repository stringclasses 156 values | issue title stringlengths 1 1.01k ⌀ | labels stringclasses 8 values | body stringlengths 1 270k ⌀ |
|---|---|---|---|
tensorflowtensorflow | link error with libtflite gpu gl so cause by tflitegpudelegatedelete absence among export symbol | Bug | system information os platform and distribution linux ubuntu 14 04 tensorflow instal from source tensorflow version 2 0 alpha bazel version 0 24 1 gcc compiler version clang 8 0 2 issue tflitegpudelegatedelete doesn t declare as export symbol in tensorflow lite delegates gpu gl delegate h this cause link error with libtflite gpu gl so build command bazel build c opt config android arm64 copt os copt dtflite gpu binary release copt fvisibility hide linkopt s strip always cxxopt std c 11 tensorflow lite delegates gpu libtflite gpu gl so reproduce set up interpreter auto model flatbuffermodel buildfromfile model path op builtin builtinopresolver op resolver std unique ptr interpreter interpreterbuilder model op resolver interpreter new prepare gpu delegate auto delegate tflitegpudelegatecreate option nullptr if interpreter modifygraphwithdelegate delegate ktfliteok return run inference writetoinputtensor interpreter type input tensor 0 if interpreter invoke ktfliteok return readfromoutputtensor interpreter type output tensor 0 clean up tflitegpudelegatedelete delegate I follow example describe at tensorflow lite delegates gpu readme md |
tensorflowtensorflow | bug attributeerror indexedslice object have no attribute copy | Bug | tensorflow version python c import tensorflow as tf print tf git version tf version v1 12 0 0 ga6d8ffae09 1 12 0 issue 2019 04 19 15 30 36 803841 I tensorflow core platform cpu feature guard cc 141 your cpu support instruction that this tensorflow binary be not compile to use avx2 fma 2019 04 19 15 30 36 879298 I tensorflow stream executor cuda cuda gpu executor cc 964 successful numa node read from sysfs have negative value 1 but there must be at least one numa node so return numa node zero 2019 04 19 15 30 36 879649 I tensorflow core common runtime gpu gpu device cc 1432 find device 0 with property name titan x pascal major 6 minor 1 memoryclockrate ghz 1 531 pcibusid 0000 02 00 0 totalmemory 11 91gib freememory 4 75gib 2019 04 19 15 30 36 879665 I tensorflow core common runtime gpu gpu device cc 1511 add visible gpu device 0 2019 04 19 15 30 37 551722 I tensorflow core common runtime gpu gpu device cc 982 device interconnect streamexecutor with strength 1 edge matrix 2019 04 19 15 30 37 551754 I tensorflow core common runtime gpu gpu device cc 988 0 2019 04 19 15 30 37 551764 I tensorflow core common runtime gpu gpu device cc 1001 0 n 2019 04 19 15 30 37 551930 I tensorflow core common runtime gpu gpu device cc 1115 create tensorflow device job localhost replica 0 task 0 device gpu 0 with 4496 mb memory physical gpu device 0 name titan x pascal pci bus i d 0000 02 00 0 compute capability 6 1 info xxx store generate index with key 1 ttt info xxx cli py loss batch 1 on 800 1 6608102321624756 traceback most recent call last file bin xxx cli py line 253 in main sys argv 1 file bin xxx cli py line 193 in main gradient tape gradient loss trainable variable file home anaconda anaconda3 lib python3 6 site package tensorflow python eager backprop py line 901 in gradient output gradient output gradient file home anaconda anaconda3 lib python3 6 site package tensorflow python eager imperative grad py line 64 in imperative grad output gradient file home anaconda anaconda3 lib python3 6 site package tensorflow python framework op py line 858 in grad fun return dresult copy device name self device attributeerror indexedslice object have no attribute copy environment information cat etc issue linux hamburg 4 15 0 47 generic 50 ubuntu smp we d mar 13 10 44 52 utc 2019 x86 64 x86 64 x86 64 gnu linux version 18 04 2 lts bionic beaver version i d 18 04 version codename bionic be we in docker no compiler c ubuntu 6 5 0 2ubuntu1 18 04 6 5 0 20181026 copyright c 2017 free software foundation inc this be free software see the source for copy condition there be no warranty not even for merchantability or fitness for a particular purpose uname a linux hamburg 4 15 0 47 generic 50 ubuntu smp we d mar 13 10 44 52 utc 2019 x86 64 x86 64 x86 64 gnu linux check pip msgpack numpy 0 4 3 2 numpy 1 16 1 numpydoc 0 8 0 protobuf 3 6 1 tensorflow estimator 1 13 0 13084 try file tls x86 64 libpthread so 0 13084 try file tls libpthread so 0 13084 try file haswell x86 64 libpthread so 0 13084 try file haswell libpthread so 0 13084 try file x86 64 libpthread so 0 13084 try file libpthread so 0 13084 search cache etc ld so cache 13084 try file lib x86 64 linux gnu libpthread so 0 13084 13084 find library libc so 6 0 search 13084 search path home anaconda anaconda3 bin lib rpath from file home anaconda anaconda3 bin python 13084 try file home anaconda anaconda3 bin lib libc so 6 13084 search path usr local cuda 9 0 lib64 tls haswell x86 64 tls haswell tls x86 64 tls haswell x86 64 haswell x86 64 ld library path 13084 try file usr local cuda 9 0 lib64 libc so 6 13084 try file tls haswell x86 64 libc so 6 13084 try file tls haswell libc so 6 13084 try file tls x86 64 libc so 6 13084 try file tls libc so 6 13084 try file haswell x86 64 libc so 6 13084 try file haswell libc so 6 13084 try file x86 64 libc so 6 13084 try file libc so 6 13084 search cache etc ld so cache 13084 try file lib x86 64 linux gnu libc so 6 13084 13084 find library libdl so 2 0 search 13084 search path home anaconda anaconda3 bin lib rpath from file home anaconda anaconda3 bin python 13084 try file usr local cuda 9 0 lib64 libm so 6 13084 try file tls haswell x86 64 libm so 6 13084 try file tls haswell libm so 6 13084 try file tls x86 64 libm so 6 13084 try file tls libm so 6 13084 try file haswell x86 64 libm so 6 13084 try file haswell libm so 6 13084 try file x86 64 libm so 6 13084 try file libm so 6 13084 search cache etc ld so cache 13084 try file lib x86 64 linux gnu libm so 6 13084 13084 13084 call init lib x86 64 linux gnu libpthread so 0 13084 13084 13084 call init lib x86 64 linux gnu libc so 6 13084 13084 13084 call init lib x86 64 linux gnu libm so 6 13084 13084 13084 call init lib x86 64 linux gnu librt so 1 13084 13084 13084 call init lib x86 64 linux gnu libutil so 1 13084 13084 try file home anaconda anaconda3 lib python3 6 lib dynload tls haswell libffi so 6 13084 try file home anaconda anaconda3 lib python3 6 lib dynload tls x86 64 libffi so 6 13084 try file home anaconda anaconda3 lib python3 6 lib dynload tls libffi so 6 13084 try file home anaconda anaconda3 lib python3 6 lib dynload haswell x86 64 libffi so 6 13084 try file home anaconda anaconda3 lib python3 6 lib dynload haswell libffi so 6 13084 try file home anaconda anaconda3 lib python3 6 lib dynload x86 64 libffi so 6 13084 try file home anaconda anaconda3 lib python3 6 lib dynload libffi so 6 13084 13084 13084 call init home anaconda anaconda3 lib python3 6 lib dynload libffi so 6 13084 13084 13084 call init home anaconda anaconda3 lib python3 6 lib dynload ctype cpython 36 m x86 64 linux gnu so 13084 13084 13084 call init home anaconda anaconda3 lib python3 6 lib dynload struct cpython 36 m x86 64 linux gnu so 13084 13084 find library libopenblasp r0 382c8f3a 3 5 dev so 0 search 13084 search path home anaconda anaconda3 lib python3 6 site package numpy core lib tls haswell x86 64 home anaconda anaconda3 lib python3 6 site package numpy core lib tls haswell home anaconda anaconda3 lib python3 6 site package numpy core lib tls x86 64 home anaconda anaconda3 lib python3 6 site package numpy core lib tls home anaconda anaconda3 lib python3 6 site package numpy core lib haswell x86 64 home anaconda anaconda3 lib python3 6 site package numpy core lib haswell home anaconda anaconda3 lib python3 6 site package numpy core lib x86 64 home anaconda anaconda3 lib python3 6 site package numpy core libs rpath from file home anaconda anaconda3 lib python3 6 site package numpy core multiarray umath cpython 36 m x86 64 linux gnu so 13084 try file home anaconda anaconda3 lib python3 6 site package numpy core lib tls haswell x86 64 libopenblasp r0 382c8f3a 3 5 dev so 13084 try file home anaconda anaconda3 lib python3 6 site package numpy core lib tls haswell libopenblasp r0 382c8f3a 3 5 dev so 13084 try file home anaconda anaconda3 lib python3 6 site package numpy core lib tls x86 64 libopenblasp r0 382c8f3a 3 5 dev so 13084 try file home anaconda anaconda3 lib python3 6 site package numpy core lib tls libopenblasp r0 382c8f3a 3 5 dev so 13084 try file home anaconda anaconda3 lib python3 6 site package numpy core lib haswell x86 64 libopenblasp r0 382c8f3a 3 5 dev so 13084 try file home an 13084 13084 13084 call init home anaconda anaconda3 lib python3 6 lib dynload libz so 1 13084 13084 13084 call init home anaconda anaconda3 lib python3 6 lib dynload zlib cpython 36 m x86 64 linux gnu so 13084 13084 13084 call init home anaconda anaconda3 lib python3 6 lib dynload bz2 cpython 36 m x86 64 linux gnu so 13084 13084 find library liblzma so 5 0 search 13084 search path home anaconda anaconda3 lib python3 6 lib dynload rpath from file home anaconda anaconda3 lib python3 6 lib dynload ctype cpython 36 m x86 64 linux gnu so 13084 try file home anaconda anaconda3 lib python3 6 lib dynload liblzma so 5 13084 13084 13084 call init home anaconda anaconda3 lib python3 6 lib dynload liblzma so 5 13084 13084 13084 call init home anaconda anaconda3 lib python3 6 lib dynload lzma cpython 36 m x86 64 linux gnu so 13084 13084 13084 call init home anaconda anaconda3 lib python3 6 lib dynload grp cpython 36 m x86 64 linux gnu so 13084 13084 13084 call init home anaconda anaconda3 lib python3 6 lib dynload decimal cpython 36 m x86 64 linux gnu so 13084 13084 13084 call init home anaconda anaconda3 lib python3 6 site package numpy fft fftpack lite cpython 36 m x86 64 linux 13084 call init home anaconda anaconda3 lib python3 6 lib dynload hashlib cpython 36 m x86 64 linux gnu so 13084 13084 13084 call init home anaconda anaconda3 lib python3 6 lib dynload blake2 cpython 36 m x86 64 linux gnu so 13084 13084 13084 call init home anaconda anaconda3 lib python3 6 lib dynload sha3 cpython 36 m x86 64 linux gnu so 13084 13084 13084 call init home anaconda anaconda3 lib python3 6 lib dynload bisect cpython 36 m x86 64 linux gnu so 13084 13084 13084 call init home anaconda anaconda3 lib python3 6 lib dynload random cpython 36 m x86 64 linux gnu so 13084 13084 find library libtensorflow framework so 0 search 13084 search path home anaconda anaconda3 lib python3 6 site package tensorflow python solib local u s stensorflow spython c upywrap utensorflow uinternal so utensorflow tls haswell x86 64 home anaconda anaconda3 lib python3 6 site package tensorflow python solib local u s stensorflow spython c upywrap utensorflow uinternal so utensorflow tls haswell home anaconda anaconda3 lib python3 6 site package tensorflow python solib local u s stensorflow spython c upywrap utensorflow uinternal so utensorflow tls x86 64 home anaconda anaconda3 lib python3 6 site package tensorflow python solib local u s stensorflow spython c upywrap utensorflow uinternal so utensorflow tls home anaconda anaconda3 lib python3 6 site package tensorflow python solib local u s stensorflow spython c upywrap utensorflow uinternal so utensorflow haswell x86 64 home anaconda anaconda3 lib python3 6 site package tensorflow python solib local u s stensorflow spython c upywrap utensorflow uinternal so utensorflow haswell home anaconda anaconda3 lib python3 6 site package tensorflow python solib local u s stensorflow spython c upywrap utensorflow uinternal so utensorflow x86 64 home anaconda anaconda3 lib python3 6 site package tensorflow python solib local u s stensorflow spython c upywrap utensorflow uinternal so utensorflow home anaconda anaconda3 lib python3 6 site package tensorflow python solib local u local uconfig ucuda s scuda ccubla uexternal slocal uconfig ucuda scuda scuda slib tls haswell x86 64 home anaconda anaconda3 lib python3 6 site package tensorflow python solib local u local uconfig ucuda s scuda ccubla uexternal slocal uconfig ucuda scuda scuda slib tls haswell home anaconda anaconda3 lib python3 6 site package tensorflow python solib local u local uconfig ucuda s scuda ccubla uexternal slocal uconfig ucuda scuda scuda slib tls x86 64 home anaconda anaconda3 lib python3 6 site package tensorflow python solib local u local uconfig ucuda s scuda ccubla uexternal slocal uconfig ucuda scuda scuda slib tls home anaconda anaconda3 lib python3 6 site package tensorflow python solib local u local uconfig ucuda s scuda ccubla uexternal slocal uconfig ucuda scuda scuda slib haswell x86 64 home anaconda anaconda3 lib python3 6 site package tensorflow python solib local u local uconfig ucuda s scuda ccubla uexternal slocal uconfig ucuda scuda scuda slib haswell home anaconda anaconda3 lib python3 6 site package tensorflow python solib local u local uconfig ucuda s scuda ccubla uexternal slocal uconfig ucuda scuda scuda slib x86 64 home anaconda anaconda3 lib python3 6 site package tensorflow python solib local u local uconfig ucuda s scuda ccubla uexternal slocal uconfig ucuda scuda scuda slib home anaconda anaconda3 lib python3 6 site package tensorflow python solib local u local uconfig ucuda s scuda ccusolver uexternal slocal uconfig ucuda scuda scuda slib tls haswell x86 64 home anaconda anaconda3 lib python3 6 site package tensorflow python solib local u local uconfig ucuda s scuda ccusolver uexternal slocal uconfig ucuda scuda scuda slib tls haswell home anaconda anaconda3 lib python3 6 site package tensorflow python solib local u local uconfig ucuda s scuda ccusolver uexternal slocal uconfig ucuda scuda scuda slib tls x86 64 home anaconda anaconda3 lib python3 6 site package tensorflow python solib local u local uconfig ucuda s scuda ccusolver uexternal slocal uconfig ucuda scuda scuda slib tls home anaconda anaconda3 lib python3 6 site package tensorflow python solib local u local uconfig ucuda s scuda ccusolver uexternal slocal uconfig ucuda scuda scuda slib haswell x86 64 home anaconda anaconda3 lib python3 6 site package tensorflow python solib local u local uconfig ucuda s scuda ccusolver uexternal slocal uconfig ucuda scuda scuda slib haswell home anaconda anaconda3 lib python3 6 site package tensorflow python solib local u local uconfig ucuda s scuda ccusolver uexternal slocal uconfig ucuda scuda scuda slib x86 64 home anaconda anaconda3 lib python3 6 site package tensorflow python solib local u local uconfig ucuda s scuda ccusolver uexternal slocal uconfig ucuda scuda scuda slib home anaconda anaconda3 lib python3 6 site package tensorflow python solib local u local uconfig ucuda s scuda ccudart uexternal slocal uconfig ucuda scuda scuda slib tls haswell x86 64 home anaconda anaconda3 lib python3 6 site package tensorflow python solib local u local uconfig ucuda s scuda ccudart uexternal slocal uconfig ucuda scuda scuda slib tls haswell home anaconda anaconda3 lib python3 6 site package tensorflow python solib local u local uconfig ucuda s scuda ccudart uexternal slocal uconfig ucuda scuda scuda slib tls x86 64 home anaconda anaconda3 lib python3 6 site package tensorflow python solib local u local uconfig ucuda s scuda ccudart uexternal slocal uconfig ucuda scuda scuda slib tls home anaconda anaconda3 lib python3 6 site package tensorflow python solib local u local uconfig ucuda s scuda ccudart uexternal slocal uconfig ucuda scuda scuda slib haswell x86 64 home anaconda anaconda3 lib python3 6 site package tensorflow python solib local u local uconfig ucuda s scuda ccudart uexternal slocal uconfig ucuda scuda scuda slib haswell home anaconda anaconda3 lib python3 6 site package tensorflow python solib local u local uconfig ucuda s scuda ccudart uexternal slocal uconfig ucuda scuda scuda slib x86 64 home anaconda anaconda3 lib python3 6 site package tensorflow python solib local u local uconfig ucuda s scuda ccudart uexternal slocal uconfig ucuda scuda scuda slib home anaconda anaconda3 lib python3 6 site package tensorflow python tls haswell x86 64 home anaconda anaconda3 lib python3 6 site package tensorflow python tls haswell home anaconda anaconda3 lib python3 6 site package tensorflow python tls x86 64 home anaconda anaconda3 lib python3 6 site package tensorflow python tls home anaconda anaconda3 lib python3 6 site package tensorflow python haswell x86 64 home anaconda anaconda3 lib python3 6 site package tensorflow python haswell home anaconda anaconda3 lib python3 6 site package tensorflow python x86 64 home anaconda anaconda3 lib python3 6 site package tensorflow python home anaconda anaconda3 lib python3 6 site package tensorflow python tls haswell x86 64 home anaconda anaconda3 l config cuda cuda lib64 tls haswell x86 64 local config cuda cuda lib64 tls haswell local config cuda cuda lib64 tls x86 64 local config cuda cuda lib64 tls local config cuda cuda lib64 haswell x86 64 local config cuda cuda lib64 haswell local config cuda cuda lib64 x86 64 local config cuda cuda lib64 local config cuda cuda extras cupti lib64 tls haswell x86 64 local config cuda cuda extras cupti lib64 tls haswell local config cuda cuda extras cupti lib64 tls x86 64 local config cuda cuda extras cupti lib64 tls local config cuda cuda extras cupti lib64 haswell x86 64 local config cuda cuda extras cupti lib64 haswell local config cuda cuda extras cupti lib64 x86 64 local config cuda cuda extras cupti lib64 rpath from file home anaconda anaconda3 lib python3 6 site package tensorflow python pywrap tensorflow internal so 13084 try file home anaconda anaconda3 lib python3 6 site package tensorflow python solib local u s stensorflow spython c upywrap utensorflow uinternal so utensorflow tls haswell x86 64 libtensorflow framework so 13084 try file home anaconda anaconda3 lib python3 6 site package tensorflow python solib local u s stensorflow spython c upywrap utensorflow uinternal so utensorflow tls haswell libtensorflow framework so 13084 try file home anaconda anaconda3 lib python3 6 site package tensorflow python solib local u s stensorflow spython c upywrap utensorflow uinternal so utensorflow tls x86 64 libtensorflow framework so 13084 try file home anaconda anaconda3 lib python3 6 site package tensorflow python solib local u s stensorflow spython c upywrap utensorflow uinternal so utensorflow tls libtensorflow framework so 13084 try file home anaconda anaconda3 lib python3 6 site package tensorflow python solib local u s stensorflow spython c upywrap utensorflow uinternal so utensorflow haswell x86 64 libtensorflow framework so 13084 try file home anaconda anaconda3 lib python3 6 site package tensorflow python solib local u s stensorflow spython c upywrap utensorflow uinternal so utensorflow haswell libtensorflow framework so 13084 try file home anaconda anaconda3 lib python3 6 site package tensorflow python solib local u s stensorflow spython c upywrap utensorflow uinternal so utensorflow x86 64 libtensorflow framework so 13084 try file home anaconda anaconda3 lib python3 6 site package tensorflow python solib local u s stensorflow spython c upywrap utensorflow uinternal so utensorflow libtensorflow framework so 13084 try file home anaconda anaconda3 lib python3 6 site package tensorflow python solib local u local uconfig ucuda s scuda ccubla uexternal slocal uconfig ucuda scuda scuda slib tls haswell x86 64 libtensorflow framework so 13084 try file home anaconda anaconda3 lib python3 6 site package tensorflow python solib local u local uconfig ucuda s scuda ccubla uexternal slocal uconfig ucuda scuda scuda slib tls haswell libtensorflow framework so 13084 try file home anaconda anaconda3 lib python3 6 site package tensorflow python solib local u local uconfig ucuda s scuda ccubla uexternal slocal uconfig ucuda scuda scuda slib tls x86 64 libtensorflow framework so 13084 try file home anaconda anaconda3 lib python3 6 site package tensorflow pyth 4 libtensorflow framework so 13084 try file home anaconda anaconda3 lib python3 6 site package tensorflow python solib local u local uconfig ucuda s scuda ccubla uexternal slocal uconfig ucuda scuda scuda slib libtensorflow framework so 13084 try file home anaconda anaconda3 lib python3 6 site package tensorflow python solib local u local uconfig ucuda s scuda ccusolver uexternal slocal uconfig ucuda scuda scuda slib tls haswell x86 64 libtensorflow framework so 13084 try file home anaconda anaconda3 lib python3 6 site package tensorflow python solib local u local uconfig ucuda s scuda ccusolver uexternal slocal uconfig ucuda scuda scuda slib tls haswell libtensorflow framework so 13084 try file home anaconda anaconda3 lib python3 6 site package tensorflow python solib local u local uconfig ucuda s scuda ccusolver uexternal slocal uconfig ucuda scuda scuda slib tls x86 64 libtensorflow framework so 13084 try file home anaconda anaconda3 lib python3 6 site package tensorflow python solib local u local uconfig ucuda s scuda ccusolver uexternal slocal uconfig ucuda scuda scuda slib tls libtensorflow framework so 13084 try file home anaconda anaconda3 lib python3 6 site package tensorflow python solib local u local uconfig ucuda s scuda ccusolver uexternal slocal uconfig ucuda scuda scuda slib haswell x86 64 libtensorflow framework so 13084 try file home anaconda anaconda3 lib python3 6 site package tensorflow python solib local u local uconfig ucuda s scuda ccusolver uexternal slocal uconfig ucuda scuda scuda slib haswell libtensorflow framework so 13084 try file home anaconda anaconda3 lib python3 6 site package tensorflow python solib local u local uconfig ucuda s scuda ccusolver uexternal slocal uconfig ucuda scuda scuda slib x86 64 libtensorflow framework so 13084 try file home anaconda anaconda3 lib python3 6 site package tensorflow python solib local u local uconfig ucuda s scuda ccusolver uexternal slocal uconfig ucuda scuda scuda slib libtensorflow framework so 13084 try file home anaconda anaconda3 lib python3 6 site package tensorflow python solib local u local uconfig ucuda s scuda ccudart uexternal slocal uconfig ucuda scuda scuda slib tls haswell x86 64 libtensorflow framework so 13084 try file home anaconda anaconda3 lib python3 6 site package tensorflow python solib local u local uconfig ucuda s scuda ccudart uexternal slocal uconfig ucuda scuda scuda slib tls haswell libtensorflow framework so 13084 try file home anaconda anaconda3 lib python3 6 site package tensorflow python solib local u local uconfig ucuda s scuda ccudart uexternal slocal uconfig ucuda scuda scuda slib tls x86 64 libtensorflow framework so 13084 try file home anaconda anaconda3 lib python3 6 site package tensorflow python solib local u local uconfig ucuda s scuda ccudart uexternal slocal uconfig ucuda scuda scuda slib tls libtensorflow framework so 13084 try file home anaconda anaconda3 lib p ll libtensorflow framework so 13084 try file home anaconda anaconda3 lib python3 6 site package tensorflow python solib local u local uconfig ucuda s scuda ccudart uexternal slocal uconfig ucuda scuda scuda slib x86 64 libtensorflow framework so 13084 try file home anaconda anaconda3 lib python3 6 site package tensorflow python solib local u local uconfig ucuda s scuda ccudart uexternal slocal uconfig ucuda scuda scuda slib libtensorflow framework so 13084 try file home anaconda anaconda3 lib python3 6 site package tensorflow python tls haswell x86 64 libtensorflow framework so 13084 try file home anaconda anaconda3 lib python3 6 site package tensorflow python tls haswell libtensorflow framework so 13084 try file home anaconda anaconda3 lib python3 6 site package tensorflow python tls x86 64 libtensorflow framework so 13084 try file home anaconda anaconda3 lib python3 6 site package tensorflow python tls libtensorflow framework so 13084 try file home anaconda anaconda3 lib python3 6 site package tensorflow python haswell x86 64 libtensorflow framework so 13084 try file home anaconda anaconda3 lib python3 6 site package tensorflow python haswell libtensorflow framework so 13084 try file home anaconda anaconda3 lib python3 6 site package tensorflow python x86 64 libtensorflow framework so 13084 try file home anaconda anaconda3 lib python3 6 site package tensorflow python libtensorflow framework so 13084 try file home anaconda anaconda3 lib python3 6 site package tensorflow python tls haswell x86 64 libtensorflow framework so 13084 try file home anaconda anaconda3 lib python3 6 site package tensorflow python tls haswell libtensorflow framework so 13084 try file home anaconda anaconda3 lib python3 6 site package tensorflow python tls x86 64 libtensorflow framework so 13084 try file home anaconda anaconda3 lib python3 6 site package