questions. Isn't ordering already implicitly captured by the query vectors, ...
0
votes
0answers
6 views

Reduce Training steps for SSD-300

I am new to deep learning and I am trying to train my SSD-300 (single shot detector) model which is taking too long. For example even though I ran 50 epochs, it is training for 108370+ global steps. I ...
0
votes
0answers
14 views

How does state information get transferred when predicting with LSTM

I have a typical mutivariate time series forecasting problem that I want to solve using an LSTM, with mutliple features in the input sequnce and one feature in the output sequence. If I train my LSTM ...
0
votes
0answers
11 views

Convert a simple NN model to LSTM model (made using Tensorflow.compat.v1)

I have a simple neural net which I would like to convert to an LSTM model. Can someone please help me with the code? The following code contains the neural net part: ...
-1
votes
0answers
12 views

What each of the state-of-the-art deep learning models in NLP are best for? [closed]

It is not easy to find an good answer on the Internet to this question so I thought about asking it here. What each of the state-of-the-art deep learning models in NLP are best for? (eg which is best ...
0
votes
0answers
10 views

Architecture for ConvLSTM

I have an input data with 2000 samples each having shape of (5, 3, 178, 178) where 5 is time dimension, 3 is a color channel, and the rest are x and y-axis. Now I want to use ConvLSTM layer to predict ...
0
votes
0answers
12 views

Is an output layer with 2 units and softmax better than one with 1 unit and sigmoid for binary classification using LSTM?

I am using an LSTM for binary classification and initially tried a model with 1 unit in the output(Dense) layer with sigmoid as the activation function. However, it didn't perform well and I saw a few ...
0
votes
0answers
23 views

How to train a policy and a value network, implementing alphazero at chess

So, I'm trying to implement alphazero's logic on the game of chess. What I understand so far of the algorithm is: Load 2 models, one of which is the best model you have so far. Both these models have ...
0
votes
1answer
19 views

Working of Dense Layer

What kind of operation does Dense Layer perform to reduce dimemsion. So basically I have used Dense layer to compress the dimension all the time like from 10000 neurons to direct 2000 neurons or even ...
0
votes
0answers
9 views

LSTM - unable to get a 3D output

I have an array with shape (55834, 250, 30) and I'd like to get an output of the same shape from this LSTM model. ...
0
votes
1answer
19 views

Error during the compilation of a neural network in Vitis AI, "Not found op in super_const_dict: name: Decoder_Section_1_UpConv_1/kernel"

I'm following a Xilinx Tutorial about the implementation of a Neural Network in a System on Chip (ARM Processor + Xilinx FPGA) and I have come up with an error during the compilation step. I've ...
0
votes
1answer
19 views

How do I feed my keras model in batches?

I am trying to feed a Sequential model in batches. To be reproducible my example, suppose my data is: X=np.random.rand(24,432) Y=np.random.rand(24,432) My goal is ...
0
votes
0answers
16 views

How to decide number of hidden layers and number of neurons for Autoencoder for dimensionality reduction function?

I have been looking into deep learning and what caught my attention is the implementation of Autoencoder as a dimensionality reduction function for anomaly detection. I found out about it through the ...
1
vote
2answers
37 views

What effect repetitive data will have on the performance of the model

I understand that my question is very broad and that the correct answer may depend on various things. I want to get an idea in general what we may expect if we have repetitive data in our dataset. ...
0
votes
0answers
4 views

Huggingface - TypeError: 'TensorSliceDataset' object is not subscriptable

I'm trying to make my own model for translate a language to another with T5ForConditionalGeneration and Huggingface using no pretrained model (I need to use my own dataset and tokenizer because no ...
0
votes
0answers
5 views

Soft-NMS during Inference

I have read the paper about the soft-NMS and the problem that tries to solve. Does it make sense to use this algorithm during inference apart from training? It smooths the predictions on the bounding ...
0
votes
0answers
21 views

Predict value from array (Assisted Sequence Prediction)

TL; DR How to develop a machine learning implementation (in python) that given an array of values is able to predict its maximum percentage of increase? [11.082, 11.052, 11.052, ... 11.202, 11.202, 11....
0
votes
0answers
7 views

Custom vocabulary with Transformer model in Trax

I'm currently working on a programming language translation problem (NMT) using a transformer model with Trax. So I basically need to convert a language A to a language B. In order to tokenize the ...
2
votes
1answer
28 views

Data Art/ Data Visualization Art/ Information Art

Few days ago, I learned about data art/ data visualization art/ information art. I think I have interest in it. I want to see how I can use my data science skills in this area. However, I don't know ...
1
vote
0answers
13 views

Can a reformer model really handle long-range dependency?

