Questions tagged [cntk]

Microsoft Cognitive Toolkit (CNTK) is an open source library for commercial-grade distributed deep learning. It became open source in January 2016. In March, 2019, Microsoft announced that version 2.7 was to be the last release of CNTK.

Microsoft Cognitive Toolkit (CNTK) is an open source library for commercial-grade distributed deep learning. It describes neural networks as a series of computational steps via a directed graph. In this directed graph, leaf nodes represent input values or network parameters, while other nodes represent matrix operations upon their inputs. CNTK allows to easily realize and combine popular model types such as feed-forward DNNs, convolutional nets (CNNs), and recurrent networks (RNNs/LSTMs).

CNTK implements stochastic gradient descent (SGD, error backpropagation) learning with automatic differentiation and parallelization across multiple GPUs and servers. CNTK has been available under an open-source license since April 2015. It is our hope that the community will take advantage of CNTK to share ideas more quickly through the exchange of open source working code.

In March 2019, Microsoft announced that version 2.7 was to be the last release of CNTK.

More information

Blogs:

503 questions
0
votes
4 answers

Installation fails for CNTK v.2.0 Beta 8 Release when VS2015 is already installed

I have VS2015 already installed on my machine, and when I try to install CNTK v.2.0 Beta 8 Release (Script-driven installation) it spawns VS installation window which immediately gets closed and I get next output: The following operations will be…
MarkoR
  • 543
  • 4
  • 12
0
votes
1 answer

CNTKLibrary-2.0 "System.AccessViolationException: Attempted to read or write protected memory. "

I am following the CNTKLibrary-2.0 example to write a simple case to check minibatch version of evaluation. The code looks like in the following way. But I got the access violation exception when I call "Evaluate" function. Any insights about such…
0
votes
3 answers

CNTK.GPU Managed C# Eval API: System.Runtime.InteropServices.SEHException when calling outputVal.CopyVariableValueTo

I have trained a model that labels a sequence by grabbing the label of the last word in the sequence: Sequential([ Embedding(emb_dim), cntk.ops.sequence.last(Recurrence(LSTM(hidden_dim), go_backwards=False)), …
malex
  • 99
  • 6
0
votes
1 answer

Incresaing the image input resolution for a pretrained Resnet model in CNTK

I am using CNTK's model edit language to load a pre-trained Resnet model, add a new last layer, and refine the model on some new dataset. I would also like to change the network architecture to accept higher resolution images as input (which is…
0
votes
1 answer

RuntimeError: Node '__v2libuid__Plus502__v2libname__Plus225' (Plus operation): DataFor: FrameRange's dynamic axis is inconsistent with matrix

the same training code that works on one machine gets the following error on a different machine. The first machine is a SurfaceBook, the other is my desktop with a GTX 1080. Any idea what's going wrong here? The error is: Traceback (most recent…
0
votes
1 answer

How can I get the feature vector (the N-1 layer's output before softmax) of input image from deep model?

How can I get the feature vector (the N-1 layer's output before softmax) of an input image from deep model? I am using Microsoft's CNTK deep learning toolkit.
Avi Thaker
  • 455
  • 3
  • 10
0
votes
1 answer

How to implement K-max pooling in cntk?

Has anyone implemented K-Max pooling in CNTK? It is a max-pooling layer that, rather than selecting the top one out of some number of units, selects the k largest outputs from all the units in the previous layer.
0
votes
1 answer

How to implement wide one dimensional convolution in cntk

I'd like to implement the wide type of one dimensional convolution (https://arxiv.org/pdf/1404.2188v1.pdf) in CNTK. Is there a built in method for that or how should I play with the parameters of Convolution() to implement it? Thanks!
0
votes
1 answer

Improve error message for ill-formed input format?

I have a map file containing data like this: |labels 0 0 1 0 0 0 |features 0 |labels 1 0 0 0 0 0 |features 2 |labels 0 0 0 1 0 0 |features 3 |labels 0 0 0 0 0 1 |features 7 Data is read into a minibatch with the following code: from cntk import…
OlavT
  • 2,496
  • 4
  • 31
  • 56
0
votes
1 answer

ImportError: cannot import name 'cntk_device' when using VSCode

I followed the instructions from this link to use VSCode for cntk. But got this error: Traceback (most recent call last): File "c:\DebugLabs\VSCodeProjects\TestPython\experiment_1.py", line 6, in from cntk import Trainer, cntk_device,…
0
votes
3 answers

Hardware requirement for installing cntk

Are there any recommended or minimum system requirements for Microsoft Cognitive Network Toolkit? I cannot find this information anywhere on the git.
cho
  • 1
  • 2
0
votes
1 answer

CNTK Progress Printer

Can we better document what the different parameters for Progress Printer's constructor mean? https://cntk.ai/pythondocs/cntk.utils.html#module-cntk.utils.progress_print
0
votes
2 answers

CNTK Distributed Crash - Beta 7

I am running a variation of the CIFAR 10 distributed to utilize my data. I get the following error: Traceback (most recent call last): File "CNTK_Train.py", line 158, in checkpoint_path =…
David Crook
  • 2,722
  • 3
  • 23
  • 49
0
votes
1 answer

Does ImageReader support Base64ImageDeserializer format in CNTK now?

I saw Base64ImageDeserializer.cpp was added into CNTK github. https://github.com/Microsoft/CNTK/tree/00bcf2cbac79fec3597a1ce0ce8445ebdc823ae3/Source/Readers/ImageReader So I want to ask when base64 encode image as imagereader input will be…
0
votes
1 answer

In CNTK implementation of ADAM optimizer, how the parameters alpha, beta1, beta2 and epsilon relate to learning rate and momentum

I am using the adam_sgd optimiser to train a neural network and I am having trouble associating the arguments in the function with the parameters reported in the paper for Adam. More specifically how do the parameters alpha, beta1, beta2 and epsilon…
Sayan Pathak
  • 870
  • 4
  • 7