tensorflow python tls libtensorflow framework so 13084 try file home anaconda anaconda3 lib python3 6 site package tensorflow python haswell x86 64 libtensorflow framework so 13084 try file home anaconda anaconda3 lib python3 6 site package tensorflow python haswell libtensorflow framework so 13084 try file home anaconda anaconda3 lib python3 6 site package tensorflow python x86 64 libtensorflow framework so 13084 try file home anaconda anaconda3 lib python3 6 site package tensorflow python libtensorflow framework so 13084 13084 find library libcubla so 9 0 0 search 13084 search path home anaconda anaconda3 lib python3 6 site package tensorflow python home anaconda anaconda3 lib python3 6 site package tensorflow python local config cuda cuda lib64 tls haswell x86 64 local config cuda cuda lib64 tls haswell local config cuda cuda lib64 tls x86 64 local config cuda cuda lib64 tls local config cuda cuda lib64 haswell x86 64 local config cuda cuda lib64 haswell local config cuda cuda lib64 x86 64 local config cuda cuda lib64 local config cuda cuda extras cupti lib64 tls haswell x86 64 local config cuda cuda extras cupti lib64 tls haswell local config cuda cuda extras cupti lib64 tls x86 64 local config cuda cuda extras cupti lib64 tls local config cuda cuda extras cupti lib64 haswell x86 64 local config cuda cuda extras cupti lib64 haswell local config cuda cuda extras cupti lib64 x86 64 local config cuda cuda extras cupti lib64 r 13084 try file local config cuda cuda extras cupti lib64 tls haswell libcubla so 9 0 13084 try file local config cuda cuda extras cupti lib64 tls x86 64 libcubla so 9 0 13084 try file local config cuda cuda extras cupti lib64 tls libcublas so 9 0 13084 try file local config cuda cuda extras cupti lib64 haswell x86 64 libcubla so 9 0 13084 try file local config cuda cuda extras cupti lib64 haswell libcublas so 9 0 13084 try file local config cuda cuda extras cupti lib64 x86 64 libcubla so 9 0 13084 try file local config cuda cuda extras cupti lib64 libcublas so 9 0 13084 search path home anaconda anaconda3 bin lib rpath from file home anaconda anaconda3 bin python 13084 try file home anaconda anaconda3 bin lib libcublas so 9 0 13084 13084 find library libcusolver so 9 0 0 search 13084 search path home anaconda anaconda3 lib python3 6 site package tensorflow python home anaconda anaconda3 lib python3 6 site package tensorflow python local config cuda cuda lib64 tls haswell x86 64 local config cuda cuda lib64 tls haswell local config cuda cuda lib64 tls x86 64 local config cuda cuda lib64 tls local config cuda cuda lib64 haswell x86 64 local config cuda cuda lib64 haswell local config cuda cuda lib64 x86 64 local config cuda cuda lib64 local config cuda cuda extras cupti lib64 tls haswell x86 64 local config cuda cuda extras cupti lib64 tls haswell local config cuda cuda extras cupti lib64 tls x86 64 local config cuda cuda extras cupti lib64 tls local config cuda cuda extras cupti lib64 haswell x86 64 local config cuda cuda extras cupti lib64 haswell local config cuda cuda extras cupti lib64 x86 64 local config cuda cuda extras cupti lib64 rpath from file home anaconda anaconda3 lib python3 6 site package tensorflow python pywrap tensorflow internal so 13084 try file home anaconda anaconda3 lib python3 6 site package tensorflow python libcusolver so 9 0 13084 try file home anaconda anaconda3 lib python3 6 site package tensorflow python libcusolver so 9 0 13084 try file local config cuda cuda lib64 tls haswell x86 64 libcusolver so 9 0 13084 try file local config cuda cuda lib64 tls haswell libcusolver so 9 0 13084 try file local config cuda cuda lib64 tls x86 64 libcusolver so 9 0 13084 try file local config cuda cuda lib64 tls libcusolver so 9 0 13084 try file local config cuda cuda lib64 haswell x86 64 libcusolver so 9 0 13084 try file local config cuda cuda lib64 haswell libcusolver so 9 0 13084 try file local config cuda cuda lib64 x86 64 libcusolver so 9 0 13084 try file local config cuda cuda lib64 libcusolver so 9 0 13084 search path home anaconda anaconda3 lib python3 6 site package tensorflow python home anaconda anaconda3 lib python3 6 site package tensorflow python local config cuda cuda lib64 tls haswell x86 64 local config cuda cuda lib64 tls haswell local config cuda cuda lib64 tls x86 64 local config cuda cuda lib64 tls local config cuda cuda lib64 haswell x86 64 local config cuda cuda lib64 haswell local config cuda cuda lib64 x86 64 local config cuda cuda lib64 local config cuda cuda extras cupti lib64 tls haswell x86 64 local config cuda cuda extras cupti lib64 tls haswell local config cuda cuda extras cupti lib64 tls x86 64 local config cuda cuda extras cupti lib64 tls local config cuda cuda extras cupti lib64 haswell x86 64 local config cuda cuda extras cupti lib64 haswell local config cuda cuda extras cupti lib64 x86 64 local config cuda cuda extras cupti lib64 rpath from file home anaconda anaconda3 lib python3 6 site package tensorflow python pywrap tensorflow internal so 13084 try file home anaconda anaconda3 lib python3 6 site package tensorflow python libcudart so 9 0 13084 try file home anaconda anaconda3 lib python3 6 site package tensorflow python libcudart so 9 0 13084 try file local config cuda cuda lib64 tls haswell x86 64 libcudart so 9 0 13084 try file local config cuda cuda lib64 tls haswell libcudart so 9 0 13084 try file local config cuda cuda lib64 tls x86 64 libcudart so 9 0 13084 try file local config cuda cuda lib64 tls libcudart so 9 0 13084 try file local config cuda cuda lib64 haswell x86 64 libcudart so 9 0 13084 try file local config cuda cuda lib64 haswell libcudart so 9 0 13084 try file local config cuda cuda lib64 x86 64 libcudart so 9 0 13084 try file local config cuda cuda lib64 libcudart so 9 0 13084 try file local config cuda cuda extras cupti lib64 tls haswell x86 64 libcudart so 9 0 13084 try file local config cuda cuda extras cupti lib64 tls haswell libcudart so 9 0 13084 try file local config cuda cuda extras cupti lib64 tls x86 64 libcudart so 9 0 13084 try file local config cuda cuda extras cupti lib64 tls libcudart so 9 0 13084 try file local config cuda cuda extras cupti lib64 haswell x86 64 libcudart so 9 0 13084 try file local config cuda cuda extras cupti lib64 haswell libcudart so 9 0 13084 try file local config cuda cuda extras cupti lib64 x86 64 libcudart so 9 0 13084 try file local config cuda cuda extras cupti lib64 libcudart so 9 0 13084 search path home anaconda anaconda3 bin lib rpath from file home anaconda anaconda3 bin python 13084 try file home anaconda anaconda3 bin lib libcudart so 9 0 13084 13084 find library libgomp so 1 0 sear 13084 try file home anaconda anaconda3 lib python3 6 site package tensorflow python libgomp so 1 13084 try file local config cuda cuda lib64 tls haswell x86 64 libgomp so 1 13084 try file local config cuda cuda lib64 tls haswell libgomp so 1 13084 try file local config cuda cuda lib64 tls x86 64 libgomp so 1 13084 try file local config cuda cuda lib64 tls libgomp so 1 13084 try file local config cuda cuda lib64 haswell x86 64 libgomp so 1 13084 try file local config cuda cuda lib64 haswell libgomp so 1 13084 try file local config cuda cuda lib64 x86 64 libgomp so 1 13084 try file local config cuda cuda lib64 libgomp so 1 13084 try file local config cuda cuda extras cupti lib64 tls haswell x86 64 libgomp so 1 13084 try file local config cuda cuda extras cupti lib64 tls haswell libgomp so 1 13084 try file local config cuda cuda extras cupti lib64 tls x86 64 libgomp so 1 13084 try file local config cuda cuda extras cupti lib64 tls libgomp so 1 13084 try file local config cuda cuda extras cupti lib64 haswell x86 64 libgomp so 1 13084 try file local config cuda cuda extras cupti lib64 haswell libgomp so 1 13084 try file local config cuda cuda extras cupti lib64 x86 64 libgomp so 1 13084 try file local config cuda cuda extras cupti lib64 libgomp so 1 13084 search path home anaconda anaconda3 bin lib rpath from file home anaconda anaconda3 bin python 13084 try file home anaconda anaconda3 bin lib libgomp so 1 13084 13084 find library libstdc so 6 0 search 13084 search path home anaconda anaconda3 lib python3 6 site package tensorflow python home anaconda anaconda3 lib python3 6 site package tensorflow python local config cuda cuda lib64 tls haswell x86 64 local config cuda cuda lib64 tls haswell local config cuda cuda lib64 tls x86 64 local config cuda cuda lib64 tls local config cuda cuda lib64 haswell x86 64 local config cuda cuda lib64 haswell local config cuda cuda lib64 x86 64 local config cuda cuda lib64 local config cuda cuda extras cupti lib64 tls haswell x86 64 local config cuda cuda extras cupti lib64 tls haswell local config cuda cuda extras cupti lib64 tls x86 64 local config cuda cuda extras cupti lib64 tls local config cuda cuda extras cupti lib64 haswell x86 64 local config cuda cuda extras cupti lib64 haswell local config cuda cuda extras cupti lib64 x86 64 local config cuda cuda extras cupti lib64 rpath from file home anaconda anaconda3 lib python3 6 site package tensorflow python pywrap tensorflow internal so 13084 try file home anaconda anaconda3 lib python3 6 site package tensorflow python libstdc so 6 13084 try file home anaconda anaconda3 lib python3 6 site package tensorflow python libstdc so 6 13084 try file local config cuda cuda lib64 tls haswell x86 64 libstdc so 6 13084 try file local config cuda cuda lib64 tls haswell libstdc so 6 13084 try file local config cuda cuda lib64 tls x86 64 libstdc so 6 13084 try file local config cuda cuda lib64 tls libstdc so 6 13084 try file local config cuda cuda lib64 haswell x86 64 libstdc so 6 13084 try file local config cuda cuda lib64 haswell libstdc so 6 13084 try file local config cuda cuda lib64 x86 64 libstdc so 6 13084 try file local config cuda cuda lib64 libstdc so 6 13084 try file local config cuda cuda extras cupti lib64 tls haswell x86 64 libstdc so 6 13084 try file local config cuda cuda extras cupti lib64 tls haswell libstdc so 6 13084 try file local config cuda cuda extras cupti lib64 tls x86 64 libstdc so 6 13084 try file local config cuda cuda extras cupti lib64 tls libstdc so 6 13084 try file local config cuda cuda extras cupti lib64 haswell x86 64 libstdc so 6 13084 try file local config cuda cuda extras cupti lib64 haswell libstdc so 6 13084 try file local config cuda cuda extras cupti lib64 x86 64 libstdc so 6 13084 try file local config cuda cuda extras cupti lib64 libstdc so 6 13084 search path home anaconda anaconda3 bin lib rpath from file home anaconda anaconda3 bin python 13084 try file home anaconda anaconda3 bin lib libstdc so 6 13084 13084 find library libgcc s so 1 0 search 13084 search path home anaconda anaconda3 lib python3 6 site package tensorflow python home anaconda anaconda3 lib python3 6 site package tensorflow python local config cuda cuda lib64 tls haswell x86 64 local config cuda cuda lib64 tls haswell local config cuda cuda lib64 tls x86 64 local config cuda cuda lib64 tls local config cuda cuda lib64 haswell x86 64 local config cuda cuda lib64 haswell local config cuda cuda lib64 x86 64 local config cuda cuda lib64 local config cuda cuda extras cupti lib64 tls haswell x86 64 local config cuda cuda extras cupti lib64 tls haswell local config cuda cuda extras cupti lib64 tls x86 64 local config cuda cuda extras cupti lib64 tls local config cuda cuda extras cupti lib64 haswell x86 64 local config cuda cuda extras cupti lib64 haswell local config cuda cuda extras cupti lib64 x86 64 local config cuda cuda extras cupti lib64 rpath from file home anaconda anaconda3 lib python3 6 site package tensorflow python pywrap tensorflow internal so 13084 try file home anaconda anaconda3 lib python3 6 site package tensorflow python libgcc s so 1 13084 try file home anaconda anaconda3 lib python3 6 site package tensorflow python libgcc s so 1 13084 try file local config cuda cuda lib64 tls haswell x86 64 libgcc s so 1 13084 try file local config cuda cuda lib64 tls haswell libgcc s so 1 13084 try file local config cuda cuda lib64 tls x86 64 libgcc s so 1 13084 try file local config cuda cuda lib64 tls libgcc s so 1 13084 try file local config cuda cuda lib64 haswell x86 64 libgcc s so 1 13084 try file local config cuda cuda lib64 haswell libgcc s so 1 13084 try file local config cuda cuda lib64 x86 64 libgcc s so 1 13084 try file local config cuda cuda lib64 libgcc s so 1 13084 try file local config cuda cuda extras cupti lib64 tls haswell x86 64 libgcc s so 1 13084 try file local config cuda cuda extras cupti lib64 tls haswell libgcc s so 1 13084 try file local config cuda cuda extras cupti lib64 tls x86 64 libgcc s so 1 13084 try file local config cuda cuda extras cupti lib64 tls libgcc s so 1 13084 try file local config cuda cuda extras cupti lib64 haswell x86 64 libgcc s so 1 13084 try file local config cuda cuda extras cupti lib64 haswell libgcc s so 1 13084 try file local config cuda cuda extras cupti lib64 x86 64 libgcc s so 1 13084 try file local config cuda cuda extras cupti lib64 libgcc s so 1 13084 search path home anaconda anaconda3 bin lib rpath from file home anaconda anaconda3 bin python 13084 try file home anaconda anaconda3 bin lib libgcc s so 1 13084 13084 find library libcuda so 1 0 search 13084 search path home anaconda anaconda3 lib python3 6 site package tensorflow python solib local u local uconfig ucuda s scuda ccuda udriver uexternal slocal uconfig ucuda scuda scuda slib tls haswell x86 64 home anaconda anaconda3 lib python3 6 site package tensorflow python solib local u local uconfig ucuda s scuda ccuda udriver uexternal slocal uconfig ucuda scuda scuda slib tls haswell home anaconda anaconda3 lib python3 6 site package tensorflow python solib local u local uconfig ucuda s scuda ccuda udriver uexternal slocal uconfig ucuda scuda scuda slib tls x86 64 home anaconda anaconda3 lib python3 6 site package tensorflow python solib local u local uconfig ucuda s scuda ccuda udriver uexternal slocal uconfig ucuda scuda scuda slib tls home anaconda anaconda3 lib python3 6 site package tensorflow python solib local u local uconfig ucuda s scuda ccuda udriver uexternal slocal uconfig ucuda scuda scuda slib haswell x86 64 home anaconda anaconda3 lib python3 6 site package tensorflow python solib local u local uconfig ucuda s scuda ccuda udriver uexternal slocal uconfig ucuda scuda scuda slib haswell home anaconda anaconda3 lib python3 6 site package tensorflow python solib local u local uconfig ucuda s scuda ccuda udriver uexternal slocal uconfig ucuda scuda scuda slib x86 64 home anaconda anaconda3 lib python3 6 site package tensorflow python solib local u local uconfig ucuda s scuda ccuda udriver uexternal slocal uconfig ucuda scuda scuda slib home anaconda anaconda3 lib python3 6 site package tensorflow python solib local u local uconfig ucuda s scuda ccudnn uexternal slocal uconfig ucuda scuda scuda slib tls haswell x86 64 home anaconda anaconda3 lib python3 6 site package tensorflow python solib local u local uconfig ucuda s scuda ccudnn uexternal slocal uconfig ucuda scuda scuda slib tls haswell home anaconda anaconda3 lib python3 6 site package tensorflow python solib local u local uconfig ucuda s scuda ccudnn uexternal slocal uconfig ucuda scuda scuda slib tls x86 64 home anaconda anaconda3 lib python3 6 site package tensorflow python solib local u local uconfig ucuda s scuda ccudnn uexternal slocal uconfig ucuda scuda scuda slib tls home anaconda anaconda3 lib python3 6 site package tensorflow python solib local u local uconfig ucuda s scuda ccudnn uexternal slocal uconfig ucuda scuda scuda slib haswell x86 64 home anaconda anaconda3 lib python3 6 site package tensorflow python solib local u local uconfig ucuda s scuda ccudnn uexternal slocal uconfig ucuda scuda scuda slib haswell home anaconda anaconda3 lib python3 6 site package tensorflow python solib local u local uconfig ucuda s scuda ccudnn uexternal slocal uconfig ucuda scuda scuda slib x86 64 home anaconda anaconda3 lib python3 6 site package tensorflow python solib local u local uconfig ucuda s scuda ccudnn uexternal slocal uconfig ucuda scuda scuda slib home anaconda anaconda3 lib python3 6 site package tensorflow python solib local u local uconfig ucuda s scuda ccufft uexternal slocal uconfig ucuda scuda scuda slib tls haswell x86 64 home anaconda anaconda3 lib python3 6 site package tensorflow python solib local u local uconfig ucuda s scuda ccufft uexternal slocal uconfig ucuda scuda scuda slib tls haswell home anaconda anaconda3 lib python3 6 site package tensorflow python solib local u local uconfig ucuda s scuda ccufft uexternal slocal uconfig ucuda scuda scuda slib tls x86 64 home anaconda anaconda3 lib python3 6 site package tensorflow python solib local u local uconfig ucuda s scuda ccufft uexternal slocal uconfig ucuda scuda scuda slib tls home anaconda anaconda3 lib python3 6 site package tensorflow python solib local u local uconfig ucuda s scuda ccufft uexternal slocal uconfig ucuda scuda scuda slib haswell x86 64 home anaconda anaconda3 lib python3 6 site package tensorflow python solib local u local uconfig ucuda s scuda ccufft uexternal slocal uconfig ucuda scuda scuda slib haswell home anaconda anaconda3 lib python3 6 site package tensorflow python solib local u local uconfig ucuda s scuda ccufft uexternal slocal uconfig ucuda scuda scuda slib x86 64 home anaconda anaconda3 lib python3 6 site package tensorflow python solib local u local uconfig ucuda s scuda ccufft uexternal slocal uconfig ucuda scuda scuda slib home anaconda anaconda3 lib python3 6 site package tensorflow python solib local u local uconfig ucuda s scuda ccurand uexternal slocal uconfig ucuda scuda scuda slib tls haswell x86 64 home anaconda anaconda3 lib python3 6 site package tensorflow python solib local u local uconfig ucuda s scuda ccurand uexternal slocal uconfig ucuda scuda scuda slib tls haswell home anaconda anaconda3 lib python3 6 site package tensorflow python solib local u local uconfig ucuda s scuda ccurand uexternal slocal uconfig ucuda scuda scuda slib tls x86 64 home anaconda anaconda3 lib python3 6 site package tensorflow python solib local u local uconfig ucuda s scuda ccurand uexternal slocal uconfig ucuda scuda scuda slib tls home anaconda anaconda3 lib python3 6 site package tensorflow python solib local u local uconfig ucuda s scuda ccurand uexternal slocal uconfig ucuda scuda scuda slib haswell x86 64 home anaconda anaconda3 lib python3 6 site package tensorflow python solib local u local uconfig ucuda s scuda ccurand uexternal slocal uconfig ucuda scuda scuda slib haswell home anaconda anaconda3 lib python3 6 site package tensorflow python solib local u local uconfig ucuda s scuda ccurand uexternal slocal uconfig ucuda scuda scuda slib x86 64 home anaconda anaconda3 lib python3 6 site package tensorflow python solib local u local uconfig ucuda s scuda ccurand uexternal slocal uconfig ucuda scuda scuda slib home anaconda anaconda3 lib python3 6 site package tensorflow python local config cuda cuda lib64 tls haswell x86 64 local config cuda cuda lib64 tls haswell local config cuda cuda lib64 tls x86 64 local config cuda cuda lib64 tls local config cuda cuda lib64 haswell x86 64 local config cuda cuda lib64 haswell local config cuda cuda lib64 x86 64 local config cuda cuda lib64 local config cuda cuda extras cupti lib64 tls haswell x86 64 local config cuda cuda extras cupti lib64 tls haswell local config cuda cuda extras cupti lib64 tls x86 64 local config cuda cuda extras cupti lib64 tls local config cuda cuda extras cupti lib64 haswell x86 64 local config cuda cuda extras cupti lib64 haswell local config cuda cuda extras cupti lib64 x86 64 local config cuda cuda extras cupti lib64 rpath from file home anaconda anaconda3 lib python3 6 site package tensorflow python p tls libcuda so 1 13084 try file home anaconda anaconda3 lib python3 6 site package tensorflow python solib local u local uconfig ucuda s scuda ccuda udriver uexternal slocal uconfig ucuda scuda scuda slib haswell x86 64 libcuda so 1 13084 try file home anaconda anaconda3 lib python3 6 site package tensorflow python solib local u local uconfig ucuda s scuda ccuda udriver uexternal slocal uconfig ucuda scuda scuda slib haswell libcuda so 1 13084 try file home anaconda anaconda3 lib python3 6 site package tensorflow python solib local u local uconfig ucuda s scuda ccuda udriver uexternal slocal uconfig ucuda scuda scuda slib x86 64 libcuda so 1 13084 try file home anaconda anaconda3 lib python3 6 site package tensorflow python solib local u local uconfig ucuda s scuda ccuda udriver uexternal slocal uconfig ucuda scuda scuda slib libcuda so 1 13084 try file home anaconda anaconda3 lib python3 6 site package tensorflow python solib local u local uconfig ucuda s scuda ccudnn uexternal slocal uconfig ucuda scuda scuda slib tls haswell x86 64 libcuda so 1 13084 try file home anaconda anaconda3 lib python3 6 site package tensorflow python solib local u local uconfig ucuda s scuda ccudnn uexternal slocal uconfig ucuda scuda scuda slib tls haswell libcuda so 1 13084 try file home anaconda anaconda3 lib python3 6 site package tensorflow python solib local u local uconfig ucuda s scuda ccudnn uexternal slocal uconfig ucuda scuda scuda slib tls x86 64 libcuda so 1 13084 try file home anaconda anaconda3 lib python3 6 site package tensorflow python solib local u local uconfig ucuda s scuda ccudnn uexternal slocal uconfig ucuda scuda scuda slib tls libcuda so 1 13084 try file home anaconda anaconda3 lib python3 6 site package tensorflow python solib local u local uconfig ucuda s scuda ccudnn uexternal slocal uconfig ucuda scuda scuda slib haswell x86 64 libcuda so 1 13084 try file home anaconda anaconda3 lib python3 6 site package tensorflow python solib local u local uconfig ucuda s scuda ccudnn uexternal slocal uconfig ucuda scuda scuda slib haswell libcuda so 1 13084 try file home anaconda anaconda3 lib python3 6 site package tensorflow python solib local u local uconfig ucuda s scuda ccudnn uexternal slocal uconfig ucuda scuda scuda slib x86 64 libcuda so 1 13084 try file home anaconda anaconda3 lib python3 6 site package tensorflow python solib local u local uconfig ucuda s scuda ccudnn uexternal slocal uconfig ucuda scuda scuda slib libcuda so 1 13084 try file home anaconda anaconda3 lib python3 6 site package tensorflow python solib local u local uconfig ucuda s scuda ccufft uexternal slocal uconfig ucuda scuda scuda slib tls haswell x86 64 libcuda so 1 13084 try file home anaconda anaconda3 lib python3 6 site package tensorflow python solib local u local uconfi 13084 try file home anaconda anaconda3 lib python3 6 site package tensorflow python solib local u local uconfig ucuda s scuda ccufft uexternal slocal uconfig ucuda scuda scuda slib libcuda so 1 13084 try file home anaconda anaconda3 lib python3 6 site package tensorflow python solib local u local uconfig ucuda s scuda ccurand uexternal slocal uconfig ucuda scuda scuda slib tls haswell x86 64 libcuda so 1 13084 try file home anaconda anaconda3 lib python3 6 site package tensorflow python solib local u local uconfig ucuda s scuda ccurand uexternal slocal uconfig ucuda scuda scuda slib tls haswell libcuda so 1 13084 try file home anaconda anaconda3 lib python3 6 site package tensorflow python solib local u local uconfig ucuda s scuda ccurand uexternal slocal uconfig ucuda scuda scuda slib tls x86 64 libcuda so 1 13084 try file home anaconda anaconda3 lib python3 6 site package tensorflow python solib local u local uconfig ucuda s scuda ccurand uexternal slocal uconfig ucuda scuda scuda slib tls libcuda so 1 13084 try file home anaconda anaconda3 lib python3 6 site package tensorflow python solib local u local uconfig ucuda s scuda ccurand uexternal slocal uconfig ucuda scuda scuda slib haswell x86 64 libcuda so 1 13084 try file home anaconda anaconda3 lib python3 6 site package tensorflow python solib local u local uconfig ucuda s scuda ccurand uexternal slocal uconfig ucuda scuda scuda slib haswell libcuda so 1 13084 try file home anaconda anaconda3 lib python3 6 site package tensorflow python solib local u local uconfig ucuda s scuda ccurand uexternal slocal uconfig ucuda scuda scuda slib x86 64 libcuda so 1 13084 try file home anaconda anaconda3 lib python3 6 site package tensorflow python solib local u local uconfig ucuda s scuda ccurand uexternal slocal uconfig ucuda scuda scuda slib libcuda so 1 13084 try file home anaconda anaconda3 lib python3 6 site package tensorflow python libcuda so 1 13084 try file local config cuda cuda lib64 tls haswell x86 64 libcuda so 1 13084 try file local config cuda cuda lib64 tls haswell libcuda so 1 13084 try file local config cuda cuda lib64 tls x86 64 libcuda so 1 13084 try file local config cuda cuda lib64 tls libcuda so 1 13084 try