Closed. This question does not meet Stack Overflow guidelines. It is not currently accepting answers.
This question does not appear to be about programming within the scope defined in the help center.
Closed last year.
Improve this question
Feed forward neural network do they belong to deep neural network or ANN?
I have a doubt whether to call my frame work as deep neural network or a artificial neural network
FFN refers to neural networks in which information flows into one direction only, as opposed to Recurrent Network for instance where information can flow back form the previous time steps.
An FFN is a type of ANN. "Deep" usually means several layers stacked. A one layer ANN would usually be qualified as "Shallow".
Related
Closed. This question does not meet Stack Overflow guidelines. It is not currently accepting answers.
This question does not appear to be about programming within the scope defined in the help center.
Closed 2 years ago.
Improve this question
I have two traffic related datasets. One contains traffic signs and the other traffic lights.
I want to merge the two datasets and train the model to detect both of them.
Will unlabeled traffic-signs from the traffic-light dataset affect the training process and vice versa?
From what I've read so far YOLO also learns contextual information about the objects and that's why this concern.
As you mentioned and as I found there, "YOLO sees the entire image during training and test time so it implicitly encodes contextual information about classes as well as their appearance", but I see that the meaning is that it considers the direct areas around the labels to add their information to the trained network, thus, you will likely only lose the information from the unlabeled items, but it will not impact the labeled items negatively.
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
I need to build a deep learning model for image classifying. I need to train a deep learning model on those data and then deploying it on real machines.
In conclusion, My main problems are:
Images are very big, which leads CUDA to memory issues. what shall I do to prevent my model running out of memory limit?
Besides, I need a very fast inference, because the model will be used on real deploy environment. The speed is very important for timely response.
I need to solve both the 2 problems to deploy my model.
I think it is important to reduce the size of the images. reshape them if necessary, which can significantly reduce the memory cost.
I think you can try different batch size. Becasue batch size is directly related to training and inference speed of deep learning. But I think better GPU machine card is more important for image classifying with deep learning network.
I think you need better GPU card as deep learning is machine hungry.
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
I'm very new to NLP and Deep Learning field and want to understand that after vectorization of a whole corpus using Word2Vec, Do I need to store the word vector values locally?
If yes I want to make a chatbot for android. Can anyone please guide me for this?
word2vec embeddings can be saved:
in first layers of your deep model. It's rare approach, because in this case you can't use this word2vec for other tasks.
as independent file on disk. It's more viable apporach for most use cases.
I'd suggest to use gensim framework for training of word2vec. Here you can learn more how to train word2vec and save them to disk: https://radimrehurek.com/gensim/models/word2vec.html
Particularly, saving is performed via:
model = Word2Vec(common_texts, size=100, window=5, min_count=1, workers=4)
model.save("word2vec.model")
Training of chatbot is much more difficult problem. I can try to suggest you a possible workflow, but you should to clarify what type of chatbot do you have in mind? E.g. should it answer on any question (open domain)? Should it generate answers or it will have predefined answers only?
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
I understand the concept of having muliple layers, backpropagation, etc. I even understand that an activation function would squash the output to a certain range based on the activation function used. But why do we even require this? What happens if we continue with the actual result without an activation function?
Please help me understand, but in pure english - no graphs/formulas please - i want to understand the concept behind it
If your activation function is just a(z)=z (a linear neuron), the activation is just the weighted input (plus bias). In this case, the activation of each layer is a linear function of the previous layer's activation. You can quite easily convince yourself that the combined effect of many layers (i.e. a deep network) is still a linear function. That means that you could get exactly the same result with just an input layer and an output layer, without any hidden neurons. In other words, you would not win any additional complexity in what your network can do by adding hidden layers, so no advantage going to "deep" neural networks.
There are few reasons to use activation function, the most common one is when the output needs to be within certain range by its nature. e.g. if the output is a probability, which is only valid in range [0, 1].
Closed. This question does not meet Stack Overflow guidelines. It is not currently accepting answers.
Questions asking us to recommend or find a book, tool, software library, tutorial or other off-site resource are off-topic for Stack Overflow as they tend to attract opinionated answers and spam. Instead, describe the problem and what has been done so far to solve it.
Closed 8 years ago.
Improve this question
Is there any working tridiagonal solver for non base 2 matrix (for example matrix as such: 500X500)?
The algorithm at https://code.google.com/p/tridiagonalsolvers/source/browse/#svn%2Ftrunk%2Ftridiagonalsolvers deals only with base 2.
Is there any difficultly in implementing such solver for non base 2 matrices?
The cuSparse library can sovle tridiagonal systems of arbitrary length.
The two functions you want to look at are:
cusparsegtsv(): http://docs.nvidia.com/cuda/cusparse/#cusparse-lt-t-gt-gtsv
cusparsegtsv_nopivot(): http://docs.nvidia.com/cuda/cusparse/#cusparse-lt-t-gt-gtsv_nopivot
There is also a batched tridiagonal solver.
You'll get best performance when the matrix side-length is a power-of-two, but you may not lose a lot of performance for non power-of-two matrices, especially if they're slightly below a power of two.