repository
stringclasses
156 values
issue title
stringlengths
1
1.01k
labels
stringclasses
8 values
body
stringlengths
1
270k
tensorflowtensorflow
broken link mac gpu py3 whl file
Bug
on page the url of the tensorflow python package the url of the tensorflow python package this link be break
tensorflowtensorflow
invalid link in
Bug
seem that the cifar 10 multi gpu trainer link in be break actually this code only exist r0 12 there be three way to fix it 1 use an old version such as cifar 10 multi gpu trainer 2 remove this link in the doc page 3 create a new cifar10 multi gpu train py file for r1 1 I can w...
tensorflowtensorflow
several link error in readme md
Bug
system information os window 10 enterprise x64 browser chrome 58 x64 location china issue the instal tensorflow link from readme md fail with error error not find the request url version r0 12 install index html be not find on this server ...
tensorflowtensorflow
memory leak from deep learning training step finalize graph
Bug
system information have I write custom code as oppose to use a stock example script provide in tensorflow yes although my code be somewhat base on the mnist deep learn tutorial os platform and distribution e g linux ubuntu 16 04 linux ubuntu 14 04 versi...
tensorflowtensorflow
documentation page be unnecessarily large
Bug
curl wc c 1841844 1 8 mb each page in the documentation now contain a huge left navbar contribute over 99 of the size 1 this would waste a lot of network traffic on load identical navbar over and over again 2 it create a big trouble when I try to build an offline version of the d...
tensorflowtensorflow
tf pow x y will freeze for negative integer y
Bug
system information have I write custom code as oppose to use a stock example script provide in tensorflow yes os platform and distribution e g linux ubuntu 16 04 window 10 and macos 10 12 4 tensorflow instal from source or binary binary tensorflow version use com...
tensorflowtensorflow
ctc greedy decoder inconsistent with ctc beam search decoder
Bug
the following extract from the the ctc beam search decoder documentation seem to be mislead the ctc greedy decoder be a special case of the ctc beam search decoder with top path 1 and beam width 1 but that decoder be fast for this special case instead the follow result can be observe ...
tensorflowtensorflow
tensorflow process and core engine documentation
Bug
hi all I be interesting to understand more detail about tf core engine my focus interest on how a process create and launch when session be be create also how the thread work and how python api translate these thread to c tf engine be there any documentation about that sincerely
tensorflowtensorflow
tutorial update logging and monitor basic
Bug
please go to stack overflow for help and support if you open a github issue here be our policy 1 it must be a bug or a feature request 2 the form below must be fill out here s why we have that policy tensorflow developer respond to issue we want to focus on work that benefit the whol...
tensorflowtensorflow
statically link library in tf binary can cause symbol collision
Bug
tensorflow currently statically link all dependency this sometimes cause hard to diagnose crash e g segfault when another version of a dependency be load into the process this can even happen within tensorflow if separate tensorflow so s be load into the same python process possible solution would be t...
tensorflowtensorflow
documentation for tf nn ctc label argument be unclear
Bug
the documentation for the connectionist temporal classifier be unclear for label argument here be what exist currently label an int32 sparsetensor label indice I b t mean label value I store the i d for batch b time t label value I must take on value in 0 num label se...
tensorflowtensorflow
how do you generate tensorflow doc so you can confirm documentation fix you make
Bug
example suppose I see a format error in a tensorflow function s argument on the web I then make a change in the args section of the function s python comment how can I generate these html doc after make this fix
tensorflowtensorflow
check fail ndim dim 2 vs 1 when I build a svm model
Bug
when I build a svm model with tf learn it get error like this f tensorflow core framework tensor shape cc 36 check fail ndim dim 2 vs 1 ask for tensor of 2 dimension from a tensor of 1 dimension I have ask a question in stackoverflow it could be a issue accord the reply all the reprod...
tensorflowtensorflow
inconsistent result when tf sqrt be apply to tensor versus element wise
Bug
system information os platform and distribution e g linux ubuntu 16 04 fedora 24 tensorflow instal from source or binary pip exact command to reproduce import tensorflow as tf import numpy as np start 0 linear index tf range start 496 1 8 ...
