Where communities thrive

  • Join over 1.5M+ people
  • Join over 100K+ communities
  • Free without limits
  • Create your own community
Repo info
    Wei Wu
    Al Vinjamur
    Am trying to use MXNET with GPU on Windows 10. Following directions on github does not work and I have errors at the very first step of installing the C++ compiler (2013 version) as indicated on github. Is there an easier way to accomplish all this??
    Xiaolin Zhang
    Hi, mxnet hackers. I am trying to load a model from my c++ program. But, the model seems contains a "Custom Operator" which is written in python. So, what is the proper way to deal with it?
    Zheng Xu
    @leoncamel oops. Could you write a C++ equivalent of the python Custom Operator?

    I am following the R instructions for using the pre-trained Inception model. (http://mxnet.io/tutorials/r/classifyRealImageWithPretrainedModel.html)

    However, whatever picture I send in, it seems to always send back "torch" or "school bus". Also, most scores are exactly zero. Anyone seen this before?

    Joeri Hermans
    Hi all, I am implementing an optimizer for a distributed KV store. However, the last step requires me to push some weights to the distributed KV_store. Anyone has some pointers to do this, or where the documentation describes this?
    Joeri Hermans
    To recap, from what I understand if you update the weight in the optimizer, e.g., weight[:] -= lr * grad then you update the weight directly in the KV store right?
    Shiv Gowda
    Is there a problem with travis-ci setup or tests? I see some of the PRs failed these tests but are still merged. Note I am new to this project and I am trying to understand the ground rules. -Thanks.
    Qingsong Liu
    Hi, are there some tutorials for distributed training with kubernetes
    Are there examples of a vae coded up in the julia interface to mxnet?
    Hi, I'm a newbie who is looking through the API and trying to figure out how to learn more about what all the options do. For example, in Convolution, there is no explanation of what dilate does or pad (although I found that term in Lasagne docs). Any suggestions on what to read? It seems there are many things that get added from separate papers and some list sources (like ADAM) and others don't.
    Eren Golge
    Guys anybody help me to compile mxnet with NNPACK support ?
    I compiled NNPACK
    but what files should I move to where ?
    Hi Every one
    Who is the industry that come up with MXNet
    Xin Li
    Hi, I found that version 0.9.1 removes Optimizer in c_api (MXOptimizer** functions), how do I get optimizer by c api now?
    a little suggestion about config.mk: specify export CC = gcc-4.9 rather than export CC = gcc system like ubuntu16.04 already used 5.4.0 and may failed the compilation process
    Spencer Lyon
    I have a couple simple (I think ) questions about the julia bindings. Is the the right place to ask?

    My question has no relation with this topic. I want to know the reason why does people use ::testing::initGoogleTest instead of testing::initGoogleTest when people use gtest to do unit test. => from this, I am confused by the difference of ::testing::initGoogleTest and testing::initGoogleTest? Sorry to take up u guys time here in such stupid question.

    Anyone can help me for this confusion?Thanks.

    @ultradian just search the keyword , for example dilate convolution, then you will find the paper, I am also reading the source now, takes some time, but worth it
    I am trying oinstall mxnet wit gpu support on windows to be used from julia ... so far it fails, can anyone here help in any way?
    Benoît Quartier
    Do you guys know if there is any way of reducing the precision of the types in mxnet? Right now the documentation says the NDArrays only supports mx_float, is that float32? And if you wanna reduce it is it possible to still use mxnet?
    Benoît Quartier
    identify vzco7dRDirsr
    Ashwin Jeyaseelan
    Hello I'm having trouble installing MXNet
    I cloned the repository and followed the instructions for quick install. But when I run my program from desktop, I get the error: Traceback (most recent call last): File "neuralnet.py", line 1, in <module> import mxnet as mx ImportError: No module named mxnet
    I have macOS
    Ashwin Jeyaseelan
    Any ideas what the error is? When I do the install, I get a brew update error, which fails to install openCV, which I can manually install. But I'm not sure why it isn't recognizing the mxnet module...
    Ashwin Jeyaseelan
    nvr mind submitted it as an issue and it got fixed
    Is there any easy way to show the outputs of internal layers in mxnet/julia?
    Hello, I'm having trouble using the library libmxnet_predict.so in android.
    The error is 'cannot locate the symbol "cblas_sgemm"'. How can I solve this problem?
    German Alfaro
    anyone experience using MXNet with Haskell?
    Chris Kennedy
    anyone have tips on building mxnet with gpu support on ubuntu 16.04? the gcc < 6 requirement is making it very tough. I don’t know enough about ubuntu to downgrade gcc to 5.3 and still be able to install the corresponding old version of libatlas, or have gcc-5 installed alongside gcc-6
    Chris Kennedy
    nevermind, finally got it to work - had to use update-alternatives to have gcc-5 beside gcc-6
    Ashwin Jeyaseelan
    in the python api, whats the difference between sym and symbol?
    Jiajia Geng
    sym ==symbol
    I think, there are many abbreviation for short, like mx.module==mx.mod, mx.monitor==mx.mon,mx.ndarray==mx.nd,mx.visualization==mx.vis .etc
    Chih-Wei (Bert) Chang
    is anyone familiar with seq2seq? Is there any ready-to-use seq2seq model in MXNET?
    Huiting Liu
    Does mxnet have something like feed_previous argument for seq2seq in Tensorflow? If setting feed_previous to False. This means that the decoder will use decoder_inputs tensors as provided. If setting feed_previous to True, the decoder would only use the first element of decoder_inputs.
    I have looked at several examples in MXNet. Many of them only have the training part and leave the predicting part empty because there is not real data for the decoder in the predicting process. So after training the model, users don't know how to use it.
    Ashwin Jeyaseelan
    I'm having trouble finding the number of epochs I need in my model fit method, could someone check my error?
    Ben Hutchison

    Hi all, Im new to mxnet.

    Im trying to understand the autoencoder example. There is no documentation.

    While I have got it running on a p2.xlarge using the AWS deep learning AMI, there are many things I dont understand, eg

    1. How many layers and how many hidden units does it use?
    2. How can I interpret or visualize the results?
    Is there any example about Visualizing and understanding convolutional networks (Zeiler, 2014)?
    when i run "python demo.py" after install rcnn ,i get the error: Traceback (most recent call last):
    File "demo.py", line 141, in <module>
    File "demo.py", line 135, in main
    symbol = get_vgg_test(num_classes=config.NUM_CLASSES, num_anchors=config.NUM_ANCHORS)
    File "/home/zhixian/new/software/mxnet/example/rcnn/rcnn/symbol/symbol_vgg.py", line 276, in get_vgg_test
    rois = mx.symbol.Proposal(
    AttributeError: 'module' object has no attribute 'Proposal'
    how can i solve it
    Ashwin Jeyaseelan
    does anyone know how to plot the AUC score?