Where communities thrive

  • Join over 1.5M+ people
  • Join over 100K+ communities
  • Free without limits
  • Create your own community
Repo info
    Rabindra Nath Nandi
    while work with MNIST data
    steps=2000, monitors=[validation_monitor])
    Vikas Raunak
    Hi, I was trying the seq2seq tutorial of tensorflow. It has 20 GB dataset. My VM on Google cloud compute is throwing me out of memory error while unpacking the data file. I have 8vCPUs, 10 GB instance. But that's ram, and I have 30GB disk. How do you guys do experiments with such data? Using AWS EMR ?
    Rishab Goel
    Could some tell me a good resource to read about the details, types, training algorithms etc in depth about a Recurrent Neural Network?
    Vikas Raunak
    @RishabGoel hi,
    Vikas Raunak
    Ilya sutskever's PhD thesis is in fact titled Training RNNs.
    Rishab Goel
    @vyraun could you post a link?
    Thomas Mundt
    @RishabGoel Did you try to read the tutorial on the official tensorflow website? There are also links on it referring to further readings.
    Rishab Goel
    @tmundt Thanks for the pointer. I saw that but it didnt give me a very good intuition or mathematical justification as to why it is doing the things as it does?
    @vyraun thanks got the link on web
    I think a project might be good way to truly understand! is someone interested in doing one with me?
    Thomas Mundt
    Sounds good.
    @vyraun Did you solve the problem with the big data? This example seems to be too huge to be processed by normal matters. Is not there out another one not that big?
    Vikas Raunak
    @tmundt I haven't tried again after upgrading my VM. But I guess, other nmt tutorials must be there too.
    Ancas Horia
    hey, I have a problem when fitting my model using tflearn. It says: IndexError: list index out of range
    What shape should X and Y have for this to work ?
    Rohith Radhakrishnan
    Hi guys, does any of you have a pretrained model (convolutional network). I am trying to visualize the hidden layers.
    Danijar Hafner
    I don't, but you can easily run the code from one of the official TF tutorials on conv nets.
    Hi Guys,
    do you know any data set which contains flower images,
    i have created a custom set using google images but getting the following error when i run "tensorflow/tensorflow/examples/image_retraining/retrain.py" CRITICAL:tensorflow:Label inception has no images in the category training.
    @chotasanjiv : it 's here
    Hey, does anyone know any good rooms to learn the math behind neural networks. Thanks in advance
    Onuray Sahin

    Hey, is anyone working on im2txt model? What was the final training loss at the end of the training process (for 1M epoch)?

    For me it is iterating 61763 and training loss is 2.82 It was also 2.9x for the 35000th iteration. I would love to hear yours to see the ~difference.

    Thanks @arita37
    Richard L. Burton III
    Leonard Lee
    Hi everone, I want to know how to classify my own images using tensorflow. Any idea?
    Xiaolin Zhang
    Hi, tf hackers. Is there any method that I can convert DT_FLOAT to DT_HALF/DF_BFLOAT16 in my graph? I am thinking of accerlate the performance for Forward path only.
    Hello, I've proposed a project for chromosomes segmentation : https://github.com/jeanpat/DeepFISH . Up to now there are datasets (82146 images+ground-truth labels) and python notebooks (way to generate datasets, scikit-image based). The notebooks could be better: the size of the datasets are limited by the amount of RAM. The datasets are waiting to train a classifier.
    Xiaolin Zhang
    @jeanpat what kind of GPU card did you use?

    A small nvdia gpu: lspci -k | grep -A 2 -E "(VGA|3D)"
    00:02.0 VGA compatible controller: Intel Corporation 4th Gen Core Processor Integrated Graphics Controller (rev 06)
    Subsystem: CLEVO/KAPOK Computer Device 5000

        Kernel driver in use: i915

    01:00.0 3D controller: NVIDIA Corporation GK208M [GeForce GT 740M] (rev ff)
    Kernel modules: nouveau, nvidia_drm, nvidia
    03:00.0 Network controller: Intel Corporation Wireless 7260 (rev 73)

    @leoncamel up to now, I don't know if I can train a cnn with the proposed data.
    Jeff James
    @onurays - I am training as we speak, at global step 275k
    I'm hoping at step 1,000,000 it will do reasonably good captioning - currently running at 1.6 global steps per second on an ec2 p2.xlarge single GPU machine (90 cents / hr on demand)
    Jeff James
    Did yours finish? Let me know how it's working
    hi tensorflow user from Japan.
    Shikhar Srivastava
    Hello !
    I am facing an issue with Deepdream using an inception-v3 net that was retrained on my own image dataset.
    The graph for the retrained inception-v3-net is quite different. And subsequently, I'm finding altering the Deepdream notebook (https://github.com/tensorflow/tensorflow/blob/master/tensorflow/examples/tutorials/deepdream/deepdream.ipynb), quite difficult.
    Is it possible to use the retrained model forDeepdream, in the first place? :O
    And if so, could you suggest some changes to the deepdream notebook .
    Any help would be appreciated !
    Hi,this is newbie of Tensorflow
    Paweł Biernat
    Hi, I'm just starting with the tensorflow and I have a question. Say I have a vector x=[1,2,3] with a shape [3], how do I produce a matrix with columns coming from x, something like [x,x,x,x], with a shape [3,4]? Mathematically it could be described as the outer product of [1,2,3] and [1,1,1,1].
    Paweł Biernat
    If that helps, I'm trying to run Edward with w = Normal(mu=tf.zeros([D, K]), sigma=sigma0 * tf.ones([D, K])), and I would like to change tf.ones([D,K]) into the outer product matrix that I described.
    Paweł Biernat

    answering my own question


    seems to do the job.