tensorflowtensorflow
terminate call after throw an instance of std bad alloc not out of memory
Bug
hi I m use kera with tensorflow back end to train a lstm network I be do a grid search over the learning rate and dropout factor with the fix batch size of 64 it run perfectly but in the middle of it be interrupt by signal 6 sigabrt with the follow error terminate call after throw an instance of std...
tensorflowtensorflow
io no opkernel be register to support op less with these attrs
Bug
hi all I have try to load the model inside io that I generate from python and right now I have the follow problem error add graph to session no opkernel be register to support op less with these attrs register device cpu register kernel device cpu t in dt float no...
tensorflowtensorflow
link error for the deprecate init py
Bug
it seem that the link look here can not point to the right page in the doc of histogram summary should we just change it to please go to stack overflow for help and support if you open a github issue here be our policy 1 it must be a bug or a feature request 2 the form below mu...
tensorflowtensorflow
improve documentation for tensorarray
Bug
the current documentation for tensorarray could use an expand introduction explain its utility and main motivation example of actual usage and further explanation of some of its method the doc here do not have a single usage example the overall description of the section be just this class be mea...
tensorflowtensorflow
non deterministic behaviour of tf train batch in case the number of thread be high than 1
Bug
this be relate to the stackoverflow question 43613376 the thread owner create a batch with the follow code batchedinput tf train batch input batch size 64 num thread 8 capacity 500 3 64 and he notice that create batch be not in every run the same they be quite si...
tensorflowtensorflow
small error in dynamicattentionwrapper
Bug
system information have I write custom code as oppose to use a stock example script provide in tensorflow not relevant os platform and distribution e g linux ubuntu 16 04 not relevant tensorflow instal from source or binary not relevant tensorflow vers...
tensorflowtensorflow
cifar10 tutorial link to example code 404 s
Bug
follow any of the link to the code for the cifar10 tutorial 404 s example link code organization any from above
tensorflowtensorflow
mac gpu link not work
Bug
hello I be try to install tensorflow for mac gpu python 3 one from github and I find that it be a broken url give I an http 404 error the url be please fix the url good daksh
tensorflowtensorflow
tensorflow python client device lib list local device bug
Bug
I be try to set up gpu configuration for tensorflow the step be very simple call tensorflow python client device lib list local device to detect the number of gpu device on the machine and then set config for tensorflow the follow be the code for reproduce from log import getlogger imp...
tensorflowtensorflow
enable fuse winograd by default
Bug
right now fuse winograd be disable by default l1970 l1976 this be even though enable this speed up model considerable in the 3x3 case see issuecomment 295779638 what remain as far as issue etc to get this fast conv enable by default cc yangzihao tfboyd
tensorflowtensorflow
tflearn incorrect comment
Bug
in the tflearn quick start guide here below the section describe the training input pipeline train input the first snippet python define the test input def get train input x tf constant training set datum y tf constant training set target return x y sh...
tensorflowtensorflow
tensorflow 60 80 slow than pytorch on train wide resnet
Bug
cc tfboyd from issuecomment 295502315 on an aw p2 xlarge use the tensorflow tensorflow 1 0 1 devel gpu docker image as a base I see 270 ms per epoch while train a wrn 16 4 without dropout on cifar 10 use a pytorch implementation from I see instead 150 ms per epoch for the same my imp...
tensorflowtensorflow
explain what tf nn softplus do to integer
Bug
tf nn softplus compute log 1 exp x naively I wouldn t expect this to work for integer but it do on integer it seem to degenerate to a poorly name version of tf relu it compute max 0 x we probably can t eliminate the integer version for backward compatibility reason but we sh...
tensorflowtensorflow
tf get collection documentation argument description be confuse
Bug
from item without a name attribute be never return if a scope be supply and the choice or re match mean that a scope without special token filter by prefix what do this mean be there a typo here
tensorflowtensorflow
evaluable doc name checkpoint path and hook should be new bullet
Bug
see final param for evaluate be not bullet make read the documentation less readable I ve see this in other doc too can comment if I come across more of they I wonder if the parser have some bug that cause it to not bullet ify some param for function
tensorflowtensorflow
segmentation fault in tensorflow filesystemregistryimpl register
Bug
syntaxnet package be build tf package ver 1 01 be instal from repo by pip as dependency fortunately tf instal follow these instruction create the pip package and install work fine new lwp 31764 thread debugging use libthread db enable use host libthread db library lib x86 64 linux gnu libthread...