file local config cuda cuda lib64 haswell x86 64 libcuda so 1 13084 try file local config cuda cuda lib64 haswell libcuda so 1 13084 try file local config cuda cuda lib64 x86 64 libcuda so 1 13084 try file local config cuda cuda lib64 libcuda so 1 13084 try file local config cuda cu 13084 try file home anaconda anaconda3 lib python3 6 site package tensorflow python libcuda so 1 13084 try file home anaconda anaconda3 lib python3 6 site package tensorflow python libcuda so 1 13084 try file local config cuda cuda lib64 tls haswell x86 64 libcuda so 1 13084 try file local config cuda cuda lib64 tls haswell libcuda so 1 13084 try file local config cuda cuda lib64 tls x86 64 libcuda so 1 13084 try file local config cuda cuda lib64 tls libcuda so 1 13084 try file local config cuda cuda lib64 haswell x86 64 libcuda so 1 13084 try file local config cuda cuda lib64 haswell libcuda so 1 13084 try file local config cuda cuda lib64 x86 64 libcuda so 1 13084 try file local config cuda cuda lib64 libcuda so 1 13084 try file local config cuda cuda extras cupti lib64 tls haswell x86 64 libcuda so 1 13084 try file local config cuda cuda extras cupti lib64 tls haswell libcuda so 1 13084 try file local config cuda cuda extras cupti lib64 tls x86 64 libcuda so 1 13084 try file local config cuda cuda extras cupti lib64 tls libcuda so 1 13084 try file local config cuda cuda extras cupti lib64 haswell x86 64 libcuda so 1 13084 try file local config cuda cuda extras cupti lib64 haswell libcuda so 1 13084 try file local config cuda cuda extras cupti lib64 x86 64 libcuda so 1 13084 try file local config cuda cuda extras cupti lib64 libcuda so 1 13084 search path home anaconda anaconda3 bin lib rpath from file home anaconda anaconda3 bin python 13084 try file home anaconda anaconda3 bin lib libcuda so 1 13084 search path usr local cuda 9 0 lib64 tls haswell x86 64 tls haswell tls x86 64 tls haswell x86 64 haswell x86 64 ld library path 13084 try file usr local cuda 9 0 lib64 libcuda so 1 13084 try file tls haswell x86 64 libcuda so 1 13084 try file tls haswell libcuda so 1 13084 try file tls x86 64 libcuda so 1 13084 try file tls libcuda so 1 13084 try file haswell x86 64 libcuda so 1 13084 try file h lib64 local config cuda cuda extras cupti lib64 tls haswell x86 64 local config cuda cuda extras cupti lib64 tls haswell local config cuda cuda extras cupti lib64 tls x86 64 local config cuda cuda extr as cupti lib64 tls local config cuda cuda extras cupti lib64 haswell x86 64 local config cuda cuda extras cupti lib64 haswell local config cuda cuda extras cupti lib64 x86 64 local config cuda cuda extras cupti lib64 rpath from file home anaconda anaconda3 lib python3 6 site package tensorflow python pywrap tensorflow internal so 13084 try file home anaconda anaconda3 lib python3 6 site package tensorflow python libcudnn so 7 13084 try file local config cuda cuda lib64 tls haswell x86 64 libcudnn so 7 13084 try file local config cuda cuda lib64 tls haswell libcudnn so 7 13084 try file local config cuda cuda lib64 tls x86 64 libcudnn so 7 13084 try file local config cuda cuda lib64 tls libcudnn so 7 13084 try file local config cuda cuda lib64 haswell x86 64 libcudnn so 7 13084 try file local config cuda cuda lib64 haswell libcudnn so 7 13084 try file local config cuda cuda lib64 x86 64 libcudnn so 7 13084 try file local config cuda cuda lib64 libcudnn so 7 13084 try file local config cuda cuda extras cupti lib64 tls haswell x86 64 libcudnn so 7 13084 try file local config cuda cuda extras cupti lib64 tls haswell libcudnn so 7 13084 try file local config cuda cuda extras cupti lib64 tls x86 64 libcudnn so 7 13084 try file local config cuda cuda extras cupti lib64 tls libcudnn so 7 13084 try file local config cuda cuda extras cupti lib64 haswell x86 64 libcudnn so 7 13084 try file local config cuda cuda extras cupti lib64 haswell libcudnn so 7 13084 try file local config cuda cuda extras cupti lib64 x86 64 libcudnn so 7 13084 try file local config cuda cuda extras cupti lib64 libcudnn so 7 13084 search path home anaconda anaconda3 lib python3 6 site package tensorflow python home anaconda anaconda3 lib python3 6 site package tensorflow python local config cuda cuda lib64 tls haswell x86 64 local config cuda cuda lib64 tls haswell local config cuda cuda lib64 tls x86 64 local config cuda cuda lib64 tls local config cuda cuda lib64 haswell x86 64 local config cuda cuda lib64 haswell local config cuda cuda lib64 x86 64 local config cuda cuda lib64 local config cuda cuda extras cupti lib64 tls haswell x86 64 local config cuda cuda extras cupti lib64 tls haswell local config cuda cuda extras cupti lib64 tls x86 64 local config cuda cuda extras cupti lib64 tls local config cuda cuda extras cupti lib64 haswell x86 64 local config cuda cuda extras cupti lib64 haswell local config cuda cuda extras cupti lib64 x86 64 local config cuda cuda extras cupti lib64 rpath from file home anaconda anaconda3 lib python3 6 site package tensorflow python pywrap tensorflow internal so 13084 try file home anaconda anaconda3 lib python3 6 site package tensorflow python libcudnn so 7 13084 try file home anaconda anaconda3 lib python3 6 site package tensorflow python libcudnn 13084 try file home anaconda anaconda3 bin lib libcudnn so 7 13084 search path usr local cuda 9 0 lib64 tls haswell x86 64 tls haswell tls x86 64 tls haswell x86 64 haswell x86 64 ld library path 13084 try file usr local cuda 9 0 lib64 libcudnn so 7 13084 try file tls haswell x86 64 libcudnn so 7 13084 try file tls haswell libcudnn so 7 13084 try file tls x86 64 libcudnn so 7 13084 try file tls libcudnn so 7 13084 try file haswell x86 64 libcudnn so 7 13084 try file haswell libcudnn so 7 13084 try file x86 64 libcudnn so 7 13084 try file libcudnn so 7 13084 search cache etc ld so cache 13084 try file usr lib x86 64 linux gnu libcudnn so 7 13084 13084 find library libcufft so 9 0 0 search 13084 search path home anaconda anaconda3 lib python3 6 site package tensorflow python local config cuda cuda lib64 tls haswell x86 64 local config cuda cuda lib64 tls haswell local config cuda cuda lib64 tls x86 64 local config cuda cuda lib64 tls local config cuda cuda lib64 haswell x86 64 local config cuda cuda lib64 haswell local config cuda cuda lib64 x86 64 local config cuda cuda lib64 local config cuda cuda extras cupti lib64 tls haswell x86 64 local config cuda cuda extras cupti lib64 tls haswell local config cuda cuda extras cupti lib64 tls x86 64 local config cuda cuda extras cupti lib64 tls local config cuda cuda extras cupti lib64 haswell x86 64 local config cuda cuda extras cupti lib64 haswell local config cuda cuda extras cupti lib64 x86 64 local config cuda cuda extras cupti lib64 rpath from file home anaconda anaconda3 lib python3 6 site package tensorflow python pywrap tensorflow internal so 13084 try file home anaconda anaconda3 lib python3 6 site package tensorflow python libcufft so 9 0 13084 try file local config cuda cuda lib64 tls haswell x86 64 libcufft so 9 0 13084 try file local config cuda cuda lib64 tls haswell libcufft so 9 0 13084 try file local config cuda cuda lib64 tls x86 64 libcufft so 9 0 13084 try file local config cuda cuda lib64 tls libcufft so 9 0 13084 try file local config cuda cuda lib64 haswell x86 64 libcufft so 9 0 13084 try file local config cuda cuda lib64 haswell libcufft so 9 0 13084 try file local config cu ell x86 64 local config cuda cuda lib64 tls haswell local config cuda cuda lib64 tls x86 64 local config cuda cuda lib64 tls local config cuda cuda lib64 haswell x86 64 local config cuda cuda lib64 h aswell local config cuda cuda lib64 x86 64 local config cuda cuda lib64 local config cuda cuda extras cupti lib64 tls haswell x86 64 local config cuda cuda extras cupti lib64 tls haswell local config cuda cuda extras cupti lib64 tls x86 64 local config cuda cuda extras cupti lib64 tls local config cuda cuda extras cupti lib64 haswell x86 64 local config cuda cuda extras cupti lib64 haswell local config cuda cuda extras cupti lib64 x86 64 local config cuda cuda extras cupti lib64 rpath from file home anaconda anaconda3 lib python3 6 site package tensorflow python pywrap tensorflow internal so 13084 try file home anaconda anaconda3 lib python3 6 site package tensorflow python libcufft so 9 0 13084 try file home anaconda anaconda3 lib python3 6 site package tensorflow python libcufft so 9 0 13084 try file local config cuda cuda lib64 tls haswell x86 64 libcufft so 9 0 13084 try file local config cuda cuda lib64 tls haswell libcufft so 9 0 13084 try file local config cuda cuda lib64 tls x86 64 libcufft so 9 0 13084 try file local config cuda cuda lib64 tls libcufft so 9 0 13084 try file local config cuda cuda lib64 haswell x86 64 libcufft so 9 0 13084 try file local config cuda cuda lib64 haswell libcufft so 9 0 13084 try file local config cuda cuda lib64 x86 64 libcufft so 9 0 13084 try file local config cuda cuda lib64 libcufft so 9 0 13084 try file local config cuda cuda extras cupti lib64 tls haswell x86 64 libcufft so 9 0 13084 try file local config cuda cuda extras cupti lib64 tls haswell libcufft so 9 0 13084 try file local config cuda cuda extras cupti lib64 tls x86 64 libcufft so 9 0 13084 try file local config cuda cuda extras cupti lib64 tls libcufft so 9 0 13084 try file local config cuda cuda extras cupti lib64 haswell x86 64 libcufft so 9 0 13084 try file local config cuda cuda extras cupti lib64 haswell libcufft so 9 0 13084 try file local config cuda cuda extras cupti lib64 x86 64 libcufft so 9 0 13084 try file local config cuda cuda extras cupti lib64 libcufft so 9 0 13084 search path home anaconda anaconda3 bin lib rpath from file home anaconda anaconda3 bin python 13084 try file home anaconda anaconda3 bin lib libcufft so 9 0 13084 13084 find library libcurand so 9 0 0 search 13084 search path home anaconda anaconda3 lib python3 6 site package tensorflow python local config cuda cuda lib64 tls haswell x86 64 local config cuda cuda lib64 tls haswell local config cuda cuda lib64 tls x86 64 local config cuda cuda lib64 tls local config cuda cuda lib64 haswell x86 64 local config cuda cuda lib64 haswell local config cuda cuda lib64 x86 64 local config cuda cuda lib64 local config cuda cuda extras cupti lib64 tls haswell x86 64 local config cuda cuda extras cupti lib64 tls haswell local config cuda cuda extras cupti lib64 tls x86 64 local config cuda cuda extras cupti lib64 tls local config cuda cuda extras cupti lib64 haswell x86 64 local config cuda cuda extras cupti lib64 haswell local config cuda cuda extras cupti lib64 x86 64 local config cuda cuda extras cupti lib64 rpath from file home anaconda anaconda3 lib python3 6 site package tensorflow python pywrap tensorflow internal so 13084 try file home anaconda anaconda3 lib python3 6 site package tensorflow python libcurand so 9 0 13084 try file local config cuda cuda lib64 tls haswell x86 64 libcurand so 9 0 13084 try file local config cuda cuda lib64 tls haswell libcurand so 9 0 13084 try file local config cuda cuda lib64 tls x86 64 libcurand so 9 0 13084 try file local config cuda cuda lib64 tls libcurand so 9 0 13084 try file local config cuda cuda lib64 haswell x86 64 libcurand so 9 0 13084 try file local config cuda cuda lib64 haswell libcurand so 9 0 13084 try file local config cuda cuda lib64 x86 64 libcurand so 9 0 13084 try file local config cuda cuda lib64 libcurand so 9 0 13084 try file local config cuda cuda extras cupti lib64 tls haswell x86 64 libcurand so 9 0 13084 try file local config cuda cuda extras cupti lib64 tls haswell libcurand so 9 0 13084 try file local config cuda cuda extras cupti lib64 tls x86 64 libcurand so 9 0 13084 try file local config cuda cuda extras cupti lib64 tls libcurand so 9 0 13084 tryi nfig cuda cuda extras cupti lib64 x86 64 local config cuda cuda extras cupti lib64 rpath from file home anaconda anaconda3 lib python3 6 site package tensorflow python pywrap tensorflow internal so 13084 try file home anaconda anaconda3 lib python3 6 site package tensorflow python libcurand so 9 0 13084 try file home anaconda anaconda3 lib python3 6 site package tensorflow python libcurand so 9 0 13084 try file local config cuda cuda lib64 tls haswell x86 64 libcurand so 9 0 13084 try file local config cuda cuda lib64 tls haswell libcurand so 9 0 13084 try file local config cuda cuda lib64 tls x86 64 libcurand so 9 0 13084 try file local config cuda cuda lib64 tls libcurand so 9 0 13084 try file local config cuda cuda lib64 haswell x86 64 libcurand so 9 0 13084 try file local config cuda cuda lib64 haswell libcurand so 9 0 13084 try file local config cuda cuda lib64 x86 64 libcurand so 9 0 13084 try file local config cuda cuda lib64 libcurand so 9 0 13084 try file local config cuda cuda extras cupti lib64 tls haswell x86 64 libcurand so 9 0 13084 try file local config cuda cuda extras cupti lib64 tls haswell libcurand so 9 0 13084 try file local config cuda cuda extras cupti lib64 tls x86 64 libcurand so 9 0 13084 try file local config cuda cuda extras cupti lib64 tls libcurand so 9 0 13084 try file local config cuda cuda extras cupti lib64 haswell x86 64 libcurand so 9 0 13084 try file local config cuda cuda extras cupti lib64 haswell libcurand so 9 0 13084 try file local config cuda cuda extras cupti lib64 x86 64 libcurand so 9 0 13084 try file local config cuda cuda extras cupti lib64 libcurand so 9 0 13084 search path home anaconda anaconda3 bin lib rpath from file home anaconda anaconda3 bin python 13084 try file home anaconda anaconda3 bin lib libcurand so 9 0 13084 13084 find library libnvidia fatbinaryloader so 390 116 0 search 13084 search path home anaconda anaconda3 lib python3 6 site package tensorflow python local config cuda cuda lib64 tls haswell x86 64 local config cuda cuda lib64 tls haswell local config cuda cuda lib64 tls x86 64 local config cuda cuda lib64 tls local config cuda cuda lib64 haswell x86 64 local config cuda cuda lib64 haswell local config cuda cuda lib64 x86 64 local config cuda cuda lib64 local config cuda cuda extras cupti lib64 tls haswell x86 64 local config cuda cuda extras cupti lib64 tls haswell local config cuda cuda extras cupti lib64 tls x86 64 local config cuda cuda extras cupti lib64 tls local config 13084 try file local config cuda cuda lib64 tls haswell libnvidia fatbinaryloader so 390 116 13084 try file local config cuda cuda lib64 tls x86 64 libnvidia fatbinaryloader so 390 116 13084 try file local config cuda cuda lib64 tls libnvidia fatbinaryloader so 390 116 13084 try file local config cuda cuda lib64 haswell x86 64 libnvidia fatbinaryloader so 390 116 13084 try file local config cuda cuda lib64 haswell libnvidia fatbinaryloader so 390 116 13084 try file local config cuda cuda lib64 x86 64 libnvidia fatbinaryloader so 390 116 13084 try file local config cuda cuda lib64 libnvidia fatbinaryloader so 390 116 13084 try file local config cuda cuda extras cupti lib64 tls haswell x86 64 libnvidia fatbinaryloader so 390 116 13084 try file local config cuda cuda extras cupti lib64 tls haswell libnvidia fatbinaryloader so 390 116 13084 try file local config cuda cuda extras cupti lib64 tls x86 64 libnvidia fatbinaryloader so 390 116 13084 try file local config cuda cuda extras cupti lib64 tls libnvidia fatbinaryloader so 390 116 13084 try file local config cuda cuda extras cupti lib64 haswell x86 64 libnvidia fatbinaryloader so 390 116 13084 try file local config cuda cuda extras cupti lib64 haswell libnvidia fatbinaryloader so 390 116 13084 try file local config cuda cuda extras cupti lib64 x86 64 libnvidia fatbinaryloader so 390 116 13084 try file local config cuda cuda extras cupti lib64 libnvidia fatbinaryloader so 390 116 13084 search path home anaconda anaconda3 lib python3 6 site package tensorflow python home anaconda anaconda3 lib python3 6 site package tensorflow python local config cuda cuda lib64 tls haswell x86 64 local config cuda cuda lib64 tls haswell local config cuda cuda lib64 tls x86 64 local config cuda cuda lib64 tls local config cuda cuda lib64 haswell x86 64 local config cuda cuda lib64 haswell local config cuda cuda lib64 x86 64 local config cuda cuda lib64 local config cuda cuda extras cupti lib64 tls haswell x86 64 local config cuda cuda extras cupti lib64 tls haswell local config cuda cuda extras cupti lib64 tls x86 64 local config cuda cuda extras cupti lib64 tls local config cuda cuda extras cupti lib64 haswell x86 64 local config cuda cuda extras cupti lib64 haswell local config cuda cuda extras cupti lib64 x86 64 local config cuda cuda extras cupti lib64 rpath from file home anaconda anaconda3 lib python3 6 site package tensorflow python pywrap tensorflow internal so 13084 try file home anaconda anaconda3 lib python3 6 site package tensorflow python libnvidia fatbinaryloader so 390 116 13084 try file home anaconda anaconda3 lib python3 6 site package tensorflow python libnvidia fatbinaryloader so 390 116 13084 try file local config cuda cuda lib64 tls haswell x86 64 libnvidia fatbinaryloader so 390 116 13084 try file local config cuda cuda lib64 tls haswell libnvidia fatbinaryloader so 390 116 13084 try file local config cuda cuda lib64 tls x86 64 libnvidia fatbinaryloader so 390 116 13084 try file local config cuda cuda lib64 tls libnvidia fatbinaryloader so 390 116 13084 try file local config cuda cuda lib64 haswell x86 64 libnvidia fat 13084 call init home anaconda anaconda3 bin lib libcufft so 9 0 13084 13084 13084 call init usr lib x86 64 linux gnu libcudnn so 7 13084 13084 13084 call init usr lib x86 64 linux gnu libcuda so 1 13084 13084 13084 call init home anaconda anaconda3 bin lib libgomp so 1 13084 13084 13084 call init home anaconda anaconda3 bin lib libcudart so 9 0 13084 13084 13084 call init home anaconda anaconda3 bin lib libcusolver so 9 0 13084 13084 13084 call init home anaconda anaconda3 bin lib libcublas so 9 0 13084 13084 13084 call init home anaconda anaconda3 lib python3 6 site package tensorflow python libtensorflow framework so 13084 13084 13084 call init home anaconda anaconda3 lib python3 6 site package tensorflow python pywrap tensorflow internal so 13084 13084 find library libhdfs so 0 search 13084 search path home anaconda anaconda3 lib python3 6 site package tensorflow python 13084 try file local config cuda cuda lib64 tls x86 64 libhdfs so 13084 try file local config cuda cuda lib64 tls libhdfs so 13084 try file local config cuda cuda lib64 haswell x86 64 libhdfs so 13084 try file local config cuda cuda lib64 haswell libhdfs so 13084 try file local config cuda cuda lib64 x86 64 libhdfs so 13084 try file local config cuda cuda lib64 libhdfs so 13084 try file local config cuda cuda extras cupti lib64 tls haswell x86 64 libhdfs so 13084 try file local config cuda cuda extras cupti lib64 tls haswell libhdfs so 13084 try file local config cuda cuda extras cupti lib64 tls x86 64 libhdfs so 13084 try file local config cuda cuda extras cupti lib64 tls libhdfs so 13084 try file local config cuda cuda extras cupti lib64 haswell x86 64 libhdfs so 13084 try file local config cuda cuda extras cupti lib64 haswell libhdfs so 13084 try file local config cuda cuda extras cupti lib64 x86 64 libhdfs so 13084 try file local config cuda cuda extras cupti lib64 libhdfs so 13084 search path home anaconda anaconda3 lib python3 6 site package tensorflow python home anaconda anaconda3 lib python3 6 site package tensorflow python local config cuda cuda lib64 tls haswell x86 64 local config cuda cuda lib64 tls haswell local config cuda cuda lib64 tls x86 64 local config cuda cuda lib64 tls local config cuda cuda lib64 haswell x86 64 local config cuda cuda lib64 haswell local config cuda cuda lib64 x86 64 local config cuda cuda lib64 local config cuda cuda extras cupti lib64 tls haswell x86 64 local config cuda cuda extras cupti lib64 tls haswell local config cuda cuda extras cupti lib64 tls x86 64 local config cuda cuda extras cupti lib64 tls local config cuda cuda extras cupti lib64 haswell x86 64 local config cuda cuda extras cupti lib64 haswell local config cuda cuda extras cupti lib64 x86 64 local config cuda cuda extras cupti lib64 rpath from file home anaconda anaconda3 lib python3 6 site package tensorflow python pywrap tensorflow internal so 13084 try file home anaconda anaconda3 lib python3 6 site package tensorflow python libhdfs so 13084 try file home anaconda anaconda3 lib python3 6 site package tensorflow python libhdfs so 13084 try file local config cuda cuda lib64 tls haswell x86 64 libhdfs so 13084 try file local config cuda cuda lib64 tls haswell libhdfs so 13084 try file local config cuda cuda lib64 tls x86 64 libhdfs so 13084 try file local config cuda cuda lib64 tls libhdfs so 13084 try file local config cuda cuda lib64 haswell x86 64 libhdfs so 13084 try file 13084 search path lib x86 64 linux gnu tls haswell x86 64 lib x86 64 linux gnu tls haswell lib x86 64 linux gnu tls x86 64 lib x86 64 linux gnu tls lib x86 64 linux gnu haswell x86 64 lib x86 64 linux gnu haswell lib x86 64 linux gnu x86 64 lib x86 64 linux gnu usr lib x86 64 linux gnu tls haswell x86 64 usr lib x86 64 linux gnu tls haswell usr lib x86 64 linux gnu tls x86 64 usr lib x86 64 linux gnu tls usr lib x86 64 linux gnu haswell x86 64 usr lib x86 64 linux gnu haswell usr lib x86 64 linux gnu x86 64 usr lib x86 64 linux gnu lib tls haswell x86 64 lib tls haswell lib tls x86 64 lib tls lib haswell x86 64 lib haswell lib x86 64 lib usr lib tls haswell x86 64 usr lib tls haswell usr lib tls x86 64 usr lib tls usr lib haswell x86 64 usr lib haswell usr lib x86 64 usr lib system search path 13084 try file lib x86 64 linux gnu tls haswell x86 64 libhdfs so 13084 try file lib x86 64 linux gnu tls haswell libhdfs so 13084 try file lib x86 64 linux gnu tls x86 64 libhdfs so 13084 try file lib x86 64 linux gnu tls libhdfs so 13084 try file lib x86 64 linux gnu haswell x86 64 libhdfs so 13084 try file lib x86 64 linux gnu haswell libhdfs so 13084 try file lib x86 64 linux gnu x86 64 libhdfs so 13084 try file lib x86 64 linux gnu libhdfs so 13084 try file usr lib x86 64 linux gnu tls haswell x86 64 libhdfs so 13084 try file usr lib x86 64 linux gnu tls haswell libhdfs so 13084 try file usr lib x86 64 linux gnu tls x86 64 libhdfs so 13084 try file usr lib x86 64 linux gnu tls libhdfs so 13084 try file usr lib x86 64 linux gnu haswell x86 64 libhdfs so 13084 try file usr lib x86 64 linux gnu haswell libhdfs so 13084 try file usr lib x86 64 linux gnu x86 64 libhdfs so 13084 try file usr lib x86 64 linux gnu libhdfs so 13084 try file lib tls haswell x86 64 libhdfs so 13084 try file lib tls haswell libhdfs so 13084 try file lib tls x86 64 libhdfs so 13084 try file lib tls libhdfs so 13084 try file lib haswell x86 64 libhdfs so 13084 try file lib haswell libhdfs so 13084 13084 call init home anaconda anaconda3 lib python3 6 lib dynload json cpython 36 m x86 64 linux gnu so 13084 13084 13084 call init home anaconda anaconda3 lib python3 6 lib dynload array cpython 36 m x86 64 linux gnu so 13084 13084 find library libssl so 1 1 0 search 13084 search path home anaconda anaconda3 lib python3 6 lib dynload rpath from file home anaconda anaconda3 lib python3 6 lib dynload ctype cpython 36 m x86 64 linux gnu so 13084 try file home anaconda anaconda3 lib python3 6 lib dynload libssl so 1 1 13084 13084 13084 call init home anaconda anaconda3 lib python3 6 lib dynload libssl so 1 1 13084 13084 13084 call init home anaconda anaconda3 lib python3 6 lib dynload ssl cpython 36 m x86 64 linux gnu so 13084 13084 find library libhdf5 5773eb11 so 103 0 0 0 search 13084 search path usr local cuda 9 0 lib64 tls haswell x86 64 tls haswell tls x86 64 tls haswell x86 64 haswell x86 64 ld library path 13084 try file usr local cuda 9 0 lib64 libhdf5 5773eb11 so 103 0 0 13084 try file tls haswell x86 64 libhdf5 5773eb11 so 103 0 0 13084 try file tls haswell libhdf5 5773eb11 so 103 0 0 13084 try file tls x86 64 libhdf5 5773eb11 so 103 0 0 13084 try file tls libhdf5 5773eb11 so 103 0 0 13084 try file haswell x86 64 libhdf5 5773eb11 so 103 0 0 13084 try file haswell libhdf5 5773eb11 so 103 0 0 13084 try file x86 64 libhdf5 5773eb11 so 103 0 0 13084 try file libhdf5 5773eb11 so 103 0 0 13084 search path home anaconda anaconda3 lib python3 6 site package h5py lib tls haswell x86 64 home anaconda anaconda3 lib python3 6 site package h5py lib tls haswell home anaconda anaconda3 lib python3 6 site package h5py lib tls x86 64 home anaconda anaconda3 l 13084 try file home anaconda anaconda3 lib python3 6 site package h5py lib x86 64 libhdf5 5773eb11 so 103 0 0 13084 try file home anaconda anaconda3 lib python3 6 site package h5py lib libhdf5 5773eb11 so 103 0 0 13084 13084 find library libhdf5 hl db841637 so 100 1 1 0 search 13084 search path usr local cuda 9 0 lib64 tls haswell x86 64 tls haswell tls x86 64 tls haswell x86 64 haswell