Where communities thrive

  • Join over 1.5M+ people
  • Join over 100K+ communities
  • Free without limits
  • Create your own community
Repo info
    so I got as far as installing tensor globally even though I had the virtualenv active :/
    problem was I was using the latest ubuntu 16 but should have been using ubuntu 14
    64bit system of course
    but I'm having another problem :/
    numpy won't install and its asking for setuptools which is already installed
    Cleaning up...
    Command /usr/bin/python -c "import setuptools, tokenize;__file__='/tmp/pip_build_root/numpy/setup.py';exec(compile(getattr(tokenize, 'open', open)(__file__).read().replace('\r\n', '\n'), __file__, 'exec'))" install --record /tmp/pip-5hYSAi-record/install-record.txt --single-version-externally-managed --compile failed with error code 1 in /tmp/pip_build_root/numpy
    Traceback (most recent call last):
      File "/usr/bin/pip", line 9, in <module>
        load_entry_point('pip==1.5.4', 'console_scripts', 'pip')()
      File "/usr/lib/python2.7/dist-packages/pip/__init__.py", line 235, in main
        return command.main(cmd_args)
      File "/usr/lib/python2.7/dist-packages/pip/basecommand.py", line 161, in main
        text = '\n'.join(complete_log)
    UnicodeDecodeError: 'ascii' codec can't decode byte 0xe2 in position 72: ordinal not in range(128)
    Hi guys
    I am in need of few ideas that can be implemented in deep learning using Tensor. Any help will be appreciated
    Does anyone know of examples on how to integrate tensorflow into java applications. From what I have read, it looks like a tightly couple application would require jep (due to the underlying C++ implementation). The alternative would be a separate service that communicates over http, which would be slower. Any ideas or source referrals?
    Hi, one basic question for a newbye, please
    I am training a network and it works great, but after some time where it is stable ate the maximum performance, then it gets crazy.
    I want to store the best network parameters in a new variable, i.e.,
    if current_performance > best_performance:
    best_params = current_params
    how can I copy the parameters?
    solved! :-)

    oh! nope... I have created another network:
    best_inputs, best_out, best_scaled_out = create_network()
    best_network_params = tf.trainable_variables()

    Then I create this op:
    update_best_network_params = [best_network_params[i].assign(original_network_params[i])
    for i in range(len(original_network_params))]

    Later in my program I invoke:

    And the problem is that it seems that it is assigning a reference rather than doing a deep copy

    Learning a new library/language by imitation, I feel like a monkey writing a book by random keys! :-D
    I am struggling for duplicating a network when some condition is met (e.g., performance > threshold) in order to get a snapshot. This way, even if the training gets worse, I can recover an intermediate snapshot of the network. Please, any idea how to achieve this?
    Hello , I'm using tensorflow as backend in keras and after upgrading to 0.12rc1 I get the following error : ValueError: The shape for while_1/Merge_2:0 is not an invariant for the loop. It enters the loop with shape (1, 1), but has shape (?, 1) after one iteration. Provide shape invariants using either the shape_invariants argument of tf.while_loop or set_shape() on the loop variables. I guess the issue is on keras itself, but I would like to revert to 0.10.0 , do u guys know if it can be done with pip . I m on macOS
    upgrading to the latest keras 1.1.2 from github fixed the issue :)
    Hey guys!
    Is it possible to store the log-files of tensorboard in another file than /tmp ? It doesn't seem to work. My problem is that I want to train a model today, but visualize the results after reboot on another day. Any suggestions?
    Okay, I solved it. Instead of ~/... one has to save the log-file as /home/$USER....
    Tiago Rodrigues
    hey guys
    imagine I want to train data for semantic similarity. I have stored the features for each sentence individually but all my features are dependent of the pair of sentences selected (eg I may have number of nouns in sentence A is X and number of nouns in sentence B is Y but the feature I want to train simmilarity is X-Y) do I need to do precalc for each sentece pair or is another way to make the features appear on the fly.
    Surya Bhupatiraju
    Hi guys, do people know how to extract the attention matrices that are computed in TensorFlow's seq2seq example code during an example decoding?
    chen wei
    Hi guys, how to implement weight normalization in tensorflow?

    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.

    Prakhar Mishra
    I am interested in learning Time Series data. Right now i have created a NN that, takes in input a set of days(window) and predicts the next days output. How can I take it further in predicting next 7 days for eg. ?
    anyone here ?
    need help loading my weights and variables
    Rizky Luthfianto
    Anyone has easy an explanation of Beam Search? I've googled it and opened all it links, but still doesn't understand it
    Nicholas Connor
    any examples floating around of multi-label classification with the contrib estimator APIs?
    Nicholas Connor
    actually I don't see many examples of multi-label period. that is surprising given all the image tagging hype lately
    Segev Malool
    Hello everyone, I have a question about feeding a scalar to a tf model, as in nExamples = tf.placeholder(tf.int32,shape=[])
    When I try to use nExamples to define a tf.Variable, I get an error:
    initial_value must have a shape specified: Tensor("random_uniform_1:0", shape=(?, 10), dtype=float32)
    It thinks the shape is '?', but I clearly defined the shape in the placeholder call above.
    What gives?
    Simon Ho
    Anyone can explain how tf.contrib.training.bucket_by_sequence_length works?
    If interested, there's a semantic segmentation problem waiting to be solved here : https://github.com/chromosome-seg/DeepFISH
    What is a good way to find out if tensorflow is supported in certain hardware (mac AMD openCL)
    Darren Garvey
    @ItchyJunk Checking out the sources and running ./configure is a pretty good way. In a general sense, mac AMD and OpenCL are all supported
    pip install tensorflow-gpu and running python -c "import tensorflow as tf; tf.InteractiveSession()" should also tell you if TF finds your GPU, if that's what you are interested in
    Shivaji Dutta
    Just do git clone and build it, that is the best way
    It will do other platform optimizations as needed.
    Any tips on using SequenceExample where each item in the sequence has an associated vector. ex. token sequence where each token has a char list.
    Ming Kim
    Hi, I'm trying to search various ways to monitor tensorflow weight tensor.
    I know we can watch these variable tensors through Session.run(), tf.Print(), tf.py_func()and tools like tensorboard, tdb, tfdbg
    But is it impossible to use IDE(like Pycharm) for this?
    I tried by myself, and couldn't find some places to set a breakpoint.
    Please tell me if you succeed tensor debugging using IDE. Thank you!
    Siddharth Jain
    I was coding a new op to be added onto the library but i compiler encountered an error when i accessed the element in the ith index of an object of tensor class using parenthesis
    i.e. object(i)
    which says:
    error: 'object' can't be used as a function.
    Please tell me what should i do? Thanx!
    Andre Pemmelaar

    I have a question regarding the use of tf.nn.dynamic_rnn.

    I have a numpy array of size x_shape = (50, 30, 10),
    where the
    batch size = 50,
    max length of series (max_time) = 30
    input vector of length = 10.

    I'm getting an error of TypeError: 'Tensor' object is not iterable.

    According to the documatation:
    If time_major == False (default), this must be a Tensor of shape: [batch_size, max_time, ...], or a nested tuple of such elements.

    How should I format my input if not an array of rank three e.g. [50, 30, 10]? Perhaps a list of 30 elements each of which each element is a vector of length 10?