tensorflowtensorflow
tutorial have error recurrent neural network
Bug
tutorial url I m go through the tutorial list above and I think there be a mistake in the very first code example lstm tf contrib rnn basiclstmcell lstm size initial state of the lstm memory state tf zero batch size lstm state size an error be report for the third line val...
tensorflowtensorflow
tensorboard filter regresson
Bug
system information docker image tensorflow tensorflow nightly or 1 1 0rc2 describe the problem start a tensorboard process tensorboard logdir efs log atari and try and filter it do not have any effect image
tensorflowtensorflow
doc update wheel url to match late tensorflow release 1 1 0
Bug
the wheel available on the website be from the previous release 1 0 1 which still contain opkernel error that be solve at head as well as on 1 1 0 and can be ignore but be confuse user prompt repeat report of this issue
tensorflowtensorflow
tensorflow on armv7 seem to be slow
Bug
I have instal tensorflow on our custom board which be base on I mx6 armv7 processor I have compile the c module alone without python support and be able to run the tensorflow pi example label image command use make f tensorflow contrib makefile makefile host os pi target pi the problem be t...
tensorflowtensorflow
stack bidirectional dynamic rnn input incorrect documentation
Bug
hi this be really a documentation problem rather than problem with the actual code the doc state that input should be of shape number of numsequence x batchsize x inputsize but in reality it s batchsize x numsequence x inputsize
tensorflowtensorflow
segfault nan s in svd
Bug
I m get failure try to run svd on a particular matrix the result be either all nan s for u matrix or it s segfault like below to reproduce run this script in python3 I can t see anything special about this matrix beside the fact that it s badly condition ie I can perform svd on this matrix in math...
tensorflowtensorflow
readme md have a deprecate api call
Bug
system information tensorflow v1 0 0 65 g4763edf dirty 1 0 1 describe the problem clearly in readme md line prediction vgg vgg16 image be train true should be prediction vgg vgg 16 image be train true see source code for vgg
tensorflowtensorflow
the running speed of window be low than that in linux
Bug
please go to stack overflow for help and support if you open a github issue here be our policy 1 it must be a bug or feature request 2 the form below must be fill out here s why we have that policy tensorflow developer respond to issue we want to focus on work that benefit the whole commu...
tensorflowtensorflow
possible doc inconsistency tf contrib framework load variable
Bug
note issue that be not bug or feature request will be close please ask usage question on stackoverflow you must complete this information or else your issue will be close have I write custom code as oppose to use a stock example script provide in tensorflow tensorflow instal from source o...
tensorflowtensorflow
tf set random seed do not reset random op state
Bug
tf version 1 1 0rc1 instal from nightly apr 10 2017 1 03 am run on cpu python 2 import tensorflow as tf import numpy as np sess tf interactivesession sess run tf global variable initializer tf set random seed 1 a tf truncated normal initializer seed none 1 prin...
tensorflowtensorflow
hi I be unable to access the documentation
Bug
note issue that be not bug or feature request will be close please ask usage question on stackoverflow you must complete this information or else your issue will be close have I write custom code as oppose to use a stock example script provide in tensorflow tensorflow instal from source o...
tensorflowtensorflow
yolodetector example need java test rule
Bug
hi when I use android studio to compile tensorflowyolodetector the follow error occur e art no implementation find for long org tensorflow contrib android runstat allocate try java org tensorflow contrib android runstat allocate and java org tensorflow contrib android runstat allocate te...
tensorflowtensorflow
documentation link issue
Bug
note issue that be not bug or feature request will be close please ask usage question on stackoverflow under this link the webpage be not find you must complete this information or else your issue will be close have I write custom code as oppose to use a stock example script provide ...
tensorflowtensorflow
problem with wide n deep tutorial py on tensorflow 1 0
Bug
use python 3 6 0 anaconda x64 tensorflow 1 0 macos sierra version 10 12 4 I get the follow error python wide deep py training datum be download to var folders h2 727s56vx40s 6n2z9ldl6kx00000gs t tmp76m50o3h test datum be download to var folders h2 727s56vx40s 6n2z9ldl6kx00000gs t tmpw...