x86 64 ld library path 13084 try file usr local cuda 9 0 lib64 libhdf5 hl db841637 so 100 1 1 13084 try file tls haswell x86 64 libhdf5 hl db841637 so 100 1 1 13084 try file tls haswell libhdf5 hl db841637 so 100 1 1 13084 try file tls x86 64 libhdf5 hl db841637 so 100 1 1 13084 try file tls libhdf5 hl db841637 so 100 1 1 13084 try file haswell x86 64 libhdf5 hl db841637 so 100 1 1 13084 try file haswell libhdf5 hl db841637 so 100 1 1 13084 try file x86 64 libhdf5 hl db841637 so 100 1 1 13084 try file libhdf5 hl db841637 so 100 1 1 13084 search path home anaconda anaconda3 lib python3 6 site package h5py lib runpath from file home anaconda anaconda3 lib python3 6 site package h5py error cpython 36 m x86 64 linux gnu so 13084 try file home anaconda anaconda3 lib python3 6 site package h5py lib libhdf5 hl db841637 so 100 1 1 13084 13084 find library libsz 1c7dd0cf so 2 0 1 0 search 13084 search path home anaconda anaconda3 lib python3 6 site package h5py lib tls haswell x86 64 home anaconda anaconda3 lib python3 6 site package h5py lib tls haswell home anaconda anaconda3 lib python3 6 site package h5py lib tls x86 64 home anaconda anaconda3 lib python3 6 site package h5py lib tls home anaconda anaconda3 lib python3 6 site package h5py lib haswell x86 64 home anaconda anaconda3 lib python3 6 site package h5py lib haswell home anaconda anaconda3 lib python3 6 site package h5py lib x86 64 home anaconda anaconda3 lib python3 6 site package h5py lib rpath from file home anaconda anaconda3 lib python3 6 site package h5py lib libhdf5 5773eb11 so 103 0 0 13084 try file home anaconda anaconda3 lib python3 6 site package h5py lib tls haswell x86 64 libsz 1c7dd0cf so 2 0 1 13084 try file home anaconda anaconda3 lib python3 6 site package h5py lib tls haswell libsz 1c7dd0cf so 2 0 1 13084 try file home anaconda anaconda3 lib python3 6 site package h5py lib tls x86 64 libsz 1c7dd0cf so 2 0 1 13084 try file home anaconda anaconda3 lib python3 6 site package h5py lib tls libsz 1c7dd0cf so 2 0 1 13084 try file home anaconda anaconda3 lib python3 6 site package h5py lib hasw 13084 13084 call init home anaconda anaconda3 lib python3 6 site package h5py h5r cpython 36 m x86 64 linux gnu so 13084 13084 13084 call init home anaconda anaconda3 lib python3 6 site package h5py object cpython 36 m x86 64 linux gnu so 13084 13084 13084 call init home anaconda anaconda3 lib python3 6 site package h5py def cpython 36 m x86 64 linux gnu so 13084 13084 13084 call init home anaconda anaconda3 lib python3 6 site package h5py h5 t cpython 36 m x86 64 linux gnu so 13084 13084 13084 call init home anaconda anaconda3 lib python3 6 site package h5py util cpython 36 m x86 64 linux gnu so 13084 13084 13084 call init home anaconda anaconda3 lib python3 6 site package h5py h5 cpython 36 m x86 64 linux gnu so 13084 13092 find library libc so 6 0 search 13092 search path usr local cuda 9 0 lib64 tls haswell x86 64 usr local cuda 9 0 lib64 tls haswell usr local cuda 9 0 lib64 tls x86 64 usr local cuda 9 0 lib64 tls usr local cuda 9 0 lib64 haswell x86 64 usr local cuda 9 0 lib64 haswell usr local cuda 9 0 lib64 x86 64 usr local cuda 9 0 lib64 tls haswell x86 64 tls haswell tls x86 64 tls haswell x86 64 haswell x86 64 ld library path 13092 try file usr local cuda 9 0 lib64 tls haswell x86 64 libc so 6 13092 try file usr local cuda 9 0 lib64 tls haswell libc so 6 13092 try file usr local cuda 9 0 lib64 tls x86 64 libc so 6 13092 try file usr local cuda 9 0 lib64 tls libc so 6 13092 try file usr local cuda 9 0 lib64 haswell x86 64 libc so 6 13092 try file usr local cuda 9 0 lib64 haswell libc so 6 13092 try file usr local cud 13084 call init home anaconda anaconda3 lib python3 6 site package h5py h5 g cpython 36 m x86 64 linux gnu so 13084 13084 13084 call init home anaconda anaconda3 lib python3 6 site package h5py h5i cpython 36 m x86 64 linux gnu so 13084 13084 13084 call init home anaconda anaconda3 lib python3 6 site package h5py h5fd cpython 36 m x86 64 linux gnu so 13084 13084 13084 call init home anaconda anaconda3 lib python3 6 site package h5py h5o cpython 36 m x86 64 linux gnu so 13084 13084 13084 call init home anaconda anaconda3 lib python3 6 site package h5py h5l cpython 36 m x86 64 linux gnu so 13084 13084 13084 call init home anaconda anaconda3 lib libuuid so 1 13084 13084 13084 call init home anaconda anaconda3 lib python3 6 lib dynload unicodedata cpython 36 m x86 64 linux gnu so 13084 13084 13084 call init home anaconda anaconda3 lib python3 6 site package cryptography hazmat binding constant time abi3 so 13084 13084 find library libffi bce22613 so 6 0 4 0 search 13084 search path home anaconda anaconda3 lib python3 6 site package lib cffi backend tls haswell x86 64 home anaconda anaconda3 lib python3 6 site package lib cffi backend tls haswell home anaconda anaconda3 lib python3 6 site package lib cffi backend tls x86 64 home anaconda anaconda3 lib python3 6 site package lib cffi backend tls home anaconda anaconda3 lib python3 6 site package lib cffi backend haswell x86 64 home anaconda anaconda3 lib python3 6 site package lib cffi backend haswell home anaconda anaconda3 lib python3 6 site package lib cffi backend x86 64 home anaconda anaconda3 lib python3 6 site package lib cffi backend rpath from file home anaconda anaconda3 l 13084 13084 13084 call init home anaconda anaconda3 lib python3 6 site package scipy sparse csparsetool cpython 36 m x86 64 linux gnu so 13084 13084 13084 call init home anaconda anaconda3 lib python3 6 site package scipy sparse csgraph short path cpython 36 m x86 64 linux gnu so 13084 13084 13084 call init home anaconda anaconda3 lib python3 6 site package scipy sparse csgraph tool cpython 36 m x86 64 linux gnu so 13084 13084 13084 call init home anaconda anaconda3 lib python3 6 site package scipy sparse csgraph traversal cpython 36 m x86 64 linux gnu so 13084 13084 13084 call init home anaconda anaconda3 lib python3 6 site package scipy sparse csgraph min span tree cpython 36 m x86 64 linux gnu so 13084 13084 13084 call init home anaconda anaconda3 lib python3 6 site package scipy sparse csgraph reorder cpython 36 m x86 64 linux gnu so 13084 13084 find library libyaml 0 so 2 0 search 13084 search path home anaconda anaconda3 lib python3 6 site package tls haswell x86 64 home anaconda anaconda3 lib python3 6 site package tls haswell home anaconda anaconda3 lib python3 6 site package tls x86 64 home anaconda anaconda3 lib python3 6 site package tls home anaconda anaconda3 lib python3 6 site package haswell x86 64 home anaconda anaconda3 lib python3 6 site package haswell home anaconda anaconda3 lib python3 6 site package x86 64 home anaconda anaconda3 lib python3 6 site package rpath from file home anaconda anaconda3 lib python3 6 site package yaml cpython 36 m x86 64 linux gnu so 13084 try file home anaconda anaconda3 lib python3 6 site package tls haswell x86 64 libyaml 0 so 2 13084 try file home anaconda anaconda3 lib python3 6 site package tls haswell libyaml 0 so 2 13084 try file home anaconda anaconda3 lib python3 6 site package tls x86 64 libyaml 0 so 2 13084 try file 13084 try file home anaconda anaconda3 lib python3 6 site package pil tls haswell x86 64 libjpeg so 9 13084 try file home anaconda anaconda3 lib python3 6 site package pil tls haswell libjpeg so 9 13084 try file home anaconda anaconda3 lib python3 6 site package pil tls x86 64 libjpeg so 9 13084 try file home anaconda anaconda3 lib python3 6 site package pil tls libjpeg so 9 13084 try file home anaconda anaconda3 lib python3 6 site package pil haswell x86 64 libjpeg so 9 13084 try file home anaconda anaconda3 lib python3 6 site package pil haswell libjpeg so 9 13084 try file home anaconda anaconda3 lib python3 6 site package pil x86 64 libjpeg so 9 13084 try file home anaconda anaconda3 lib python3 6 site package pil libjpeg so 9 13084 13084 find library libtiff so 5 0 search 13084 search path home anaconda anaconda3 lib python3 6 site package pil rpath from file home anaconda anaconda3 lib python3 6 site package pil image cpython 36 m x86 64 linux gnu so 13084 try file home anaconda anaconda3 lib python3 6 site package pil libtiff so 5 13084 13084 13084 call init home anaconda anaconda3 lib python3 6 site package pil libjpeg so 9 13084 13084 13084 call init home anaconda anaconda3 lib python3 6 site package pil libtiff so 5 13084 13084 13084 call init home anaconda anaconda3 lib python3 6 site package pil image cpython 36 m x86 64 linux gnu so 13084 13084 find library libmkl rt so 0 search 13084 search path home anaconda anaconda3 lib python3 6 site package scipy linalg tls haswell x86 64 home anaconda anaconda3 lib python3 6 site package scipy linalg tls haswell home anaconda anaconda3 lib python3 6 site package scipy linalg tls x86 64 home anaconda anaconda3 lib python3 6 site package scipy linalg tls home anaconda anaconda3 lib python3 6 site package scipy linalg haswell x86 64 home anaconda anaconda3 lib python3 6 site package scipy linalg haswell home anaconda anaconda3 lib python3 6 site package scipy linalg x86 64 home anaconda anaconda3 lib python3 6 site packag 13084 13084 call init home anaconda anaconda3 lib python3 6 site package scipy linalg flinalg cpython 36 m x86 64 linux gnu so 13084 13084 13084 call init home anaconda anaconda3 lib python3 6 site package scipy linalg solve toeplitz cpython 36 m x86 64 linux gnu so 13084 13084 13084 call init home anaconda anaconda3 lib python3 6 site package scipy linalg decomp update cpython 36 m x86 64 linux gnu so 13084 13084 13084 call init home anaconda anaconda3 lib python3 6 site package scipy linalg cython blas cpython 36 m x86 64 linux gnu so 13084 13084 13084 call init home anaconda anaconda3 lib python3 6 site package scipy linalg cython lapack cpython 36 m x86 64 linux gnu so 13084 13084 13084 call init home anaconda anaconda3 lib python3 6 site package scipy ndimage nd image cpython 36 m x86 64 linux gnu so 13084 13084 find library libgfortran so 4 0 search 13084 search path home anaconda anaconda3 lib python3 6 site package scipy special tls haswell x86 64 home anaconda anaconda3 lib python3 6 site package scipy special tls haswell home anaconda anaconda3 lib python3 6 site package scipy special tls x86 64 home anaconda anaconda3 lib python3 6 site package scipy special tls home anaconda anaconda3 lib python3 6 site package scipy special haswell x86 64 home anaconda anaconda3 lib python3 6 site package scipy special haswell home anaconda anaconda3 lib python3 6 site package scipy special x86 64 home anaconda anaconda3 lib python3 6 site package scipy special rpath from file home anaconda anaconda3 lib python3 6 site package scipy special ufunc cpython 36 m x86 64 linux gnu so 13084 try file home anaconda anaconda3 lib python3 6 site package scipy special tls haswell x86 64 libgfortran so 4 13084 try file home anaconda anaconda3 lib python3 6 site package scipy special tls haswell libgfortran so 4 13084 try file home anaconda anaconda3 lib python3 6 site package scipy special tls x86 64 libgfortran so 4 13084 try file home anaconda anaconda3 lib python3 6 site package scipy special tls libgfortran so 4 13084 try file home anaconda anaconda3 lib python3 6 site package scipy special haswell x86 64 libgfortran so 4 13084 try file home anaconda anaconda3 lib python3 6 site package scipy special haswell libgfortran so 4 13084 try file home anaconda anaconda3 lib python3 6 site package scipy special x86 64 libgfortran so 4 13084 try file home anaconda anaconda3 lib python3 6 site package scipy special libgfortran so 4 13084 13084 find library libquadmath so 0 0 search 13084 search path home rdonnelly mc conda bld compilers linux 64 1534627447954 work gcc build x86 64 conda cos6 linux gnu lib lib tls haswell x86 64 home rdonnelly mc conda bld compilers linux 64 1534627447954 work gcc build x86 64 conda cos6 linux gnu lib lib tls haswell home rdonnelly mc conda bld compilers linux 64 1534627447954 work gcc build x86 64 conda cos6 linux gnu lib lib tls x86 64 home rdonnelly mc conda bld compilers linux 64 1534627447954 work gcc build x86 64 conda cos6 linux gnu lib lib tls home rdonnelly mc conda bld compilers linux 64 1534627447954 work gcc build x86 64 conda cos6 linux gnu lib lib haswell x86 64 home rdonnelly mc conda bld compilers linux 64 1534627447954 work gcc build x86 64 conda cos6 linux gnu lib lib haswell home rdonnelly mc conda bld compilers linux 64 1534627447954 work gcc build x86 64 conda cos6 linux gnu lib lib x86 64 home rdonnelly mc conda bld compilers linux 64 1534627447954 work gcc build x86 64 conda cos6 linux gnu lib lib rpath from file home anaconda anaconda3 lib python3 6 site package scipy special libgfortran so 4 13084 try file home rdonnelly mc conda bld compilers linux 64 1534627447954 work gcc build x86 64 conda cos6 linux gnu lib lib tls haswell x86 64 libquadmath so 0 13084 try file home rdonnelly mc conda bld compilers linux 64 1534627447954 work gcc build x86 64 conda cos6 linux gnu lib lib tls haswell libquadmath so 0 13084 try file home rdonnelly mc conda bld compilers linux 64 1534627447954 work gcc build x86 64 conda cos6 linux gnu lib lib tls x86 64 libquadmath so 0 13084 try file home rdonnelly mc conda bld compilers linux 64 1534627447954 work gcc build x86 64 conda cos6 linux gnu lib lib tls libquadmath so 0 13084 try file home rdonnelly mc conda bld compilers linux 64 1534627447954 work gcc build x86 64 conda cos6 linux gnu lib lib haswell x86 64 libquadmath so 0 13084 try 13084 call init home anaconda anaconda3 lib python3 6 site package scipy special ellip harm 2 cpython 36 m x86 64 linux gnu so 13084 13084 13084 call init home anaconda anaconda3 lib python3 6 site package scipy interpolate fitpack cpython 36 m x86 64 linux gnu so 13084 13084 13084 call init home anaconda anaconda3 lib python3 6 site package scipy interpolate dfitpack cpython 36 m x86 64 linux gnu so 13084 13084 13084 call init home anaconda anaconda3 lib python3 6 site package scipy interpolate bspl cpython 36 m x86 64 linux gnu so 13084 13084 13084 call init home anaconda anaconda3 lib python3 6 site package scipy interpolate ppoly cpython 36 m x86 64 linux gnu so 13084 13084 13084 call init home anaconda anaconda3 lib python3 6 site package scipy interpolate interpnd cpython 36 m x86 64 linux gnu so 13084 13084 13084 call init home anaconda anaconda3 lib python3 6 site package scipy spatial ckdtree cpython 36 m x86 64 linux gnu so 13084 13084 13084 call init home anaconda anaconda3 lib python3 6 site package scipy spatial qhull cpython 36 m x86 64 linux gnu so 13084 13084 13084 call init home anaconda anaconda3 lib python3 6 site package scipy lib messagestream cpython 36 m x86 64 linux gnu so 13084 13084 13084 13084 13084 13084 call init home anaconda anaconda3 lib python3 6 site package panda lib miss cpython 36 m x86 64 linux gnu so 13084 13084 13084 call init home anaconda anaconda3 lib python3 6 site package panda libs lib cpython 36 m x86 64 linux gnu so 13084 13084 13084 call init home anaconda anaconda3 lib python3 6 site package panda libs algo cpython 36 m x86 64 linux gnu so 13084 13084 13084 call init home anaconda anaconda3 lib python3 6 site package panda libs property cpython 36 m x86 64 linux gnu so 13084 13084 13084 call init home anaconda anaconda3 lib python3 6 site package panda lib hash cpython 36 m x86 64 linux gnu so 13084 13084 13084 call init home anaconda anaconda3 lib python3 6 site package bottleneck reduce cpython 36 m x86 64 linux gnu so 13084 13084 13084 call init home anaconda anaconda3 lib python3 6 site package bottleneck nonreduce cpython 36 m x86 64 linux gnu so 13084 13084 13084 call init home anaconda anaconda3 lib python3 6 site package bottleneck nonreduce axis cpython 36 m x86 64 linux gnu so 13084 13084 13084 call init home anaconda anaconda3 lib python3 6 site package bottleneck move cpython 36 m x86 64 linux gnu so 13084 13084 call init home anaconda anaconda3 lib python3 6 site package panda libs groupby cpython 36 m x86 64 linux gnu so 13084 13084 13084 call init home anaconda anaconda3 lib python3 6 site package panda libs reshape cpython 36 m x86 64 linux gnu so 13084 13084 13084 call init home anaconda anaconda3 lib python3 6 site package panda lib parser cpython 36 m x86 64 linux gnu so 13084 13084 13084 call init home anaconda anaconda3 lib python3 6 site package panda lib json cpython 36 m x86 64 linux gnu so 13084 13084 13084 call init home anaconda anaconda3 lib python3 6 site package panda lib writer cpython 36 m x86 64 linux gnu so 13084 13084 13084 call init home anaconda anaconda3 lib python3 6 site package panda io msgpack packer cpython 36 m x86 64 linux gnu so 13084 13084 13084 call init home anaconda anaconda3 lib python3 6 site package panda io msgpack unpacker cpython 36 m x86 64 linux gnu so 13084 13084 13084 call init home anaconda anaconda3 lib python3 6 site package panda util move cpython 36 m x86 64 linux gnu so 13084 13084 13084 call init home anaconda anaconda3 lib python3 6 site package panda libs testing cpython 36 m x86 64 linux gnu so 13084 13084 13084 call fini home anaconda 13084 call fini home anaconda anaconda3 lib python3 6 lib dynload grp cpython 36 m x86 64 linux gnu so 0 13084 13084 13084 call fini home anaconda anaconda3 lib python3 6 lib dynload decimal cpython 36 m x86 64 linux gnu so 0 13084 13084 13084 call fini home anaconda anaconda3 lib python3 6 site package numpy fft fftpack lite cpython 36 m x86 64 linux gnu so 0 13084 13084 13084 call fini home anaconda anaconda3 lib python3 6 site package numpy random mtrand cpython 36 m x86 64 linux gnu so 0 13084 13084 13084 call fini home anaconda anaconda3 lib python3 6 lib dynload hashlib cpython 36 m x86 64 linux gnu so 0 13084 13084 13084 call fini home anaconda anaconda3 lib python3 6 lib dynload blake2 cpython 36 m x86 64 linux gnu so 0 13084 13084 13084 call fini home anaconda anaconda3 lib python3 6 lib dynload sha3 cpython 36 m x86 64 linux gnu so 0 13084 13084 13084 call fini home anaconda anaconda3 lib python3 6 lib dynload bisect cpython 36 m x86 64 linux gnu so 0 13084 13084 13084 call fini home anaconda anaconda3 lib python3 6 lib dynload random cpython 36 m x86 64 linux gnu so 0 13084 13084 13084 call fini home anaconda anaconda3 lib python3 6 site package tensor 13084 call fini home anaconda anaconda3 lib python3 6 site package tensorflow python framework fast tensor util so 0 13084 13084 13084 call fini home anaconda anaconda3 lib python3 6 lib dynload json cpython 36 m x86 64 linux gnu so 0 13084 13084 13084 call fini home anaconda anaconda3 lib python3 6 lib dynload array cpython 36 m x86 64 linux gnu so 0 13084 13084 13084 call fini home anaconda anaconda3 lib python3 6 lib dynload ssl cpython 36 m x86 64 linux gnu so 0 13084 13084 13084 call fini home anaconda anaconda3 lib python3 6 site package h5py error cpython 36 m x86 64 linux gnu so 0 13084 13084 13084 call fini home anaconda anaconda3 lib python3 6 site package h5py conv cpython 36 m x86 64 linux gnu so 0 13084 13084 13084 call fini home anaconda anaconda3 lib python3 6 site package h5py h5r cpython 36 m x86 64 linux gnu so 0 13084 13084 13084 call fini home anaconda anaconda3 lib python3 6 site package h5py object cpython 36 m x86 64 linux gnu so 0 13084 13084 13084 call fini home anaconda anaconda3 lib python3 6 site package h5py def cpython 36 m x86 64 linux gnu so 0 13084 13084 13084 call fini home anaconda anaconda3 lib python3 6 site package h5 13084 call fini home anaconda anaconda3 lib python3 6 site package h5py lib libhdf5 5773eb11 so 103 0 0 0 13084 13084 13084 call fini home anaconda anaconda3 lib python3 6 site package h5py lib libz a147dcb0 so 1 2 3 0 13084 13084 13084 call fini home anaconda anaconda3 lib python3 6 site package h5py lib libsz 1c7dd0cf so 2 0 1 0 13084 13084 13084 call fini home anaconda anaconda3 lib python3 6 site package h5py lib libaec 2147abcd so 0 0 4 0 13084 13084 13084 call fini home anaconda anaconda3 lib libuuid so 1 0 13084 13084 13084 call fini home anaconda anaconda3 lib python3 6 lib dynload unicodedata cpython 36 m x86 64 linux gnu so 0 13084 13084 13084 call fini home anaconda anaconda3 lib python3 6 site package cryptography hazmat binding constant time abi3 so 0 13084 13084 13084 call fini home anaconda anaconda3 lib python3 6 site package cffi backend cpython 36 m x86 64 linux gnu so 0 13084 13084 13084 call fini home anaconda anaconda3 lib python3 6 site package lib cffi backend libffi bce22613 so 6 0 4 0 13084 13084 13084 call fini home anaconda anaconda3 lib python3 6 site package cryp 13084 call fini home anaconda anaconda3 lib python3 6 lib dynload libz so 1 0 13084 13084 13084 call fini lib x86 64 linux gnu librt so 1 0 13084 13084 13084 call fini home anaconda anaconda3 lib python3 6 site package scipy linalg fblas cpython 36 m x86 64 linux gnu so 0 13084 13084 13084 call fini home anaconda anaconda3 lib python3 6 site package scipy linalg flapack cpython 36 m x86 64 linux gnu so 0 13084 13084 13084 call fini home anaconda anaconda3 lib python3 6 site package scipy linalg flinalg cpython 36 m x86 64 linux gnu so 0 13084 13084 13084 call fini home anaconda anaconda3 lib python3 6 site package scipy linalg solve toeplitz cpython 36 m x86 64 linux gnu so 0 13084 13084 13084 call fini home anaconda anaconda3 lib python3 6 site package scipy linalg decomp update cpython 36 m x86 64 linux gnu so 0 13084 13084 13084 call fini home anaconda anaconda3 lib python3 6 site package scipy linalg cython blas cpython 36 m x86 64 linux gnu so 0 13084 13084 13084 call fini home anaconda anaconda3 lib python3 6 site package scipy linalg cython lapack cpython 36 m x86 64 linux gnu so 0 13084 13084 13084 call fini home anaconda anaconda3 lib python3 6 site package scipy ndimage nd image cpython 36 m 13084 call fini lib x86 64 linux gnu libdl so 2 0 13084 13084 13084 call fini home anaconda anaconda3 lib python3 6 site package scipy lib messagestream cpython 36 m x86 64 linux gnu so 0 13084 13084 13084 call fini home anaconda anaconda3 lib python3 6 site package scipy spatial voronoi cpython 36 m x86 64 linux gnu so 0 13084 13084 13084 call fini home anaconda anaconda3 lib python3 6 site package scipy spatial distance wrap cpython 36 m x86 64 linux gnu so 0 13084 13084 13084 call fini home anaconda anaconda3 lib python3 6 site package scipy spatial hausdorff cpython 36 m x86 64 linux gnu so 0 13084 13084 13084 call fini home anaconda anaconda3 lib python3 6 site package scipy ndimage ni label cpython 36 m x86 64 linux gnu so 0 13084 13084 13084 call fini home anaconda anaconda3 lib python3 6 site package panda libs tslib cpython 36 m x86 64 linux gnu so 0 13084 13084 13084 call fini home anaconda anaconda3 lib python3 6 site package panda libs tslib conversion cpython 36 m x86 64 linux gnu so 0 13084 13084 13084 call fini home anaconda anaconda3 lib python3 6 site package panda libs tslib np datetime cpython 36 m x86 64 linux gnu so 0 13084 13084 13084 call fini home anaconda anaconda3 lib python3 6 site package panda libs tslib nattype cpyth 13084 13084 call fini home anaconda anaconda3 lib python3 6 site package bottleneck nonreduce axis cpython 36 m x86 64 linux gnu so 0 13084 13084 13084 call fini home anaconda anaconda3 lib python3 6 site package bottleneck move cpython 36 m x86 64 linux gnu so 0 13084 13084 13084 call fini home anaconda anaconda3 lib python3 6 site package panda libs index cpython 36 m x86 64 linux gnu so 0 13084 13084 13084 call fini home anaconda anaconda3 lib python3 6 site package panda libs tslib period cpython 36 m x86 64 linux gnu so 0 13084 13084 13084 call fini home anaconda anaconda3 lib python3 6 site package panda lib tslib frequency cpython 36 m x86 64 linux gnu so 0 13084 13084 13084 call fini home anaconda anaconda3 lib python3 6 site package panda libs tslib resolution cpython 36 m x86 64 linux gnu so 0 13084 13084 13084 call fini home anaconda anaconda3 lib python3 6 site package panda libs tslib offset cpython 36 m x86 64 linux gnu so 0 13084 13084 13084 call fini home anaconda anaconda3 lib python3 6 site package panda libs join cpython 36 m x86 64 linux gnu so 0 13084 13084 13084 call fini home anaconda anaconda3 lib python3 6 site package panda lib op cpython 36 m x86 64 linux gnu so 0 13084 1308 13084 13084 call fini home anaconda anaconda3 lib python3 6 site package panda io msgpack packer cpython 36 m x86 64 linux gnu so 0 13084 13084 13084 call fini home anaconda anaconda3 lib python3 6 site package panda io msgpack unpacker cpython 36 m x86 64 linux gnu so 0 13084 13084 13084 call fini home anaconda anaconda3 lib python3 6 site package panda util move cpython 36 m x86 64 linux gnu so 0 13084 13084 13084 call fini home anaconda anaconda3 lib python3 6 site package panda libs testing cpython 36 m x86 64 linux gnu so 0 13084 13084 13084 call fini lib x86 64 linux gnu libpthread so 0 0 13084 env ld library path usr local cuda 9 0 lib64 usr local cuda 9 0 lib64 dyld library path be unset nvidia smi fri apr 19 15 35 28 2019 nvidia smi 390 116 driver version 390 116 gpu name persistence m bus i d disp a volatile uncorr ecc fan temp perf pwr usage cap memory usage gpu util compute m 0 titan x pascal off 00000000 02 00 0 off n a 47 77c p2 153w 250w 7177mib 12194mib 92 default process gpu memory gpu pid type process name usage 0 2059 g usr lib xorg xorg 16mib 0 28211 c python3 2299mib 0 28214 c python3 1275mib 0 28216 c python3 1275mib 0 28217 c python3 2299mib cuda lib usr local cuda 9 0 lib64 libcudart so 9 0 176 usr local cuda 9 0 lib64 libcudart static a usr local cuda 9 0 doc man man7 libcudart so 7 usr local cuda 9 0 doc man man7 libcudart 7 |
