Réponse apportée
Tutorial / classes / training for developing asynchronous CUDA and MEX code?
Make all the variables involved gpuArray objects and those lines of code will run as asynchronously as the GPU allows. This mean...

plus de 2 ans il y a | 0

| A accepté

Réponse apportée
Error with averagePooling1dLayer (pooling dimension lower than pooling size)
The time dimension is variable so just because you happen to be passing a long sequence into the network doesn't mean you're not...

plus de 2 ans il y a | 0

| A accepté

Réponse apportée
pixelLabelImageDatastore not partitionable despite underlying imageDatastore and pixelLabelDatastore are
For parallel training of a semantic segmentation network, you're supposed to use combine to join the two input datastores rather...

plus de 2 ans il y a | 0

| A accepté

Réponse apportée
Error in helperModClassTrainingOptions (line 29) 'CheckpointPath',checkpointPath,...
You need to pass the checkpointPath variable to your function.

plus de 2 ans il y a | 0

| A accepté

Réponse apportée
How to increase gpu usage?
To get GPU utilization high you need to give it a lot of high resolution work to do. Often RL problems, especially the ones in o...

plus de 2 ans il y a | 1

| A accepté

Réponse apportée
Increasing effective GPU memory through parallel GPUs?
The only way to do this is to divide your computation up manually and spread the work across your GPUs by using a parallel pool ...

plus de 2 ans il y a | 0

Réponse apportée
Parallel CPU computing for recurrent Neural Networks (LSTMs)
That doc page is about shallow networks (using train) rather than deep networks (using trainNetwork). Parallel training in train...

plus de 2 ans il y a | 1

Réponse apportée
CNN-LSTM validation data underperforming compared to training data
I don't know why your model seems to be overfitting, but I can confirm that your validation data is being run through the exact ...

plus de 2 ans il y a | 1

Réponse apportée
Error using pagefun with mldivide
This is a known bug in cublas introduced in CUDA 11 and fixed in the next MATLAB release. The workaround, as you correctly disco...

plus de 2 ans il y a | 0

| A accepté

Réponse apportée
Is it possible to use variable length arrays with SequenceInputLayer in a custom training loop, i.e. as a dlarray?
dlnetwork objects do not take your entire dataset as input, they expect to receive a single batch at a time. You need to loop ov...

plus de 2 ans il y a | 0

| A accepté

Réponse apportée
Can I use my CPU and GPU simultaneously for Numerical Simulation?
canUseGPU is the easiest way to write code that diverges based on whether a supported GPU is available. A good design should nee...

plus de 2 ans il y a | 2

Réponse apportée
Force CPU Parallel pool
You should use 'ExecutionEnvironment','cpu' for training on your local machine. This is multithreaded and will use all your core...

plus de 2 ans il y a | 2

Réponse apportée
Increrasing the validation accuracy
There's no way to do this automatically, but in your case just using the default 'OutputNetwork' setting ( last-iteration ) woul...

plus de 2 ans il y a | 0

Réponse apportée
gpu limit on 3070 with a simple CNN
You have 2-D data so you can't use a resize3dLayer. Use a resize2dLayer. But actually it doesn't look like that's what you want....

plus de 2 ans il y a | 0

| A accepté

Réponse apportée
Indexing after knnsearch with GPU is slow
wait just asks the GPU to finish executing any pending operations, in this case, the call to knnsearch. Your previous timing cod...

plus de 2 ans il y a | 0

| A accepté

Réponse apportée
How to fix the GPU warning?
It's all pretty clear from the documentation. Ampere cards are not supported in R2019a. You are using forward compatibility, whi...

plus de 2 ans il y a | 0

| A accepté

Réponse apportée
ExecutionEnviroment problem on optimizer at traininOptions
It looks like you typed ExecutionEnviroment instead of ExecutionEnvironment (ie. you missed the second 'n').

plus de 2 ans il y a | 0

Réponse apportée
augmentedImageDatastore, understanding the flow of how it works
1) Yes 2) What is that 3rd dimension? Is this an index image? I think augmentedImageDatastore only understands grayscale and RG...

plus de 2 ans il y a | 1

| A accepté

Réponse apportée
What is different between AvailableMemory and FreeMemory?
Both are valid and supported, but FreeMemory was replaced many years ago by AvailableMemory as the documented way to determine h...

plus de 2 ans il y a | 0

Réponse apportée
Loadlibrary C library (dll) changing Matrix in place?
MATLAB is being exception safe and preserving your workspace by not silently modifying variables as a side effect. The only cert...

plus de 2 ans il y a | 0

| A accepté

Réponse apportée
gpu memory cannot be released
This memory contains device binaries, i.e. program code, as well as workspace set aside for the CUDA libraries and not available...

plus de 2 ans il y a | 0

| A accepté

Réponse apportée
sparse in cuda matlab shows bad performance
It is often slower. The problem is that sparse factorizations create dense matrices...basically, it's hard to parallelize. We...

plus de 2 ans il y a | 0

| A accepté

Réponse apportée
Using multi-gpu in minibatchqueue
https://uk.mathworks.com/help/deeplearning/ug/train-network-in-parallel-with-custom-training-loop.html will show you how to use ...

presque 3 ans il y a | 0

Réponse apportée
How to use multiple gpu on CGAN?
SPMD wraps the whole epoch loop. After you open it you partition the data and then start looping. See https://uk.mathworks.com/h...

presque 3 ans il y a | 0

Réponse apportée
Memory increases on GPU while performing modification inplace
MATLAB cannot perform this operation in place because |data| is a workspace variable. MATLAB has no way of knowing there won't b...

presque 3 ans il y a | 0

Réponse apportée
Frame per second computingfor GPU code generation
This is a smoothing filter to eliminate noise, to ensure it gives a good result from the first frame you could set it to |1/newt...

presque 3 ans il y a | 0

| A accepté

Réponse apportée
Error Message: Brace indexing is not supported for variables of this type. during TrainNetwork
You can definitely use your GPU on multiple MATLAB workers in an SPMD block. You should show your code, looks like you have a sy...

presque 3 ans il y a | 0

Réponse apportée
Doesn't the GPU Coder support singleton expansion of ARRAYFUN?
Support for dimension expansion is actually restricted to the GPU version of arrayfun, and GPU Coder supports the CPU version's ...

presque 3 ans il y a | 0

| A accepté

Réponse apportée
From CPU code to GPU
You need to vectorize your code. The GPU is not intended for performing this kind of looping series of operations on scalar vari...

presque 3 ans il y a | 0

Réponse apportée
I have a problem about " out of memory" for running VGG16 for image classification
It doesn't look as though you're setting the MiniBatchSize on the call to activations which means it's using the default value o...

presque 3 ans il y a | 0