tensorflowtensorflow
cpp protobuf instruction out of date for macos
Bug
instruction to upgrade to cpp protobuf implementation on mac from protobuf pip package 31 don t work work make tf fail with follow stacktrace traceback most recent call last file kronecker benchmark py line 3 in import tensorflow as tf file user yaroslav anaconda envs mar1 lib ...
tensorflowtensorflow
documentation break link and image in documentation
Bug
the url have a break link label the section on segmentation also there be a broken image on that page as well same problem here too except not all have an image to be break the page I believe they should link back to be here segmentation I come her...
tensorflowtensorflow
bezel build problem for osx no toolchain correspond to local darwin find for cpu darwin
Bug
note issue that be not bug or feature request will be close please ask usage question on stackoverflow you must complete this information or else your issue will be close have I write custom code as oppose to use a stock example script provide in tensorflow no this be an installation issue ...
tensorflowtensorflow
biasgradop mistakenly put on cpu
Bug
note issue that be not bug or feature request will be close please ask usage question on stackoverflow you must complete this information or else your issue will be close have I write custom code as oppose to use a stock example script provide in tensorflow custom code tensorflow instal f...
tensorflowtensorflow
break link in readme md
Bug
in in addition to the type of scope mechanism in tensorflow name scope variable scope tf slim add a link of variable scope variable scope be break
tensorflowtensorflow
problem with exemple in api documentation for tf contrib distribution bijector scaleandshift
Bug
hello I be currently use tensorflow version 1 0 0 from conda vanilla installation and I be run a snippet code fromapi documentation I get an attributeerror which seem an accurate error since I can t seem to find scaleandshift on the github repo for tensorflow r1 0 ds transformeddistribu...
tensorflowtensorflow
output indifferent to input when apply merge duplicate node
Bug
I think I might have locate the root of 8698 it seem that merge duplicate nodes be the reason that quantize node malfunction whatever I do whenever I apply a merge duplicate nodes somewhere during a graph transformation the output become completely indifferent to the input unfortunately I...
tensorflowtensorflow
tensorboard not show datum on window
Bug
I have a simple example of a mnist classifier that work well for the dataset however tensorboard be unable to read the datum from the save log I have inspect the datum use tensorboard inspect logdir and the file seem to be in order I have test this and a much simple model on a different machine ...
tensorflowtensorflow
an update for the tf contrib learn quickstart example be need
Bug
hi all just try to start the script from here find one issue for python3 user python import urllib raw urllib urlopen iris training url read this return bash attributeerror ...
tensorflowtensorflow
why use unknown batch size in basicdecoder class
Bug
I find that in the source code python property def batch size self return self helper batch size def rnn output size self size self cell output size if self output layer be none return size else to use layer s compute output shape we nee...
tensorflowtensorflow
python stagingarea documentation miss
Bug
stagingarea and perhaps other structure in datum flow op py be not present in the python documentation they be however mention in the 1 0 0 and 1 1 0 release note what other attempt solution have you try a search of the doc for stagingarea doesn t reveal any of the docstring present in ...
tensorflowtensorflow
tf contrib seq2seq sequence loss documentation incorrect
Bug
the function return a tensor with rank 0 1 2 depend on its argument this be inconsistent with its docstring s description of the return l58 I will submit a pr soon to fix the doc note only file github issue for bug and feature request all other topic will be close for general support...
tensorflowtensorflow
tfdbg pyfunc be not find in partition graph
Bug
when use tensorflow debugger the program crash with the follow message traceback most recent call last file farn train py line 470 in train file farn train py line 448 in train loss value sess run train op loss file home qiqi anaconda2 li...
tensorflowtensorflow
incorrect reference for tf learn in linear model tutorial
Bug
this page and this mention tf learn multiple time however look at the link and source code I believe that tf learn should actually be either tf contrib or tf contrib learn not tf learn I guess tf learn be the old name but get rename but the tutorial be still outdate
tensorflowtensorflow
tensorflow hang during training while use with tf device device cpu 0
Bug
description I m try to fine tune an inception v1 model and my late implementation be base on train image classifier py I ve notice that sometimes tf hang might be similar to issue 2788 and I m try to figure out why the follow snippet be my load batch function I ve also use this sessi...