tensorflowtensorflow | tflite crash on android run a model for nnpi | Bug | system information have I write custom code as oppose to use a stock example script provide in tensorflow no os platform and distribution n a mobile device android q emulator and xiaomi a2 tensorflow instal from binary tensorflow version 1 13 1 gpu model and memory describe the current behavior simple program wrok with cpu but crash when nnapi be turn on with word e tflite op code 67 be currently not delegate to nnapi e tflite return error since tflite return failure nnapi delegate cc 739 expect behavior I would expect tflite to fall back to cpu for the opcode that can not work with nnapi code to reproduce the issue github repo |
tensorflowtensorflow | wrong document for dynamic rnn | Bug | please make sure that this be a documentation issue as per our github policy we only address code doc bug performance issue feature request and build installation issue on github tag doc template system information tensorflow version tf r1 13 and many other version doc link tf nn dynamic rnn describe the documentation issue it describe the argument sequence length as it s more for performance than correctness but this be wrong this argument enable this api to extract the last valid state of rnn instead of a padded time step so it be for correctness this argument can not reduce computation because even with this argument computation of new state of rnn cell still occur the only difference be that tf nn dynamic rnn choose to copy through old state instead of keep the new state and this lead to even large computation so it be not for performance the initial document be correct see this commit it s more for correctness than performance I don t really understand why you replace the correct document with a wrong one we welcome contribution by user will you be able to update submit a pr use the doc style guide to fix the doc issue |
tensorflowtensorflow | grammar | Bug | combine pr s of grammar |
tensorflowtensorflow | all diagram miss | Bug | system information tensorflow version n a doc link describe the documentation issue all of the diagram image be miss from both of the link above we welcome contribution by user will you be able to update submit a pr use the doc style guide to fix the doc issue no I don t know what the diagram be suppose to be |
tensorflowtensorflow | tensorflow 2 0 have miss feature relate to checkpointe dataset | Bug | system information tensorflow version 2 0 alpha 0 doc link describe the documentation issue with 2 0 alpha 0 tf datum dataset do not have any make xx iterator method so I can t save restore iterate state while training by use tf datum experimental make saveable from iterator how to save restore internal state of dataset |
tensorflowtensorflow | tensorflow 2 0 no gradient provide for any variable but only when tf math square and tf function be use | Bug | I post this in so but now be wonder if it be actually a bug in 176 sys version out 176 3 7 3 default mar 27 2019 16 54 48 n clang 4 0 1 tag release 401 final in 177 tf version out 177 2 0 0 alpha0 python import tensorflow as tf import tensorflow kera as keras import tensorflow kera backend as k import numpy as np m 1000 n 1 x np random randn m n astype np float32 y 3 0 np random randn m astype np float32 def create model a input keras layer input shape n dtype np float32 a k expand dim a input axis 2 q kera layer conv1d 1 1 a q tf math square q this break thing but only when use tf function model keras model model input a input output q return model model create model model predict x class trainer def init self epoch 10 self epoch epoch self model create model self optimizer tf optimizer adam self step 0 def train self x y epoch 10 x tf convert to tensor x dtype tf float32 y tf convert to tensor y dtype tf float32 for epoch in range epoch l self train one step x y return l tf function def train one step self x y with tf gradienttape as tape yp self model x loss tf reduce mean tf math square y yp gradient tape gradient loss self model trainable variable l self optimizer apply gradient zip gradient self model trainable variable d dict loss loss tf print yp 0 loss self step 1 trainer trainer l trainer train x y epoch 100 lib python3 7 site package tensorflow python keras optimizer v2 optimizer v2 py in filter grad grad and var 922 if not filter 923 raise valueerror no gradient provide for any variable s 924 v name for v in grad and var 925 if var with empty grad 926 log warning valueerror no gradient provide for any variable conv1d 47 kernel 0 conv1d 47 bias 0 |
tensorflowtensorflow | nameerror name mycapper be not define | Bug | for gardient clip in case of adam optimizer I m use below code create an optimizer opt adamoptimizer learning rate 0 1 compute the gradient for a list of variable grad and var opt compute gradient loss grad and var be a list of tuple gradient variable do whatever you need to the gradient part for example cap they etc cap grad and var mycapper gv 0 gv 1 for gv in grad and var ask the optimizer to apply the cap gradient opt apply gradient cap grad and var but I m get error nameerror name mycapper be not define fulltraceback grad mycapper gv 0 gv 1 for gv in grad and var nameerror name mycapper be not define pleae help 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 linux mobile device e g iphone 8 pixel 2 samsung galaxy if the issue happen on mobile device na tensorflow instal from source or binary binary tensorflow version use command below 1 10 0 python version 3 6 5 bazel version if compile from source na gcc compiler version if compile from source 7 2 0 cuda cudnn version cuda compilation tool release 9 0 v9 0 176 gpu model and memory tesla v100 you can collect some of this information use our environment capture script you can also obtain the tensorflow version with python c import tensorflow as tf print tf git version tf version describe the current behavior describe the expect behavior code to reproduce the issue provide a reproducible test case that be the bare minimum necessary to generate the problem other info log include any log or source code that would be helpful to diagnose the problem if include traceback please include the full traceback large log and file should be attach |
tensorflowtensorflow | tf 2 0 attributeerror module tensorflow have no attribute optimizer | Bug | system information os platform and distribution ubuntu 18 04 lt tensorflow instal from source or binary source tensorflow version use command below v2 0 0 alpha0 4 g2c2d508 2 0 0 alpha0 python version 3 5 7 bazel version if compile from source from docker image gcc compiler version if compile from source from docker image cuda cudnn version na gpu model and memory na describe the current behavior this documentation suggest there be 4 way to load an optimizer however tf 35 mark science python tf optimizer py traceback most recent call last file tf optimizer py line 7 in opt4 tf optimizer adagrad attributeerror module tensorflow have no attribute optimizer describe the expect behavior opt tf optimizers adagrad doesn t work the first 3 method do minimal code to reproduce the issue import tensorflow as tf tf version opt1 tf compat v2 keras optimizers adagrad opt2 tf compat v2 optimizer adagrad opt3 tf keras optimizers adagrad opt4 tf optimizer adagrad I don t know if this be a documentation issue or a tensorflow bug |
tensorflowtensorflow | modelcheckpoint callback error | Bug | please make sure that this be a bug as per our github policy we only address code doc bug performance issue feature request and build installation issue on github tag bug template system information have I write custom code as oppose to use a stock example script provide in tensorflow no os platform and distribution e g linux ubuntu 16 04 window 10 mobile device e g iphone 8 pixel 2 samsung galaxy if the issue happen on mobile device tensorflow instal from source or binary anaconda tensorflow version use command below 1 13 1 python version 3 5 bazel version if compile from source gcc compiler version if compile from source cuda cudnn version 7 3 4 gpu model and memory gtx 2080 ti describe the current behavior use tf kera when I fit the model with train dataset and validation dataset create from the tf dataset and use modelcheckpoint with default policy an error show up the error seem to happen because the callback try to save the model with the same name twice one at the end of each training epoch one at the end of each validation epoch twice use validation dataset this should not happen because I think the source code already set overwrite true but it still happen the error information traceback most recent call last file c user none iclouddrive course dl ass8 cnn mnist eager tf keras py line 180 in fit model and evaluate model file c user none iclouddrive course dl ass8 cnn mnist eager tf keras py line 172 in fit model and evaluate validation step 40 verbose verbose file c user none anaconda3 envs tensorflow lib site package tensorflow python keras engine training py line 851 in fit initial epoch initial epoch file c user none anaconda3 envs tensorflow lib site package tensorflow python keras engine training generator py line 232 in model iteration callback on epoch end epoch epoch log mode mode file c user none anaconda3 envs tensorflow lib site package tensorflow python keras callbacks py line 251 in on epoch end callback on epoch end epoch log file c user none anaconda3 envs tensorflow lib site package tensorflow python keras callbacks py line 624 in on epoch end self model save filepath overwrite true file c user none anaconda3 envs tensorflow lib site package tensorflow python keras engine network py line 1334 in save save model self filepath overwrite include optimizer file c user none anaconda3 envs tensorflow lib site package tensorflow python keras engine save py line 152 in save model name val shape dtype val dtype file c user none anaconda3 envs tensorflow lib site package h5py hl group py line 119 in create dataset self name dset file c user none anaconda3 envs tensorflow lib site package h5py hl group py line 287 in setitem h5o link obj i d self i d name lcpl lcpl lapl self lapl file h5py object pyx line 54 in h5py object with phil wrapper file h5py object pyx line 55 in h5py object with phil wrapper file h5py h5o pyx line 202 in h5py h5o link runtimeerror unable to create link name already exist describe the expect behavior at least the error should not happen when save model with the same name twice well if we can explicitly control when a model be save code to reproduce the issue provide a reproducible test case that be the bare minimum necessary to generate the problem import os import matplotlib pyplot as plt import tensorflow as tf from sklearn model selection import train test split tf enable eager execution mnist dataset image row image col 28 28 num class 10 batch size 32 temp dir temp def get input dataset use bfloat16 false create train and test dataset object for mnist dataset args use bfloat16 boolean to determine if input should be cast to bfloat16 return train dataset test dataset and input shape and class name cast dtype tf bfloat16 if use bfloat16 else tf float32 class name t shirt top trouser pullover dress coat sandal shirt sneaker bag ankle boot the datum split between train and test set x train y train x test y test tf keras datasets mnist load datum if tf keras backend image datum format channel first x train x train reshape x train shape 0 1 image row image col x test x test reshape x test shape 0 1 image row image col input shape 1 image row image col else x train x train reshape x train shape 0 image row image col 1 x test x test reshape x test shape 0 image row image col 1 input shape image row image col 1 preprocess x train x train astype float32 x test x test astype float32 x train 255 x test 255 y train tf keras util to categorical y train num class y test tf keras util to categorical y test num class build dataset ds train tf datum dataset from tensor slice image x train label y train ds test tf datum dataset from tensor slice image x test label y test x train x valid y train y valid train test split x train y train test size 0 1 ds train tf datum dataset from tensor slice x train y train ds valid tf datum dataset from tensor slice x valid y valid ds test tf datum dataset from tensor slice x test y test preprocess dataset ds train preprocess dataset ds train cast dtype ds valid preprocess dataset ds valid cast dtype ds test preprocess dataset ds test cast dtype return ds train ds valid ds test input shape class name def preprocess dataset dataset cast dtype dataset dataset map lambda x y tf cast x cast dtype y dataset dataset shuffle buffer size 6000 dataset dataset repeat dataset dataset batch batch size dataset dataset prefetch buffer size 1000 return dataset def plot image image plt figure plt imshow image plt colorbar plt grid false def test ds dataset fname save dir os path join temp dir s png fname os makedirs temp dir exist ok true plt figure for image label in dataset take 1 for index in range 4 plt subplot 2 2 index 1 plt imshow image index numpy reshape 28 28 plt xlabel label index numpy plt grid false plt savefig save dir bbox inch tight plt clf def get optimizer optimizer choice sgd learn rate 0 01 momentum 0 9 return sgd tf keras optimizer sgd lr learning rate momentum momentum adam tf keras optimizer adam lr learning rate get optimizer choice sgd return sgd tf train momentumoptimizer learning rate learning rate momentum momentum adam tf train adamoptimizer learning rate learning rate get optimizer choice sgd def create model input shape model tf keras sequential model add tf keras layer conv2d filter 64 kernel size 3 3 activation relu input shape input shape model add tf keras layer conv2d filter 64 kernel size 3 3 activation relu model add tf keras layer maxpooling2d pool size 2 2 model add tf keras layers dropout 0 25 model add tf keras layer flatten model add tf keras layer dense 128 activation relu model add tf keras layer dropout 0 5 model add tf keras layer dense num class activation softmax return model def create model functional input shape kernel size 3 3 dropout rate 0 l2 regularizer 0 1 input tensor tf keras input shape input shape layer tf keras layer conv2d filter 64 kernel size kernel size activation relu input tensor layer tf keras layer conv2d filter 64 kernel size kernel size activation relu layer layer tf keras layer maxpooling2d pool size 2 2 layer layer tf keras layers dropout rate dropout rate layer layer tf keras layer flatten layer layer tf keras layer dense 128 activation relu layer layer tf keras layers dropout rate dropout rate layer prediction tf keras layer dense num class activation softmax kernel regularizer tf keras regularizer l2 l l2 regularizer layer model tf keras model model input input tensor output prediction return model def fit model and evaluate model optimizer choice sgd learn rate 0 01 verbose 1 optimizer get optimizer optimizer choice optimizer choice learn rate learning rate os makedirs graph exist ok true os makedirs checkpoint exist ok true file path checkpoint model epoch 02d val loss 4f val acc 4f hdf5 model checkpoint tf keras callback modelcheckpoint file path log dir os path join board tf keras callbacks tensorboard log dir graph histogram freq 0 write graph true write image true model compile loss tf keras loss categorical crossentropy optimizer optimizer metric accuracy tf keras metric precision tf keras metrics recall model fit x ds train validation datum ds valid epoch 20 step per epoch 468 callback board model checkpoint validation step 40 verbose verbose score model evaluate ds test step 10 verbose 0 print test loss score 0 print test accuracy score 1 if name main ds train ds valid ds test input shape class name get input dataset model create model functional input shape fit model and evaluate model test ds ds train train test ds ds test test other info log include any log or source code that would be helpful to diagnose the problem if include traceback please include the full traceback large log and file should be attach |
tensorflowtensorflow | tf 2 0 documentation in guide for feature column in particular numeric column | Bug | system information tensorflow version 2 0 0 alpha0 doc link guide example that use feature column follow a bad pattern for numeric feature in particular the example in the guide always create a list of scalar numeric column s instead of one numeric column with all numeric feature this result in really bad performance when training see this thread location in the doc there may be other I miss base feature column structure of a pre make estimator program choose which column to use create feature column and input function create feature column input fn and the train the estimator the feature column guide from tf 1 or anything analogous be not available in the tf 2 doc that guide do mention the shape parameter but do not discuss performance at all and also describe create a separate numeric column for each feature in the iris dataset in this section numeric column |
tensorflowtensorflow | typeerror the train input config must be a input reader pb2 inputreader | Bug | please make sure that this be a bug as per our github policy we only address code doc bug performance issue feature request and build installation issue on github tag bug template system information have I write custom code as oppose to use a stock example script provide in tensorflow no os platform and distribution e g linux ubuntu 16 04 linux ubuntu 16 04 mobile device e g iphone 8 pixel 2 samsung galaxy if the issue happen on mobile device n a tensorflow instal from source or binary binary tensorflow version use command below v1 12 0 10232 g9a43dfe 1 13 1 python version python 3 7 1 bazel version if compile from source n a gcc compiler version if compile from source n a cuda cudnn version n a gpu model and memory n a describe the current behavior I ve be follow example locate here instead of run it in the google cloud I want to run it locally I will attach training config file below so when I run this command object detection model main py pipeline config path home konsof01 work model research pipiline config proto model dir home konsof01 tmp alsologtostderr I get the follow error traceback most recent call last file object detection model main py line 109 in tf app run file home konsof01 anaconda3 lib python3 7 site package tensorflow python platform app py line 40 in run run main main argv argv flag parser parse flag tolerate undef file home konsof01 anaconda3 lib python3 7 site package absl app py line 300 in run run main main args file home konsof01 anaconda3 lib python3 7 site package absl app py line 251 in run main sys exit main argv file object detection model main py line 105 in main tf estimator train and evaluate estimator train spec eval spec 0 file home konsof01 anaconda3 lib python3 7 site package tensorflow estimator python estimator training py line 471 in train and evaluate return executor run file home konsof01 anaconda3 lib python3 7 site package tensorflow estimator python estimator training py line 611 in run return self run local file home konsof01 anaconda3 lib python3 7 site package tensorflow estimator python estimator training py line 712 in run local saving listener save listener file home konsof01 anaconda3 lib python3 7 site package tensorflow estimator python estimator estimator py line 358 in train loss self train model input fn hook save listener file home konsof01 anaconda3 lib python3 7 site package tensorflow estimator python estimator estimator py line 1124 in train model return self train model default input fn hook save listener file home konsof01 anaconda3 lib python3 7 site package tensorflow estimator python estimator estimator py line 1151 in train model default input fn model fn lib modekeys train file home konsof01 anaconda3 lib python3 7 site package tensorflow estimator python estimator estimator py line 992 in get feature and label from input fn self call input fn input fn mode file home konsof01 anaconda3 lib python3 7 site package tensorflow estimator python estimator estimator py line 1079 in call input fn return input fn kwargs file home konsof01 work model research object detection input py line 486 in train input fn raise typeerror the train input config must be a typeerror the train input config must be a input reader pb2 inputreader the problem occur in this piece of code in input py if not isinstance train input config input reader pb2 inputreader raise typeerror the train input config must be a input reader pb2 inputreader however model pb2 detectionmodel type model config but isinstance train input config input reader pb2 inputreader be false pipeline config file pipiline config proto zip why be that thank you describe the expect behavior code to reproduce the issue provide a reproducible test case that be the bare minimum necessary to generate the problem other info log include any log or source code that would be helpful to diagnose the problem if include traceback please include the full traceback large log and file should be attach |
tensorflowtensorflow | crash on tf 1 13 when custom rnncell which have template | Bug | please make sure that this be a bug as per our github policy we only address code doc bug performance issue feature request and build installation issue on github tag bug template 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 arch linux mobile device e g iphone 8 pixel 2 samsung galaxy if the issue happen on mobile device tensorflow instal from source or binary binary pip tensorflow version use command below v1 13 0 rc2 5 g6612da8 1 13 1 python version 3 7 bazel version if compile from source gcc compiler version if compile from source cuda cudnn version 10 1 7 5 0 gpu model and memory gtx 1080 8 g you can collect some of this information use our environment capture script you can also obtain the tensorflow version with python c import tensorflow as tf print tf git version tf version describe the current behavior a code below work well on python 3 6 with tf 1 12 on python 3 7 with tf 1 13 however the code crash describe the expect behavior do not crash code to reproduce the issue provide a reproducible test case that be the bare minimum necessary to generate the problem python import tensorflow as tf class cell tf nn rnn cell rnncell def init self state size reuse none super cell self init reuse reuse self state size state size self encoder tf make template encoder self encoder property def state size self return self state size property def output size self return self state size def zero state self batch size dtype return tf zero batch size self state size def encoder self prev state obs input tf concat prev state obs 1 return tf layer dense input self state size none def call self input prev state state self encoder prev state input return state state input tf placeholder tf float32 32 2 20 rnn cell cell 20 output state tf nn dynamic rnn rnn cell input dtype tf float32 traceback warn tensorflow from test py 33 dynamic rnn from tensorflow python op rnn be deprecate and will be remove in a future version instruction for update please use keras layer rnn cell which be equivalent to this api warn tensorflow from pythondirectory site package tensorflow python op tensor array op py 162 colocate with from tensorflow python framework op be deprecate and will be remove in a future version instruction for update colocation handle automatically by placer warning tensorflow from test py 24 dense from tensorflow python layers core be deprecate and will be remove in a future version instruction for update use kera layer dense instead traceback most recent call last file test py line 33 in output state tf nn dynamic rnn rnn cell input dtype tf float32 file pythondirectory site package tensorflow python util deprecation py line 324 in new func return func args kwargs file pythondirectory site package tensorflow python op rnn py line 671 in dynamic rnn dtype dtype file pythondirectory site package tensorflow python op rnn py line 879 in dynamic rnn loop swap memory swap memory file pythondirectory site package tensorflow python op control flow op py line 3556 in while loop return same structure file pythondirectory site package tensorflow python op control flow op py line 3087 in buildloop pre body original loop var loop var shape invariant file pythondirectory site package tensorflow python op control flow op py line 3022 in buildloop body result body pack var for body file pythondirectory site package tensorflow python op control flow op py line 3525 in body lambda I lv I 1 orig body lv file pythondirectory site package tensorflow python op rnn py line 847 in time step output new state call cell file pythondirectory site package tensorflow python op rnn py line 833 in call cell lambda cell input t state file pythondirectory site package tensorflow python op rnn cell impl py line 234 in call return super rnncell self call input state file pythondirectory site package tensorflow python layers base py line 534 in call add element to collection self update op graphkey update op file pythondirectory site package tensorflow python keras engine base layer py line 653 in update return self update self gather child attribute update file pythondirectory site package tensorflow python keras engine base layer py line 1647 in gather child attribute getattr layer attribute for layer in self layer file pythondirectory site package tensorflow python keras engine base layer py line 1647 in getattr layer attribute for layer in self layer attributeerror template object have no attribute update other info log include any log or source code that would be helpful to diagnose the problem if include traceback please include the full traceback large log and file should be attach |