I read this article about new attention model called Reformer. Here is the main strength of this model: The Reformer pushes the limit of longe sequence modeling by its ability to process up to half a ...
2
votes
1answer
42 views
+50

Is the explained variance a good metric for autoencoders?

I want to evaluate how an autoencoder will perform on my data. Now, I can do this with the mean squared error of the decoded data compared to the original data, and this is fine when comparing this ...
0
votes
0answers
25 views

How to use Softmax and CrossEntropyLoss in a classification problem?

I am new in the deep learning field and I would like to go in to understand some concepts. I started learn with pytorch and I have the next question: I constructed my own dataset that contains images ...
1
vote
1answer
28 views

Robustness vs Generalization

I don't quite understand the difference between robustness and generalisability in relation to image processing (CNN). If my model generalises well, it is also robust to changes in the image material. ...
1
vote
0answers
19 views

Should I reshuffle the training set when benchmarking neural networks?

I'm trying to set up a fair benchmarking between various RNN models, where each of them is trained until convergence with a fixed random seed. Because the task is very costly, I am only able to run ...
0
votes
0answers
10 views

How to represent source code in NLP for tasks like code retrieval?

How to represent code features in NLP? Can code be treated like a language and pour into the neural network? Are there some work, paper or material around this topic?
0
votes
0answers
11 views

how to in enhance A3C entropy?

I'm trying to implement this A3C code in my custom environment, and I have a basic understanding of the algorithm. The algorithm worked, but it did not give me a good performance. I looked into ...
0
votes
0answers
11 views

How many layers/ convolutions does DensNet-169 have?

I am a little confused by the numbers. From the Digram in the paper I count 312 convolutions without the transition layers. Can anybody help understand how the 169 in the name is calculated?
0
votes
0answers
17 views

How can I create an XOR Gate using two hidden layers and one output layer (besides the or gate)?

Here is my nueron class with activation function (sigmoid), fn for printing truth table and inputs Mathematically and_gate + nor_gate + nor_gate are giving me the correct output. But the function is ...
0
votes
2answers
49 views

High accuracy in mode.fit but low precision and recall. Overfit? Unbalanced? Error?

Hello ive been training a CNN with keras. A binnary clasificator where it says if a depth image has a manhole or not. Ive labeled manually the datasets with 0 (no manhole) and 1(it has a manhole). I ...
1
vote
0answers
11 views
7
votes
1answer
101 views

Why deep learning models still use RELU instead of SELU, as their activation function?

I am a trying to understand the SELU activation function and I was wondering why deep learning practitioners keep using RELU, with all its issues, instead of SELU, which enables a neural network to ...
1
vote
0answers
21 views

How do dilated convolutions used for upsampling the inputs in FCNs?

I am reading the paper (Long et al., 2015) on fully convolutional networks (FCNs), and I came across the section where the authors describe dilated convolution as the trick to compensate for the cost ...
0
votes
1answer
16 views

predict a binary vector of size 40

I have a dataset of shape (2600, 95) with first 55 columns are features and 40 columns are label. Label is a binary matrix of size 10x4 that flattened, and features are real valued numbers ranging (0....
1
vote
1answer
41 views

Splitting the dataset manually for k-Fold Cross-Validation

I manually divided the dataset into three sets: train, test, and validation. Each set includes several folders, one for each patient. Each patient has many images from a different point of view. As a ...
0
votes
0answers
11 views

How to improve accuracy? BERT

Dataframe: ...

1
2 3 4 5
83