tensorflowtensorflow
tensorflow cause whole python to crash with error python frozen importlib bootstrap
Bug
if I try to run a computer vision script python crash with no error message at all if I run with python m trace python console go in loop with a message repeat continuously that be unreadable because line be print endless too fastly I can see some line relate to sre compile py text format py inspect ...
tensorflowtensorflow
dynamic rnn s docstre refer to a non exist rnn function
Bug
here l392 the docstring of the dynamic rnn function refer to a non exist rnn function I don t know if it get delete or move somewhere else
tensorflowtensorflow
freeze graph not initialize table
Bug
I be not sure if this be an actual bug or if its expected but undocumented behavior I have a model that use multiple lookup table create via string to index I freeze the model like so bazel bin tensorflow python tool freeze graph input graph tmp tf graph pbtxt input checkpoint tmp tf model ckpt 0...
tensorflowtensorflow
inconsistent behaviour between cpu and gpu gradient step operation
Bug
what related github issue or stackoverflow thread have you find by search the web for your problem none environment info operating system linux mint 17 2 rafaela instal version of cuda and cudnn please attach the output of ls l path to cuda lib libcud ls l usr local cuda lib...
tensorflowtensorflow
attributeerror nonetype object have no attribute tf newstatus
Bug
in the late source the issue be fix in the thensorflow python client session py but I now receive the follow error instead exception ignore in traceback most recent call last file home local lib python3 5 site package tensorflow python client session py line 595 in del attribu...
tensorflowtensorflow
please provide an example how to use a model train from scratch for image classification
Bug
the follow documentation of tensorflow slim contain how to train a model from scratch but it s not explain how to use the result checkpoint file for image classification train a model from scratch although it s possible to load the pre train model pre train model and use it for image classification with ...
tensorflowtensorflow
create variable in a while loop
Bug
I understand why this fail give the way control input be insert in non variable op create in while loop but maybe the limitation should be document somewhere python import tensorflow as tf I tf constant 0 def body I w tf variable tf constant 1 return I w loop tf while loop lam...
tensorflowtensorflow
tf test compute gradient give error when computation involve tensorarray
Bug
if possible provide a minimal reproducible example we usually don t have time to read hundred of line of your code python a tf tensorarray tf float32 size 1 x tf one 5 a a write 0 x y a stack grad tf gradient y x with tf session as sess print y ...
tensorflowtensorflow
ctc greedy decoder output sum of logit and not logprob
Bug
ctc greedy decoder output sum of logit and not logprob I have be write my own ctc greedy decoder function in order to decode the output of a rnn train with the ctc loss to check my version I decide to compare my output with tensorflow s one we have the same first output the decode label but we have bi...
tensorflowtensorflow
tenor flow distribute support for spectrum lsf
Bug
note only file github issue for bug and feature request all other topic will be close for general support from the community see stackoverflow to make bug and feature request more easy to find and organize we close issue that be deem out of scope for github issue and point people to stackoverflow ...
tensorflowtensorflow
incorrect hessian of a quadratic function
Bug
note only file github issue for bug and feature request all other topic will be close for general support from the community see stackoverflow to make bug and feature request more easy to find and organize we close issue that be deem out of scope for github issue and point people to stackoverflow ...
tensorflowtensorflow
there be a learning rate parameter in the adadelta implementation although there isn t such one in the original paper
Bug
hi I see in the documentation that tf train adadeltaoptimizer have a learning rate parameter but in the original paper that parameter be eliminate from the update rule see equation 14 in the paper or step 5 in the pseudo code of the algorithm be that a bug in the implementation if you deliber...
tensorflowtensorflow
localclidebugwrappersession hide exception message
Bug
environment info operating system osx instal version of cuda and cudnn no if instal from binary pip package provide 1 a link to the pip package you instal nightly python3 today 2 the output from python c import tensorflow print tensorflow version 1 0 1 pytho...