tensorflowtensorflow | convert inceptionv3 to tflite miss identityn | Bug | system information google colab tensorflow instal from source or binary instal from tensorflow version or github sha if from source provide the text output from tflite convert some of the operator in the model be not support by the standard tensorflow lite runtime if those be native tensorflow operator you might be able to use the extended runtime by pass enable select tf op or by set target op tflite builtin select tf op when call tf lite tfliteconverter otherwise if you have a custom implementation for they you can disable this error with allow custom op or by set allow custom op true when call tf lite tfliteconverter here be a list of builtin operator you be use average pool 2d concatenation conv 2d fully connect max pool 2d mean reshape softmax here be a list of operator for which you will need custom implementation identityn also please include a link to a graphdef or the model if possible google colab notebook I update the google colab tutorial and update it to use inceptionv3 base on the link include in the notebook and modify the file size to be 299 299 |
tensorflowtensorflow | tf summary tensor summary no long appear in the repl see it in the code doc need update | Bug | not sure what the intention be as I see a lot of v1 v2 name change tf summary scalar be there but not tf summary tensor summary in 80 tf version out 80 2 0 0 alpha0 |
tensorflowtensorflow | get error message when use tf summary scalar with tf 2 0 | Bug | please make sure that this be a bug as per our github policy we only address code doc bug performance issue feature request and build installation issue on github tag bug template 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 ubuntu 18 04 mobile device e g iphone 8 pixel 2 samsung galaxy if the issue happen on mobile device n a tensorflow instal from source or binary binary tensorflow version use command below tf nightly gpu 2 0 preview 2 0 0 dev20190413 python version 3 6 7 bazel version if compile from source n a gcc compiler version if compile from source n a cuda cudnn version 10 0 gpu model and memory yes 11 gb you can collect some of this information use our environment capture script you can also obtain the tensorflow version with python c import tensorflow as tf print tf git version tf version describe the current behavior I get the follow error message when I use tensorflow in gpu mode when I use tensorflow in cpu mode such error message doesn t occur attributeerror module tensorboard summary tf summary have no attribute summary scope describe the expect behavior I expect no error when I call tf summary scalar in tf2 0 in both gpu and cpu mode code to reproduce the issue provide a reproducible test case that be the bare minimum necessary to generate the problem python usr bin python3 import tensorflow as tf def main log tf summary create file writer checkpoint with log as default tf summary scalar loss 1 step 0 if name main main other info log include any log or source code that would be helpful to diagnose the problem if include traceback please include the full traceback large log and file should be attach |
tensorflowtensorflow | problem with running ganestimator type error in eval util impl py | Bug | 55 if grid shape 0 grid shape 1 int input tensor shape 0 56 raise valueerror grid shape s incompatible with minibatch size I 57 grid shape int input tensor shape 0 typeerror int return non int type nonetype source tensorflow contrib gan python eval python eval util impl py I be rebuild gan estimator with mnist dataset input function be define as python def preprocess img label img tf cast img dtype tf float32 img tf divide img 255 0 if len img shape 2 img tf expand dim img 2 prior tf random normal 64 return prior img def get input fn be train if be train dataset train else dataset test tf dataset tf datum dataset from tensor slice dataset tf dataset tf dataset map preprocess def input fn datum tf dataset batch batch size da iterator datum make one shot iterator prior d datum da iterator get next return prior d datum return input fn input fn get input fn true gan generator and discriminator python def generator fn prior output tf keras layer dense 1024 prior output tf keras layer dense 7 7 128 output output tf reshape output 1 7 7 128 output tf keras layer conv2dtranspose 64 4 2 padding same output output tf keras layer conv2dtranspose 32 4 2 padding same output output tf keras layer conv2dtranspose 1 4 padding same activation tf nn tanh output return output def discriminator fn img unused conditioning output tf keras layer conv2d 64 4 2 img output tf keras layer conv2d 128 4 2 output output tf layer flatten output output tf layer dense 1024 output output tf contrib layer layer norm output output tf layer dense 1 output return output estimator be create as python gan estimator tfgan estimator ganestimator generator fn generator fn discriminator fn discriminator fn generator loss fn tfgan loss wasserstein generator loss discriminator loss fn tfgan loss wasserstein discriminator loss generator optimizer tf train adamoptimizer 0 001 0 5 discriminator optimizer tf train adamoptimizer 0 0001 0 5 add summary tfgan estimator summarytype image gan estimator train input fn max step 1000 the environment I be run on be ubuntu 16 01 python 3 6 6 tensorflow 1 13 |
tensorflowtensorflow | error filename log with tf log warn | Bug | please make sure that this be a bug as per our github policy we only address code doc bug performance issue feature request and build installation issue on github tag bug template system information have I write custom code as oppose to use a stock example script provide in tensorflow no os platform and distribution e g linux ubuntu 16 04 window 7 mobile device e g iphone 8 pixel 2 samsung galaxy if the issue happen on mobile device tensorflow instal from source or binary binary tensorflow version use command below 1 13 1 python version 3 7 0 you can collect some of this information use our environment capture script you can also obtain the tensorflow version with python c import tensorflow as tf print tf git version tf version describe the current behavior stdout print tf log py warn test warn test warning describe the expect behavior stdout print test warn test warn code to reproduce the issue python import log import sys h log streamhandler sys stdout formatter log formatter filename s funcname s message s h setformatter formatter log log getlogger log addhandler h import tensorflow as tf tf logging warn test warn tf log warning test warn other info log since the logging warn of python3 be the wrapper of log warning code be here l1459 this add one frame make the get caller 4 l49 of tf logging miss the warn method |
tensorflowtensorflow | bug tfdbg session can not be use with sessionrunhook | Bug | please make sure that this be a bug as per our github policy we only address code doc bug performance issue feature request and build installation issue on github tag bug template 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 archlinux mobile device e g iphone 8 pixel 2 samsung galaxy if the issue happen on mobile device tensorflow instal from source or binary binary tensorflow version use command below b v1 13 0 rc2 5 g6612da8 1 13 1 python version 3 7 bazel version if compile from source n a gcc compiler version if compile from source n a cuda cudnn version n a gpu model and memory n a the follow code python code utf 8 file import numpy as np import tensorflow as tf from tensorflow python import debug as tf debug a tf placeholder tf float32 10 b a 1 c b 2 class hook tf train sessionrunhook def before run self return tf train sessionrunarg fetch c class hook2 tf train sessionrunhook def before run self return tf train sessionrunarg fetch b sess tf session sess tf debug localclidebugwrappersession sess class sessioncreator def create session self return sess final sess tf train monitoredsession session creator sessioncreator hook hook hook2 final sess run b feed dict a np arange 10 throw traceback most recent call last file tfdbg py line 30 in final sess run b feed dict a np arange 10 file home wyx local lib python3 7 site package tensorflow python training monitor session py line 676 in run run metadata run metadata file home wyx local lib python3 7 site package tensorflow python training monitor session py line 1171 in run run metadata run metadata file home wyx local lib python3 7 site package tensorflow python training monitor session py line 1270 in run raise six reraise original exc info file usr lib python3 7 site package six py line 693 in reraise raise value file home wyx local lib python3 7 site package tensorflow python training monitor session py line 1255 in run return self sess run args kwargs file home wyx local lib python3 7 site package tensorflow python training monitor session py line 1327 in run run metadata run metadata file home wyx local lib python3 7 site package tensorflow python training monitor session py line 1091 in run return self sess run args kwargs file home wyx local lib python3 7 site package tensorflow python debug wrapper framework py line 463 in run empty fetch not nest flatten fetch file home wyx local lib python3 7 site package tensorflow python pywrap tensorflow internal py line 2156 in flatten return pywrap tensorflow internal flatten nest typeerror not support between instance of hook and str I believe this issue be introduce in a year ago flatten can not handle fetch create by hook the fix will be to obtain empty fetch in a smart way |
tensorflowtensorflow | quantization with multiple gpu training | Bug | when I try to quantization with multiple gpu training a error happen keyerror the name tower 0 gradient tower 0 mobilenet conv ds 13 pointwise conv batchnorm batchnorm 1 rsqrt refer to an operation not in the graph my multiple gpu code with tf variable scope tf get variable scope for I in range 2 with tf device gpu d I with tf name scope tower d I as scope1 x x int I batch size 0 5 int I 1 batch size 0 5 y y int I batch size 0 5 int I 1 batch size 0 5 with slim arg scope mobilenet mobilenet arg scope weight decay 0 00004 logit end point mobilenet mobilenetv1 input x num class 136 be training be train width multipli 0 5 scope mobilenet with tf variable scope tf get variable scope reuse tf auto reuse tf contrib quantize create training graph quant delay get quant delay loss tf loss huber loss label y prediction logit delta 1 0 scope scope1 tf get variable scope reuse variable if quantize tf contrib quantize create training graph quant delay get quant delay update op tf get collection tf graphkey update op with tf control dependency update op grad train step compute gradient loss tower grad append grad grad average gradient tower grad train op train step apply gradient grad global step global step |
tensorflowtensorflow | can not initialize two different model with the same checkpoint | Bug | I m try to use tf slim resnet v1 101 pre train model to initialize two input branch of model with different input rgb feature 0 and depth map feature 1 respectively to far combine they and feed their joint result to some spatial convolution layer in order to achieve that I create separate scope for each of the branch and initialize the model with pre train checkpoint exclude the last fully connect layer logit system information os platform and distribution linux ubuntu 16 04 tensorflow version 1 13 python version 3 5 cuda cudnn version 10 1 gpu model and memory gv100 32 gb describe the current behavior with tf variable scope rgb branch with tf contrib slim arg scope resnet v1 resnet arg scope rgb logit end point resnet v1 resnet v1 101 feature 0 self num class be training be train rgb variable to restore tf contrib slim get variable to restore exclude rgb branch resnet v1 101 logit strip scope name rgb assignment map rgb variable to restore 0 name split 0 rgb variable to restore 0 rgb assignment map update v name split 0 split 1 1 v for v in rgb variable to restore 1 tf train init from checkpoint self pre train model path rgb assignment map with tf variable scope depth branch with tf contrib slim arg scope resnet v1 resnet arg scope depth logit end point resnet v1 resnet v1 101 feature 1 self num class be training be train depth variable to restore tf contrib slim get variable to restore exclude depth branch resnet v1 101 logit depth assignment map depth variable to restore 0 name split 0 depth variable to restore 0 depth assignment map update v name split 0 split 1 1 v for v in depth variable to restore 1 tf train init from checkpoint self pre train model path depth assignment map during the second initialization the depth branch tf complain about shape of the logit layer as if it be never remove valueerror shape of variable rgb branch resnet v1 101 logit bias 0 3 doesn t match with shape of tensor resnet v1 101 logit bias 1000 from checkpoint reader the problem doesn t occur when I initialize only one branch and be tie only to the first of the branch define if I change the order of the branch the above error would change to depth branch resnet v1 101 logit bias 0 describe the expect behavior expect have two separate graph both initialize by the same checkpoint |
tensorflowtensorflow | com ai tensorflow a libc fatal signal 11 sigsegv code 1 fault addr 0x28 in tid 17491 m ai tensorflow | Bug | it have a crash in tensorflow lite android app beginning of crash com ai tensorflow a libc fatal signal 11 sigsegv code 1 fault addr 0x28 in tid 17491 m ai tensorflow system information have I use tensorflow demo app tensorflow tensorflow lite example android lite android the issue happen on mobile device android 5 0 2 how this error be happen tfliteobjectdetectionapimodel java 1 tflite new interpreter loadmodelfile assetmanager modelfilename after get the tflite and then to detect the image 2 when it want to stop the tflite by tflite close 3 if it s run image detect by tflite runformultipleinputsoutput inputarray outputmap and this timing I call the tflite close then this error happen all the error information 04 15 03 34 43 581 17441 17491 com ai tensorflow a libc fatal signal 11 sigsegv code 1 fault addr 0x28 in tid 17491 m ai tensorflow 04 15 03 34 43 586 17441 17488 com ai tensorflow a libc fatal signal 11 sigsegv code 1 fault addr 0x28 in tid 17488 m ai tensorflow 04 15 03 34 43 596 17441 17490 com ai tensorflow a libc fatal signal 11 sigsegv code 1 fault addr 0x10 in tid 17490 m ai tensorflow 04 15 03 34 43 685 11849 11849 I debug 04 15 03 34 43 685 11849 11849 I debug revision 0 04 15 03 34 43 685 11849 11849 I debug abi arm 04 15 03 34 43 686 11849 11849 I debug pid 17441 tid 17491 name m ai tensorflow com ai tensorflow 04 15 03 34 43 686 11849 11849 I debug signal 11 sigsegv code 1 segv maperr fault addr 0x28 04 15 03 34 43 708 11849 11849 I debug r0 00000028 r1 0000000c r2 0000000f r3 00000000 04 15 03 34 43 709 11849 11849 I debug r4 0000000f r5 b908ef30 r6 9f26b1f4 r7 9f565cf8 04 15 03 34 43 709 11849 11849 I debug r8 b910d888 r9 0000000c sl 00000001 fp 00000000 04 15 03 34 43 710 11849 11849 I debug ip 00000001 sp 9f565cc0 lr a42ee2af pc a42ed98c cpsr 200d0030 04 15 03 34 43 710 11849 11849 I debug backtrace 04 15 03 34 43 710 11849 11849 I debug 00 pc 0005198c datum app com ai tensorflow 1 lib arm libtensorflowlite jni so 04 15 03 34 43 710 11849 11849 I debug 01 pc 000522ab datum app com ai tensorflow 1 lib arm libtensorflowlite jni so 04 15 03 34 43 710 11849 11849 I debug 02 pc 000d6bd3 datum app com ai tensorflow 1 lib arm libtensorflowlite jni so 04 15 03 34 43 710 11849 11849 I debug 03 pc 000f5a7f datum app com ai tensorflow 1 lib arm libtensorflowlite jni so 04 15 03 34 43 710 11849 11849 I debug 04 pc 00012ff7 system lib libc so pthread start void 30 04 15 03 34 43 710 11849 11849 I debug 05 pc 00010fcf system lib libc so start thread 6 |
tensorflowtensorflow | update tf upgrade v2 py | Bug | the script print forsafety because of the miss blank space here |
tensorflowtensorflow | update link to correct example project | Bug | currently the link be open the image classification example not object detection |
tensorflowtensorflow | nnapi doesn t support tensor with rank 0 index 3 name padv2 constant value | Bug | please make sure that this be a bug as per our github policy we only address code doc bug performance issue feature request and build installation issue on github tag bug template system information os platform and distribution window 10 and android 9 0 mobile device xiaomi 8 tensorflow version org tensorflow tensorflow lite 0 0 0 nightly use 1 11 0 have the same problem model resnet18 convert from pytorch log from android 9 0 e tflite nnapi doesn t support tensor with rank 0 index 3 name padv2 constant value e tflite return error since tflite return failure nnapi delegate cc 736 fail to build graph for nnapi other if I use resnet50 from tensorflow contrib slim python slim net it also have similar problem like e tflite nnapi doesn t support tensor with rank 0 index 7 name inference model mean reduction indice return error since tflite return failure nnapi delegate cc 736 fail to build graph for nnapi |
tensorflowtensorflow | sparkfun edge build failure | Bug | on macos 10 14 4 I m follow the direction at this location 0 everything go fine until step 4 when I try make f tensorflow lite experimental micro tool make makefile target sparkfun edge micro speech bin and get the error in file include from tensorflow lite experimental micro kernel fully connect cc 16 0 tensorflow lite kernel internal reference fully connect h 18 10 fatal error fixedpoint fixedpoint h no such file or directory include fixedpoint fixedpoint h compilation terminate any idea why fixedpoint fixedpoint h be not find |
tensorflowtensorflow | autograph fail after restore from the checkpoint | Bug | system information have I write custom code as oppose to use a stock example script provide in tensorflow no tensorflow instal from source or binary tensorflow version use command below bash conda install tensorflow gpu 2 0 alpha python version 3 7 1 bazel version if compile from source gcc compiler version if compile from source cuda cudnn version not relevent gpu model and memory not relevent describe the current behavior run the follow code for the first time show no warning however run it the second time result in this warning message w0412 15 23 38 694277 140471310108480 tf log py 161 entity could not be transform and will be stage without change error detail can be find in the log when run with the env variable autograph verbosity 1 please report this to the autograph team cause object conversion be not yet support if you be try to convert code that use an exist object try include the creation of that object in the conversion for example instead of convert the method of a class try convert the entire class instead see use the functional api for more information w0412 15 23 38 697970 140471310108480 tf log py 161 entity could not be transform and will be stage without change error detail can be find in the log when run with the env variable autograph verbosity 1 please report this to the autograph team cause object conversion be not yet support if you be try to convert code that use an exist object try include the creation of that object in the conversion for example instead of convert the method of a class try convert the entire class instead see ensorflow blob master tensorflow python autograph readme md use the functional api for more information describe the expect behavior autograph should work just find even after restore from a checkpoint code to reproduce the issue provide a reproducible test case that be the bare minimum necessary to generate the problem python import os import tensorflow as tf from tensorflow keras import layer model os environ tf cpp min log level 2 network model sequential layer dense 1 activation relu tf function def forward batch network batch if name main ckpt tf train checkpoint model network manager tf train checkpointmanager ckpt tmp ckptt max to keep 10 if manager late checkpoint print f resume training use manager late checkpoint ckpt restore manager late checkpoint batch tf zero 1 1 dtype tf float32 forward batch manager save other info log include any log or source code that would be helpful to diagnose the problem if include traceback please include the full traceback large log and file should be attach |
tensorflowtensorflow | shuffling of test dataset be not explicitly do | Bug | system information tensorflow version 2 0 doc link input pipeline cell documentation code issue cell have the follow code 1 test dataset tf datum dataset list file path test jpg 2 shuffle so that for every epoch a different image be generate 3 to predict and display the progress of our model 4 train dataset train dataset shuffle buffer size 5 test dataset test dataset map load image test 6 test dataset test dataset batch 1 from line no 4 it seem we be not shuffle the test dataset should it have be test dataset test dataset shuffle buffer size |
tensorflowtensorflow | the text generation tutorial doesn t seem to be pass along hidden state while generate text | Bug | system information tensorflow version tensorflow gpu 1 10 1 doc link the prediction loop describe the documentation issue I follow the text generation tutorial and train my model until it achieve a cross entropy loss of 0 5 after 30 epoch however the generate text be mostly garbage read the code link in the section above I see python we pass the predict word as the next input to the model along with the previous hidden state input eval tf expand dim predict i d 0 which doesn t make sense to I the comment say we re pass along the hidden state which maybe the model be do for we but the only text that s pass along be the last predict character after modify the text generation function to pass to the model last at most seq length generate character the output text start to look like actual shakespeare english my question be be the model suppose to implicitly propagate hidden state and I ve do something wrong follow the tutorial or be the tutorial accidentally omit the part where we pass along the last set of generate text note since I m run tensorflow 1 10 and the tutorial seem to require 1 13 I have to replace the loss function with tf nn sparse softmax cross entropy with logit label label logit logit I don t think that change be relate but be it possible that tf 1 10 doesn t propagate the hidden state in the model we welcome contribution by user will you be able to update submit a pr use the doc style guide to fix the doc issue I m happy to contribute my code change to the tutorial but want to make sure I m understand what s go wrong first |