tensorflowtensorflow
segfault check fail num element kmaxelement
Bug
I m run the current nightly build as of today 2017 03 17 I m work with very large sparse matrix and my program seem to have hit a variable overflow it crash with the following log entry 2017 03 17 10 48 34 554715 f tensorflow core framework tensor shape cc 212 check fail num element kmaxe...
tensorflowtensorflow
version web page need update
Bug
this really be a small issue but I couldn t submit a pr as I couldn t find the file for it at the current version be state as 0 12 when it should be 1 0 the doc at root I e in tensorflow org api doc refer to the most recent stable branch in this case r0 12 that s it sorry for raise an issue...
tensorflowtensorflow
cpu resource of tensorflow s docker container could not be control by cup share
Bug
environment info os ubuntu14 04lt gpu nvidia pascal titun x cuda8 0 cudnn cuda8 0 v5 1 docker verison 17 03 0 ce nvidia docker 1 0 1 cpu intel core i7 6900k hyper threading off turbo boost off docker file I instal tensorflow by dockerfile as below ...
tensorflowtensorflow
tensorflow not work in zeppelin 0 7 0
Bug
there seem to be some issue use tensorflow in zeppelin 0 7 0 and it throw this error nameerror name interactive be not define another user report the same issue on so the fix suggest there seem to be really hacky the issue seem to be that in tf log py file anaconda3 lib site package tensor...
tensorflowtensorflow
protobufs be into multiple share library load from python
Bug
issue description tensorflow currently fail with the follow error if compile use clang in c opt mode when try to import tensorflow contrib package in python python code reproduce the problem be very simple import tensorflow contrib program output libprotobuf error ex...
tensorflowtensorflow
tf lbeta error when feed with placeholder
Bug
hi there seem to be a bug in tf lbeta python tf lbeta tf placeholder tf float32 3 2 3 none valueerror can not squeeze dim 0 expect a dimension of 1 get 3 for lbeta 1 cond squeeze op squeeze with input shape 3 2 3
tensorflowtensorflow
documentation format break
Bug
see source code format leak into general text or return and raise get fold into parameter either the doc generator need to understand python doc comment well or the doc comment need to be update to work well with markdown extra newline etc what do you think thank andreas
tensorflowtensorflow
tutorial use gpu doc possibly other should use print since windows need python 3 x
Bug
should the python sample in tutorial use gpu and other doc be in 3 5 x form in install install window we re tell tensorflow only support version 3 5 x of python on window in tutorial use gpu there be sample python code in 2 x style print sess run c this code snippet choke when run on window...
tensorflowtensorflow
tfdbg error causality violate in timing relation of debug dump
Bug
what related github issue or stackoverflow thread have you find by search the web for your problem nothing find environment info operating system ubuntu 14 04 instal version of cuda and cudnn cuda 8 0 cudnn 5 1 please attach the output of ls l path to cuda lib libcud rw r...
tensorflowtensorflow
tf contrib distributions multinomial initialization problem
Bug
hi I can t initialize a multinomial distribution with more than 1 dimension probability the version of instal tf be 1 0 0 I try the example code in tensorflow website it get the same problem and the error doesn t make any sense to require probability to be integer I m not sure if the guide in the website be ...
tensorflowtensorflow
feedback on a guide to tf layer build a convolutional neural network tutorial
Bug
this tutorial seem to contain some error convolutional layer 1 the filter argument specify the number of filter to apply here 32 be the filter learn in that case you could mention it else how do you choose the type of the filter our output tensor produce by conv2d have a s...
tensorflowtensorflow
tf importgraphdef crash for the follow graph
Bug
what related github issue or stackoverflow thread have you find by search the web for your problem none environment info operating system ubuntu 16 04 instal version of cuda and cudnn please attach the output of ls l path to cuda lib libcud cuda rw r r 1 root r...
tensorflowtensorflow
different code path take in conv2d for constant vs variable filter
Bug
it appear that a different code path be take in conv2d for constant vs variable filter on a box with gpu this work python image tf constant np arange 2 15 15 2 reshape 2 2 15 15 astype np float32 filter tf variable initial value 1 np one 1 1 2 1 np float32 with tf ses...
tensorflowtensorflow
documentation incorrect for rnn tutorial
Bug
be it possible the documentation be incorrect on word tf placeholder tf int32 batch size num step for I in range num step the value of state be update after process each batch of word output state lstm word I state it seem the correct output should be ...
tensorflowtensorflow
minor but confusing mistake in doc about shape of output tensor of gather nd
Bug
accord to the doc of gather nd param indice the output tensor produce by the function have shape d 0 d q 2 param shape k param shape p 1 where d 0 d q 2 be the dimension of index without the last one p be the rank of param and k be the leng...
tensorflowtensorflow
memory leak when train complex neural network
Bug
when train some complex seq to seq neural network the memory cost of my program will keep grow and this only happen on gpu on cpu everything be ok I use to report this problem in this issue after that I solve this problem by encapsulate my encoding method as an rnn cell so this issue be close ...
tensorflowtensorflow
operation documentation have within it
Bug
l1154 this wasn t cause a problem because tf be strip out the documentation before hit python and the python protobuf text format py I want to expose the documentation in the python a patch I be not recommend per but run into numerous issue in the protobuf see in the above issue post to protobuf the l...
tensorflowtensorflow
tf py func treat result different on window and ubuntu in tensorflow 1 0
Bug
description I m use tf py func in my datum fetch pipeline the apply function basically calculate an int from an int when I run the tensorflow code it work on my window 10 development laptop but fail on the ubuntu server with an error that the python function would return an int64 instead of the expect ...
tensorflowtensorflow
unclear documentation for tf layer dense flatten behavior
Bug
documentation state note if the input tensor have a rank great than 2 then it be flatten prior to the initial matrix multiply by w however the follow return tensor with shape shape 2 2 2 400 as if the input have not be flatten tf layer dense tf placeholder tf float32 ...
tensorflowtensorflow
tfdbg require external ncurse on osx
Bug
osx 10 11 6 tf from source at 8746f8ac9 master head from a few hour ago without gpu xcode 8 2 1 when load it tfdbg crash with a curse relate error be solve with brew install homebrew dupe ncurse if this be really a dependency it would be useful to mention it in the documentation or inst...
tensorflowtensorflow
well document when to use tf sparse tensor dense matmul vs embed lookup
Bug
tf sparse tensor dense matmul have a regular non sparse tensor gradient source l153 python gradient w r t dense b grad sparse op sparse tensor dense matmul sp t grad adjoint a not adj a where sparse op sparse tensor...
tensorflowtensorflow
tf upgrade doesn t update rnn cell
Bug
migration script tf upgrade py do not update rnn cell location from tf nn rnn cell to tf contrib rnn lead to error attributeerror module tensorflow python op nn have no attribute rnn cell upgrading python import tensorflow as tf rnn tf nn rnn cell grucell 128 tf initialize ...
tensorflowtensorflow
scatter nd doc not clear about concurrent update
Bug
the doc on the scatter nd do not specify the consequence of multiple update that reference the same location I ve test this use the follow code python indice tf constant 4 3 1 1 update tf constant 9 10 11 12 shape tf constant 8 scatter tf scatter nd indi...
tensorflowtensorflow
miss image in doc
Bug
the page have a break link to an image also have a break link in addition the image seem to be miss from the tensorflow repo could those be put in there it be hard to build the doc locally for offline use without the image
tensorflowtensorflow
error in tensorflow debugger tfdbg while execute session run call in child thread
Bug
I run tensorflow debugger use the command python m debug but get the follow error I e signal only work in main thread screenshot3png the error be throw when a session run call be execute in a child thread spawn from main thread be tensorflow debugger only support for single thread applica...
tensorflowtensorflow
multiple break g3doc link
Bug
there be by my count at least 20 break g3doc link scatter among documentation code and build script cause by this commit migrating g3doc to the docs src source code documentation generating system so far in relation to this the follow issue have be open 8014 7999 7989 and the follow pu...
tensorflowtensorflow
quantize graph round and quantize mode be break
Bug
look at code in master cc petewarden a keyerror be always produce consider the round mode already visit python if self mode round self already visit for output node in output node self round node recursively output node which will f...
tensorflowtensorflow
typeerror fetch argument none have invalid type
Bug
feature request for a well error description or for well summary handle the follow code work fine if some summary where define before op op tf summary merge all session run op however if there be no summary we get typeerror fetch argument none have invalid type which be re...