tensorflowtensorflow | attributeerror module tensorflow api v1 nn have no attribute softmax cross entryopy with logit v2 | Bug | system info geoff ubuntu uname a linux ubuntu 4 15 0 47 generic 50 ubuntu smp we d mar 13 10 44 52 utc 2019 x86 64 x86 64 x86 64 gnu linux geoff ubuntu work interview coursera dl python version python 3 6 6 anaconda inc geoff ubuntu python c import tensorflow as tf print tf git version tf version b v1 13 1 0 g6612da8951 1 13 1 geoff ubuntu work interview coursera dl python c import kera print keras version use tensorflow backend 2 2 4 describe issue no luck get this to work so far and I can t find any similar issue reproduce issue reproduce txt tf env txt |
tensorflowtensorflow | tf 2 0 kera unable save and load weight for double nest model | 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 mac tensorflow instal from source or binary binary tensorflow version use command below 2 0 0 python version 3 7 describe the current behavior load weight throw exception on a doubly nest model describe the expect behavior load weight should work this problem only happen on two layer of nest model with non trainable weight the reason be save weight and load weight handle nest model differently save weight call layer weight for each layer load weight recursively call model weight if layer be a nested model code to reproduce the issue import tensorflow as tf from tensorflow keras import model from tensorflow keras layers import input conv2d batchnormalization shape none none 3 def bnmodel x input input shape x conv2d 3 1 x x batchnormalization x return model input x x inner input input shape x bnmodel x x bnmodel x inner model model inner input x input input shape model model input inner model input inner model save weight test h5 inner model load weight test h5 work fine model save weight test h5 model load weight test h5 exception axis don t match array other info log this bug be also report on upstream kera here be a detailed analysis on why this be happen issuecomment 482438283 full exception file test py line 27 in model load weight test h5 exception axis don t match array file usr local anaconda3 lib python3 7 site package tensorflow python keras engine network py line 1497 in load weight hdf5 format load weight from hdf5 group f self layer file usr local anaconda3 lib python3 7 site package tensorflow python keras save hdf5 format py line 751 in load weight from hdf5 group layer weight value original keras version original backend file usr local anaconda3 lib python3 7 site package tensorflow python keras save hdf5 format py line 377 in preprocess weight for loading weight convert nested model weight file usr local anaconda3 lib python3 7 site package tensorflow python keras save hdf5 format py line 365 in convert nest model original backend original backend file usr local anaconda3 lib python3 7 site package tensorflow python keras save hdf5 format py line 377 in preprocess weight for loading weight convert nested model weight file usr local anaconda3 lib python3 7 site package tensorflow python keras save hdf5 format py line 353 in convert nest model original backend original backend file usr local anaconda3 lib python3 7 site package tensorflow python keras save hdf5 format py line 459 in preprocess weight for loading weight 0 np transpose weight 0 3 2 0 1 file usr local anaconda3 lib python3 7 site package numpy core fromnumeric py line 598 in transpose return wrapfunc a transpose axis file usr local anaconda3 lib python3 7 site package numpy core fromnumeric py line 51 in wrapfunc return getattr obj method args kwds valueerror axis don t match array |
tensorflowtensorflow | issue in code of image captioning notebook | Bug | please make sure that this be a documentation issue as per our github policy we only address code doc bug performance issue feature request and build installation issue on github tag doc template system information tensorflow version any doc link describe the documentation issue the code in this cell scrollto uewm9xrycg45 which set vocab size len tokenizer word index 1 be wrong vocab size be set to 8235 1 but actually need to be set to 5000 1 as top k 5000 we welcome contribution by user will you be able to update submit a pr use the doc style guide to fix the doc issue yes |
tensorflowtensorflow | build error tensorflow lite gpu delegate | Bug | please make sure that this be a build installation issue as per our github policy we only address code doc bug performance issue feature request and build installation issue on github tag build template system information os platform and distribution e g linux ubuntu 16 04 debian 4 19 20 1rodete1 2019 02 12 2018 x86 64 gnu linux mobile device e g iphone 8 pixel 2 samsung galaxy if the issue happen on mobile device build for android by add the follow code android sdk repository name androidsdk api level 26 build tool version 26 0 2 path android ndk repository name androidndk path android ndk r19c api level 19 tensorflow instal from source or binary source tensorflow version master commit number 7bdb14a0b python version 2 7 16rc1 instal use virtualenv pip conda no bazel version if compile from source 0 24 1 gcc compiler version if compile from source android 5058415 base on r339409 clang version 8 0 2 cuda cudnn version n a gpu model and memory n a describe the problem try tensorflow lite gpu delegate and get a build error after add the follow code into workspace under tensorflow source code s root directory android sdk repository name androidsdk api level 26 build tool version 26 0 2 path android ndk repository name androidndk path android ndk r19c api level 19 I run the follow command bazel build cxxopt std c 11 tensorflow lite java demo app src main tflitecamerademo it report bazel out android armeabi v7a opt bin external fp16 virtual include fp16 fp16 h 5 10 fatal error fp16 fp16 h file not find include 1 error generate target tensorflow lite java demo app src main tflitecamerademo fail to build provide the exact sequence of command step that you execute before run into the problem any other info log include any log or source code that would be helpful to diagnose the problem if include traceback please include the full traceback large log and file should be attach |
tensorflowtensorflow | tf 2 method estimator model to estimator fail but model fit work for tf keras create model | Bug | system information have I write custom code as oppose to use a stock example script provide in tensorflow no os platform and distribution e g linux ubuntu 16 04 win 10 colab linux tensorflow instal from source or binary binary tensorflow version use command below tf 2 0 0 alpha python version 3 5 describe the current behavior the below code have be take from the tensorflow without phd series rnn time series prediction def compile keras sequential model list of layer msg a tf keras sequential model be a sequence of layer model tf keras sequential list of layer keras do not have a pre define metric for root mean square error let s define one def rmse y true y pre root mean square error return tf sqrt tf reduce mean tf square y pre y true print nmodel msg optimizer sgd tf keras optimizer sgd lr 0 1 decay 1e 6 momentum 0 9 nesterov true to finalize the model specify the loss the optimizer and metric model compile loss mean squared error optimizer sgd optimizer tf keras optimizer sgd lr 0 0001 momentum 0 9 metric rmse this print a description of the model model summary return model create keras model def model fn keras rnn model rmse 0 164 after 10 epoch model layer rnn l reshape seqlen 1 input shape seqlen batchsize seqlen 1 be necessary for rnn model l gru rnn cellsize return sequence true output shape batchsize seqlen rnn cellsize l gru rnn cellsize keep only last output in sequence output shape batchsize rnn cellsize l dense 1 output shape batchsize 1 model rnn compile keras sequential model model layer rnn rnn return model rnn while convert the keras model to estimator version below line give error estimator tf keras estimator model to estimator keras model model fn keras model rnn model sequential 27 layer type output shape param reshape 27 reshape none 16 1 0 unify gru 57 unifiedgru none 16 32 3360 unify gru 58 unifiedgru none 32 6336 dense 27 dense none 1 33 total param 9 729 trainable param 9 729 non trainable param 0 invalidargumenterror traceback most recent call last c user hrafiq appdata local program python python35 lib site package tensorflow python client session py in do call self fn args 1334 try 1335 return fn args 1336 except error operror as e c user hrafiq appdata local program python python35 lib site package tensorflow python client session py in run fn feed dict fetch list target list option run metadata 1317 ensure any change to the graph be reflect in the runtime 1318 self extend graph 1319 return self call tf sessionrun c user hrafiq appdata local program python python35 lib site package tensorflow python client session py in extend graph self 1352 with self graph session run lock pylint disable protect access 1353 tf session extendsession self session 1354 invalidargumenterror node training sgd gradient unified gru 58 statefulpartitionedcall grad statefulpartitionedcall connect to invalid output 4 of source node unified gru 58 statefulpartitionedcall which have 4 output during handling of the above exception another exception occur invalidargumenterror traceback most recent call last in 5 convert keras model to estimator 6 tf disable eager execution 7 estimator tf keras estimator model to estimator keras model model fn keras 8 estimator model fn keras 9 c user hrafiq appdata local program python python35 lib site package tensorflow python keras estimator init py in model to estimator keras model keras model path custom object model dir config 71 custom object custom object 72 model dir model dir 73 config config 74 75 lint thenchange tensorflow estimator python estimator keras py c user hrafiq appdata local program python python35 lib site package tensorflow estimator python estimator keras py in model to estimator keras model keras model path custom object model dir config 488 if keras model be graph network 489 warm start path save first checkpoint keras model custom object 490 config 491 elif keras model build 492 log warning you be create an estimator from a keras model manually c user hrafiq appdata local program python python35 lib site package tensorflow estimator python estimator keras py in save first checkpoint keras model custom object config 365 pylint disable protect access 366 model make train function 367 k initialize variable sess 368 pylint enable protect access 369 saver saver lib saver c user hrafiq appdata local program python python35 lib site package tensorflow python keras backend py in initialize variable session 760 mark as initialize 761 be initialize session run 762 variable module be variable initialize v for v in candidate var 763 uninitialized var 764 for flag v in zip be initialize candidate var c user hrafiq appdata local program python python35 lib site package tensorflow python client session py in run self fetch feed dict option run metadata 928 try 929 result self run none fetch feed dict option ptr 930 run metadata ptr 931 if run metadata 932 proto datum tf session tf getbuffer run metadata ptr c user hrafiq appdata local program python python35 lib site package tensorflow python client session py in run self handle fetch feed dict option run metadata 1151 if final fetch or final target or handle and feed dict tensor 1152 result self do run handle final target final fetch 1153 feed dict tensor option run metadata 1154 else 1155 result c user hrafiq appdata local program python python35 lib site package tensorflow python client session py in do run self handle target list fetch list feed dict option run metadata 1327 if handle be none 1328 return self do call run fn feed fetch target option 1329 run metadata 1330 else 1331 return self do call prun fn handle feed fetch c user hrafiq appdata local program python python35 lib site package tensorflow python client session py in do call self fn args 1347 pass 1348 message error interpolation interpolate message self graph 1349 raise type e node def op message 1350 1351 def extend graph self invalidargumenterror node training sgd gradient unified gru 58 statefulpartitionedcall grad statefulpartitionedcall connect to invalid output 4 of source node unified gru 58 statefulpartitionedcall which have 4 output describe the expect behavior should not give any error other info log include any log or source code that would be helpful to diagnose the problem if include traceback please include the full traceback large log and file should be attach |
tensorflowtensorflow | negative pixel value for ssim input | Bug | ssim and ms ssim take image as input and max val which be define as the dynamic range imply that image can be eg in 1 1 with maxval 2 however they call convert image dtype which assume pixel be in 0 so it be not clear whether ssim ms ssim can take image with negative pixel value |
tensorflowtensorflow | distribution strategy and conditional parameter update fail with operation have be mark as not fetchable | Bug | system information have I write custom code yes os platform and distribution ubuntu 16 04 tensorflow instal from binary tensorflow version 2 0 0a0 python version 3 5 2 describe the current behavior when use conditional parameter update and a distribution strategy in graph mode the first sess run fail with the error operation have be mark as not fetchable describe the expect behavior distribution strategy should support conditional parameter update code to reproduce the issue I try to compile a small and representative snippet that be the base logic for gradient accumulation python import tensorflow as tf def train op accum step none step tf variable initial value 0 dtype tf int32 trainable false optimizer tf optimizer sgd learn rate 0 1 variable tf variable tf zero 4 5 dtype tf float32 gradient tf variable tf one 4 5 dtype tf float32 trainable false if accum step be none return optimizer apply gradient zip gradient variable else return tf cond tf equal step 1 accum step 0 true fn lambda optimizer apply gradient zip gradient variable false fn tf no op device gpu 2 gpu 3 strategy tf distribute mirroredstrategy device device strategy tf distribute onedevicestrategy device device 0 accum count 8 with tf graph as default with strategy scope train op strategy extend call for each replica train op args accum count config tf compat v1 configproto allow soft placement true with tf compat v1 session config config as sess sess run tf compat v1 global variable initializer sess run train op the above code do not fail if onedevicestrategy be use or accum count be set to none other info log text traceback most recent call last file cond apply py line 27 in sess run tf compat v1 global variable initializer file usr local lib python3 5 dist package tensorflow python client session py line 930 in run run metadata ptr file usr local lib python3 5 dist package tensorflow python client session py line 1138 in run self graph fetch feed dict tensor feed handle feed handle file usr local lib python3 5 dist package tensorflow python client session py line 480 in init self assert fetchable graph fetch file usr local lib python3 5 dist package tensorflow python client session py line 497 in assert fetchable operation r have be mark as not fetchable op name valueerror operation init have be mark as not fetchable for an additional context I m port a code from v1 tf estimator and manual graph replication to v2 tf estimator and distribution strategy I m struggle to port the gradient accumulation code that we have work in v1 l199 l261 |
tensorflowtensorflow | link to the definiton tf image break | Bug | on the page the link under head module tf image be break |
tensorflowtensorflow | keras subclasse and explicit dtype of input | Bug | system information have I write custom code as oppose to use a stock example script provide in tensorflow os platform and distribution e g linux ubuntu 16 04 arch linux tensorflow version 2 0 0 alpha0 description when use kera subclasse there be no apparent way of define the dtype of the input node of the network in some case it would be neccecary to use tf float16 instead of 32 but as of now I can not find any way to adjust this also try to set the dtype use self dtype tf float16 be not permit |
tensorflowtensorflow | tensorflow core framework op def pb h 10 40 fatal error google protobuf port def inc no such file or directory | Bug | please make sure that this be a bug as per our github policy we only address code doc bug performance issue feature request and build installation issue on github tag bug template system information have I write custom code as oppose to use a stock example script provide in tensorflow use plugin os platform and distribution e g linux ubuntu 16 04 linux ubuntu 16 04 mobile device e g iphone 8 pixel 2 samsung galaxy if the issue happen on mobile device tensorflow instal from source or binary tensorflow tensorflow late gpu py3 tensorflow version use command below 1 14 1 dev20190409 python version 3 5 bazel version if compile from source gcc compiler version if compile from source cuda cudnn version gpu model and memory you can collect some of this information use our environment capture script you can also obtain the tensorflow version with python c import tensorflow as tf print tf git version tf version describe the current behavior as of 1 14 1 dev20190409 custom plugin fail to build with the follow error in file include from usr local lib python2 7 dist package tensorflow include tensorflow core framework op def builder h 24 0 from usr local lib python2 7 dist package tensorflow include tensorflow core framework op h 23 from horovod tensorflow mpi op cc 22 usr local lib python2 7 dist package tensorflow include tensorflow core framework op def pb h 10 40 fatal error google protobuf port def inc no such file or directory include compilation terminate this seem to be relate to it appear that not all the file be place in the proper location after the installation describe the expect behavior plugin should build successfully code to reproduce the issue provide a reproducible test case that be the bare minimum necessary to generate the problem docker run it rm tensorflow tensorflow late gpu py3 apt install y mpich horovod with tensorflow 1 pip install v horovod other info log include any log or source code that would be helpful to diagnose the problem if include traceback please include the full traceback large log and file should be attach cc martinwicke gunan |
tensorflowtensorflow | tf keras estimator model to estimator crashing when use tf distribute mirroredstrategy with only tensorflow native optimizer be support with distributionstrategy | Bug | system information have I write custom code as oppose to use a stock example script provide in tensorflow os platform and distribution e g linux ubuntu 16 04 mac os mobile device e g iphone 8 pixel 2 samsung galaxy if the issue happen on mobile device tensorflow instal from source or binary tensorflow version use command below 2 0 0 alpha0 python version 3 6 bazel version if compile from source gcc compiler version if compile from source cuda cudnn version gpu model and memory describe the current behavior I be use a keras model with tf 2 0 I be convert the model to estimator tf keras estimator model to estimator then it be crash when run estimator train model train when use tf distribute mirroredstrategy after migrate the code to tf 2 0 of course it work fine when use none as strategy I try to follow the instruction describe the expect behavior the same be work with tf 1 x code to reproduce the issue work in progress notebook can be find here I be use a very basic kera model with optimiser tf keras optimizer adam lr 0 01 beta 1 0 9 epsilon 1e 07 compile model model compile loss categorical crossentropy optimizer optimiser metric accuracy strategy none work strategy tf distribute mirroredstrategy crash with tf 2 0 but work with tf 1 x config tf estimator to use a give strategy training config tf estimator runconfig train distribute strategy model dir flags model dir save summary step 1 save checkpoint step 100 keep checkpoint max 3 log step count step 10 transfor kera model to estimator model estimator train model tf keras estimator model to estimator keras model model opt tf config training config fit the model use estimator train and datum dataset estimator train model train input fn lambda mnist v1 input mnist tfrecord dataset fn path train tfrecord flag mode tf estimator modekeys train batch size flag batch size step 1000 other info log i0409 22 06 27 293305 4531660224 model py 211 input dataset fn train train i0409 22 06 27 469371 123145492971520 estimator py 1126 call model fn i0409 22 06 27 475003 123145492971520 coordinator py 219 error report to coordinator only tensorflow native optimizer be support with distributionstrategy traceback most recent call last file user tarrade anaconda3 envs env gcp dl 2 0 alpha lib python3 6 site package tensorflow python training coordinator py line 297 in stop on exception yield file user tarrade anaconda3 envs env gcp dl 2 0 alpha lib python3 6 site package tensorflow python distribute mirror strategy py line 882 in run self main result self main fn self main args self main kwargs file user tarrade anaconda3 envs env gcp dl 2 0 alpha lib python3 6 site package tensorflow estimator python estimator estimator py line 1127 in call model fn model fn result self model fn feature feature kwargs file user tarrade anaconda3 envs env gcp dl 2 0 alpha lib python3 6 site package tensorflow estimator python estimator keras py line 278 in model fn raise valueerror only tensorflow native optimizer be support with valueerror only tensorflow native optimizer be support with distributionstrategy valueerror traceback most recent call last in anaconda3 envs env gcp dl 2 0 alpha lib python3 6 site package tensorflow estimator python estimator estimator py in train self input fn hook step max step save listener 357 358 save listener check listener type save listener 359 loss self train model input fn hook save listener 360 log info loss for final step s loss 361 return self anaconda3 envs env gcp dl 2 0 alpha lib python3 6 site package tensorflow estimator python estimator estimator py in train model self input fn hook save listener 1135 def train model self input fn hook save listener 1136 if self train distribution 1137 return self train model distribute input fn hook save listener 1138 else 1139 return self train model default input fn hook save listener anaconda3 envs env gcp dl 2 0 alpha lib python3 6 site package tensorflow estimator python estimator estimator py in train model distribute self input fn hook save listener 1198 self config train distribute configure self config session config 1199 return self actual train model distribute 1200 self config train distribute input fn hook save listener 1201 pylint enable protect access 1202 anaconda3 envs env gcp dl 2 0 alpha lib python3 6 site package tensorflow estimator python estimator estimator py in actual train model distribute self strategy input fn hook save listener 1267 label although this will be none it seem 1268 modekey train 1269 self config 1270 loss strategy reduce reduce util reduceop sum 1271 group estimator spec loss anaconda3 envs env gcp dl 2 0 alpha lib python3 6 site package tensorflow python distribute distribute lib py in call for each replica self fn args kwargs 1076 if kwargs be none 1077 kwargs 1078 return self call for each replica fn args kwargs 1079 1080 def call for each replica self fn args kwargs anaconda3 envs env gcp dl 2 0 alpha lib python3 6 site package tensorflow python distribute mirror strategy py in call for each replica self fn args kwargs 663 def call for each replica self fn args kwargs 664 return call for each replica self container strategy self device map 665 fn args kwargs 666 667 def configure self anaconda3 envs env gcp dl 2 0 alpha lib python3 6 site package tensorflow python distribute mirror strategy py in call for each replica distribution device map fn args kwargs 191 for t in thread 192 t should run set 193 coord join thread 194 195 return value regroup device map tuple t main result for t in thread anaconda3 envs env gcp dl 2 0 alpha lib python3 6 site package tensorflow python training coordinator py in join self thread stop grace period sec ignore live thread 387 self register thread set 388 if self exc info to raise 389 six reraise self exc info to raise 390 elif straggler 391 if ignore live thread anaconda3 envs env gcp dl 2 0 alpha lib python3 6 site package six py in reraise tp value tb 691 if value traceback be not tb 692 raise value with traceback tb 693 raise value 694 finally 695 value none anaconda3 envs env gcp dl 2 0 alpha lib python3 6 site package tensorflow python training coordinator py in stop on exception self 295 296 try 297 yield 298 except pylint disable bare except 299 self request stop ex sys exc info anaconda3 envs env gcp dl 2 0 alpha lib python3 6 site package tensorflow python distribute mirror strategy py in run self 880 self capture var scope reuse self replica i d 0 881 variable scope variable creator scope self variable creator fn 882 self main result self main fn self main args self main kwargs 883 self do true 884 finally anaconda3 envs env gcp dl 2 0 alpha lib python3 6 site package tensorflow estimator python estimator estimator py in call model fn self feature label mode config 1125 1126 log info calling model fn 1127 model fn result self model fn feature feature kwargs 1128 log info do call model fn 1129 anaconda3 envs env gcp dl 2 0 alpha lib python3 6 site package tensorflow estimator python estimator keras py in model fn feature label mode 276 not isinstance keras model optimizer 277 tf optimizer module optimizer optimizer tfoptimizer 278 raise valueerror only tensorflow native optimizer be support with 279 distributionstrategy 280 valueerror only tensorflow native optimizer be support with distributionstrategy |
tensorflowtensorflow | convert tensor to numpy array extremely slow in tf 2 0 | Bug | current issue bug report I try to convert a list of tensor to a numpy array this be no issue in tensorflow 1 13 x but be now order of magnitude slow in tensorflow 2 0 0 dev20190405 I link a piece of code to reproduce the issue in colab no hw acceleration require and to also show the difference in execution time between 1 13 x and 2 0 0 nightly I test the issue with np array list of tensor np shape list of tensor system information colab w o hw acceleration and tf 2 0 0 dev20190405 code to reproduce the issue code example |
tensorflowtensorflow | problem with keras sparse input | Bug | in keras backend graphexecutionfunction call function when this function receive sparse input it will transform sparse input into coo matrix and extract indice data shape from it but this function will try to invoke numpy asarray function on value tuple indice data shape and will cause error here def call self input input nest flatten input session get session feed array array val feed symbol symbol val for tensor value in zip self input input if value be none continue if be sparse tensor sparse coo value tocoo indice np concatenate np expand dim sparse coo row 1 np expand dim sparse coo col 1 1 value indice sparse coo datum sparse coo shape if tensor util be tensor value case feed symbolic tensor feed symbol append tensor symbol val append value else case feed numpy array feed array append tensor we need to do array conversion and type casting at this level since callable fn only support exact match tensor type dtype module as dtype tensor dtype should be that if be sparse tensor array val append value else array val append np asarray value dtype tensor type as numpy dtype |
tensorflowtensorflow | runtimeerror unable to create link name already exist during model save with modelcheckpoint | 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 arch linux mobile device e g iphone 8 pixel 2 samsung galaxy if the issue happen on mobile device no tensorflow instal from source or binary tf 2 0 download from repo tensorflow version use command below tf nightly gpu 2 0 preview 2 0 0 dev20190314 tensorflow hub 0 4 0 python version 3 6 cuda cudnn version cuda version 10 0 130 cudnn 7 5 0 gpu model and memory nvidia rtx 2080 ti 11 gb and gtx 1060 6 gb describe the current behavior I ve download an inception model from tf hub specifically this one I have add to it two keras layer a dropout layer and a dense layer and during the training I m try to save the model use the modelcheckpoint keras callback unfortunately after one epoch and during the model save I receive the follow error python traceback most recent call last file run medium federico xdata virtualenvs python36 tf2preview lib python3 6 site package ipython core interactiveshell py line 3291 in run code exec code obj self user global ns self user n file line 1 in runfile run medium federico xdata pycharmprojectsxdata ash ash prova gan plain test py wdir run medium federico xdata pycharmprojectsxdata ash ash file opt pycharm professional helper pydev pydev bundle pydev umd py line 197 in runfile pydev import execfile filename global var local var execute the script file opt pycharm professional helper pydev pydev imp pydev execfile py line 18 in execfile exec compile content n file exec glob loc file run medium federico xdata pycharmprojectsxdata ash ash prova gan plain test py line 98 in main file run medium federico xdata pycharmprojectsxdata ash ash prova gan plain test py line 90 in main logdir file run medium federico xdata pycharmprojectsxdata ash ash tester gan plain py line 85 in init self model self download and train model file run medium federico xdata pycharmprojectsxdata ash ash tester gan plain py line 255 in download and train model callback cback checkpoint file run medium federico xdata virtualenvs python36 tf2preview lib python3 6 site package tensorflow python keras engine training py line 1508 in fit generator step name step per epoch file run medium federico xdata virtualenvs python36 tf2preview lib python3 6 site package tensorflow python keras engine training generator py line 324 in model iteration callback on epoch end epoch epoch log file run medium federico xdata virtualenvs python36 tf2preview lib python3 6 site package tensorflow python keras callbacks py line 290 in on epoch end callback on epoch end epoch log file run medium federico xdata virtualenvs python36 tf2preview lib python3 6 site package tensorflow python keras callbacks py line 892 in on epoch end self model save weight filepath overwrite true file run medium federico xdata virtualenvs python36 tf2preview lib python3 6 site package tensorflow python keras engine network py line 1395 in save weight hdf5 format save weight to hdf5 group f self layer file run medium federico xdata virtualenvs python36 tf2preview lib python3 6 site package tensorflow python keras save hdf5 format py line 693 in save weight to hdf5 group param dset g create dataset name val shape dtype val dtype file run medium federico xdata virtualenvs python36 tf2preview lib python3 6 site package h5py hl group py line 139 in create dataset self name dset file run medium federico xdata virtualenvs python36 tf2preview lib python3 6 site package h5py hl group py line 371 in setitem h5o link obj i d self i d name lcpl lcpl lapl self lapl file h5py object pyx line 54 in h5py object with phil wrapper file h5py object pyx line 55 in h5py object with phil wrapper file h5py h5o pyx line 202 in h5py h5o link runtimeerror unable to create link name already exist highlight the error runtimeerror unable to create link name already exist describe the expect behavior I m expect to be able to use the modelcheckpoint callback to save the good model alternatively I m also expect to be able to save the model with the model save filepath function code to reproduce the issue you could reproduce the error on the follow tf hub colab page scrollto ccpdfxpsh47q add a cell with the modelcheckpoint code python cback checkpoint tf keras callbacks modelcheckpoint filepath good h5 verbose 1 save well only true and then add the callback to the fit generator function of the model python step per epoch train generator sample train generator batch size validation step valid generator sample valid generator batch size hist model fit generator train generator epoch 5 step per epoch step per epoch validation datum valid generator validation step validation step callback cback checkpoint history other info log I ve find some issue online regard a similar problem like this topic kera user my13tue2dlu and issue 5280 6844 and the more recent 26811 many of they speak about some problem regard the naming of the layer or weight or about create the modelcheckpoint with save well only true or save weight only true I have try all the propose approach but without success even with model save filepath keras function I face the same problem edit please follow this colab link to have all the code set up to be reproduce |
tensorflowtensorflow | nan appear on false fork of tf where propgate to e | Bug | same issue but that one be close the fix be to protect every single argument against the fork python import tensorflow as tf import numpy as np x tf convert to tensor np random randn 100 astype np float32 alpha tf variable 1 name asdf dtype tf float32 trainable true with tf gradienttape persistent true as t safe x tf where x 0 x tf one like x x safe x this fix it s tf where x 0 alpha tf math log x tf zero like x gradient t gradient s alpha in 326 reload d d gradient out 326 in 327 reload d d gradient out 327 in 346 sys version out 346 3 7 3 default mar 27 2019 16 54 48 n clang 4 0 1 tag release 401 final in 347 platform platform out 347 darwin 17 7 0 x86 64 i386 64bit in 348 tf version out 348 2 0 0 alpha0 |
tensorflowtensorflow | tf datum dataset shuffle produce the same result at each dataset iteration in tensorflow 2 alpha | Bug | system information have I write custom code as oppose to use a stock example script provide in tensorflow yes create a simple dataset shuffle it and iterate through it os platform and distribution e g linux ubuntu 16 04 ubuntu 18 04 tensorflow instal from source or binary pip tensorflow version use command below 2 0 0 alpha0 python version python 3 7 3 cuda cudnn version cuda10 0 not relevant gpu model and memory not relevant describe the current behavior when use tf datum dataset shuffle and iterate through the dataset multiple time the shuffled order be always the same describe the expect behavior the order should change for every iteration code to reproduce the issue provide a reproducible test case that be the bare minimum necessary to generate the problem import tensorflow as tf input array tf range 10 print input array dataset tf datum dataset from tensor slice input array shuffle 5 def print value for val in dataset print val numpy end print print value print value output tf tensor 0 1 2 3 4 5 6 7 8 9 shape 10 dtype int32 2 1 4 6 0 8 9 5 7 3 2 1 4 6 0 8 9 5 7 3 other info log behaviour be correct in tensorflow 1 13 |
tensorflowtensorflow | fix typo | Bug | ld lirary path ld library path 27620 |
tensorflowtensorflow | tf 2 0 api docs tf keras activation selu | Bug | system information tensorflow version 2 0 alpha doc link link should format the reference paper get rid of and add the author year of publish self normalize neural network klambauer et al 2017 definition the current definition do not specify the fix value for alpha and scale constant it also assume knowledge of the elu activation function we should also include a link I e propose modify definition the scale exponential linear unit selu activation function be scale x if x 0 and scale alpha exp x 1 if x 0 where alpha and scale be pre define constant alpha 1 6732632423543772848170429916717 and scale 1 0507009873554804934193349852946 the selu activation function multiplie scale 1 with the elu exponential linear unit elu to ensure a slope large than one for positive net input follow by what be already in the doc with the format lecun normal initialization bit and a link to it the value of alpha and scale be choose so that the mean and variance of the input be preserve between two consecutive layer as long as the weight be initialize correctly see lecun normal initialization and the number of input be large enough see reference for more information example can add a modify example fro the intro to cnns tutorial use selu instead of relu model model sequential model add layer conv2d 32 3 3 activation selu input shape 28 28 1 model add layer maxpooling2d 2 2 model add layer conv2d 64 3 3 activation selu model add layer maxpooling2d 2 2 model add layer conv2d 64 3 3 activation selu return can modify to include the word function the scale exponential unit activation function scale elu x alpha and format markdown for plain text raise not define visual should be add similar to we welcome contribution by user will you be able to update submit a pr use the doc style guide to fix the doc issue yes |
tensorflowtensorflow | minor doc fix | Bug | from 27627 |
tensorflowtensorflow | tf 2 0 api docs tf keras activation elu | Bug | system information tensorflow version 2 0 alpha doc link link link exist should format the link to the original paper delete also should add author and year of publish e g reference fast and accurate deep network learn by exponential linear unit elus clevert et al 2015 definition since it just say exponential linear unit we can modify it to the follow similar to the original paper the exponential linear unit elu with alpha 0 be x if x 0 and alpha exp x 1 if x 0 the elu hyperparameter alpha control the value to which an elu saturate for negative net input elus diminish the vanish gradient effect follow by word for word stuff from the original paper elu have negative value which push the mean of the activation close to zero mean activation that be close to zero enable fast learn as they bring the gradient close to the natural gradient elus saturate to a negative value when the argument get small saturation mean a small derivative which decrease the variation and the information that be propagate to the next layer example no example give can add a modify example from the intro to cnns tutorial where elu replace relu relu e g model model sequential model add layer conv2d 32 3 3 activation elu input shape 28 28 1 model add layer maxpooling2d 2 2 model add layer conv2d 64 3 3 activation elu model add layer maxpooling2d 2 2 model add layer conv2d 64 3 3 activation elu parameter both param define already but to aid the user it should state that alpha should be set at 1 0 by default and that alpha control the value to which an elu saturate for negative net input source paper instead of just alpha a scalar slope of negative section return define but for clarity should say the exponential linear unit elu activation function x if x 0 and alpha exp x 1 if x 0 instead of simply the exponential linear activation equation raise not define visual should be add to help the user example see p 5 of the original paper we welcome contribution by user will you be able to update submit a pr use the doc style guide to fix the doc issue yes |
tensorflowtensorflow | typo in spacetobatchnd and batchtospacend documentation | Bug | system information tensorflow version 1 13 doc link describe the documentation issue example 2 in say the output tensor have shape 4 1 1 3 and value 1 2 3 4 5 6 7 8 9 10 11 12 however that value s shape be 4 1 3 it appear the shape be correct but the value should be 1 2 3 4 5 6 7 8 9 10 11 12 the same issue be present in example 2 we welcome contribution by user will you be able to update submit a pr use the doc style guide to fix the doc issue yes |
tensorflowtensorflow | miss documentation for variable op | Bug | please make sure that this be a documentation issue as per our github policy we only address code doc bug performance issue feature request and build installation issue on github tag doc template system information tensorflow version v1 13 doc link describe the documentation issue hi I be look for documentation on the readvariableop and varhandleop specifically as to what be the input that they take and what do they produce exactly I dig around the source code for mention of these op and find some call here l196 and here l864 while this hint at the signature of these op I would love to look at the implementation which probably exist in this l37 module but I can not find it in this public repo s master branch could someone point I to this alternately be there any documentation on these op write this way thank we welcome contribution by user will you be able to update submit a pr use the doc style guide to fix the doc issue |
tensorflowtensorflow | doc keras incorrect comment in init of tf keras layers averagepooling1d | Bug | doc link argument describe the documentation issue see the comment below pool size integer size of the max pool window it should be average instead of max since this be average pooling furthermore the format of the description for input shape and output shape be break it be not render in a list format |
tensorflowtensorflow | doc keras incorrect comment in the example for tf keras layer add | Bug | doc link class add describe the documentation issue see the code example class add python add keras layer add x1 x2 equivalent to add keras layer add x1 x2 it should be add keras layer add x1 x2 equivalent to add keras layer add x1 x2 |
tensorflowtensorflow | doc keras predict of tf keras model sequential be render incorrectly | Bug | doc link predict describe the documentation issue the argument list be not render correctly see it below argument x input sample it could be a numpy array or array like or a list of array in case the model have multiple input a tensorflow tensor or a list of tensor in case the model have multiple input a tf datum dataset or a dataset iterator a generator or keras util sequence instance batch size integer or none number of sample per gradient update if unspecified batch size will default to 32 do not specify the batch size be your datum be in the form of symbolic tensor dataset dataset iterator generator or keras util sequence instance since they generate batch verbose verbosity mode 0 or 1 step total number of step batch of sample before declare the prediction round finish ignore with the default value of none max queue size integer use for generator or keras util sequence input only maximum size for the generator queue if unspecified max queue size will default to 10 worker integer use for generator or keras util sequence input only maximum number of process to spin up when use process base threading if unspecified worker will default to 1 if 0 will execute the generator on the main thread use multiprocesse boolean use for generator or keras util sequence input only if true use process base threading if unspecified use multiprocessing will default to false note that because this implementation rely on multiprocesse you should not pass non picklable argument to the generator as they can t be pass easily to child process it should be render in a list item format |
tensorflowtensorflow | doc keras init of tf keras loss binarycrossentropy be not render correctly | Bug | doc link class binarycrossentropy describe the documentation issue the init function of tf keras loss binarycrossentropy be not render correctly see the link here method |
tensorflowtensorflow | ld lirary path instead of ld library path | Bug | hello in tensorflow stream executor platform default dso loader cc line 50 ld lirary path be call instead of ld library path good |
tensorflowtensorflow | not support matrixbandpart in tflite | Bug | system information os platform and distribution e g linux ubuntu 16 04 ubuntu 16 04 tensorflow instal from source or binary binary tensorflow version 2 0alpha provide the text output from tflite convert here be a list of operator for which you will need custom implementation matrixbandpart |
tensorflowtensorflow | incomplete instruction for use image as metadata in | Bug | system information doc link describe the documentation issue the guide say to use image as metadata you must produce a single sprite image consist of small thumbnail one for each vector in the embed it then go on to explain what the sprite image should look like but it doesn t make clear if it need to be square where each row contain a number of thumbnail close to the square root of the total number of vector be any standard image format ok what do one do with the image file it doesn t work to choose it in the dialog that say step 2 optional load a tsv file of metadata if one create a config json file how should one refer to the sprite image url |
tensorflowtensorflow | e1102 non callable raise by pylint when extend from tf keras layer layer | Bug | please make sure that this be a bug as per our github policy we only address code doc bug performance issue feature request and build installation issue on github tag bug template 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 custom op container mobile device e g iphone 8 pixel 2 samsung galaxy if the issue happen on mobile device tensorflow instal from source or binary binary tensorflow version use command below nightly 2 0 preview python version the one with custom op container bazel version if compile from source gcc compiler version if compile from source cuda cudnn version gpu model and memory you can collect some of this information use our environment capture script you can also obtain the tensorflow version with python c import tensorflow as tf print tf git version tf version describe the current behavior as give in this page I be try to change the code in seq2seq as follow in decoder py class basedecoder tf keras layers layer however while run the pylint stage of sanity check it produce the follow error tensorflow addon seq2seq decoder test py 63 e1102 not callable decodernntest testdecodernn my decoder be not callable tensorflow addon seq2seq decoder test py 140 e1102 not callable decodernntest testdynamicdecodernnwithtraininghelpermatchesdynamicrnn my decoder be not callable describe the expect behavior the e1102 warning should not be raise code to reproduce the issue to reproduce in this pull request go to decoder py and change the signature of class in line 132 the warning disappear for pylint2 in current pull request as currently it import layer from tensorflow python keras and then extend layer layer other info log include any log or source code that would be helpful to diagnose the problem if include traceback please include the full traceback large log and file should be attach |
tensorflowtensorflow | tf upgrade v2 fail on google colab and late jupyter notebook | Bug | in all fairness tensorflow 2 be in alpha but the doc indicate the tf1 tf2 script should work system information use the code from I get syntax error on google colab and on jupyter notebook tensorflow instal from source or binary pip install tensorflow 2 0 0 alpha0 tensorflow version use command below mesh tensorflow 0 0 5 tensorflow 2 0 0a0 tensorflow estimator 1 13 0 tensorflow hub 0 4 0 tensorflow metadata 0 13 0 tensorflow probability 0 6 0 python version google api python client 1 6 7 ipython 5 5 0 ipython genutil 0 2 0 ipython sql 0 3 9 opencv contrib python 3 4 3 18 opencv python 3 4 5 20 python apt 1 6 3 ubuntu1 python chess 0 23 11 python dateutil 2 5 3 python louvain 0 13 python rtmidi 1 2 1 python slugify 3 0 2 python util 2 3 0 describe the current behavior test on jupyter notebook jupyter version 4 4 0 and google colab describe the expect behavior tf upgrade v2 should work in in jupyter and google colab code to reproduce the issue try the code yourself here other info log here be the source code here be the video show the result |
tensorflowtensorflow | no simple function optimisation example for tensorflow 2 0 | Bug | I don t see any simple function optimisation example for tensorflow 2 0 something like log x 2 could be good |
tensorflowtensorflow | doc keras model for predict function the list of argument be not display properly fullblock instead of list | Bug | system information tensorflow version 2 0 doc link describe the documentation issue in the documentation kera model for predict function the list of argument be not display properly fullblock instead of list on both chrome and firefox |
tensorflowtensorflow | update doc with code format for sparseconditionalaccumulator | Bug | small documentation improvement add code format for sparseconditionalaccumulator relate documentation |
tensorflowtensorflow | state build require bazel 0 23 0 for v2 0 0 alpha0 | Bug | please make sure that this be a documentation issue as per our github policy we only address code doc bug performance issue feature request and build installation issue on github tag doc template system information tensorflow version v2 0 0 alpha0 doc link install bazel describe the documentation issue I download the late at the time of write 0 24 1 bazel binary and when I try to configure py I get this you have bazel 0 24 1 instal please downgrade your bazel installation to version 0 23 0 or low to build tensorflow to downgrade download the installer for the old version from then run the installer I think it would be useful to state explicitly that 0 23 0 be the version require to build tensorflow to save some time I even just download 0 23 2 because I didn t see that the message say 0 23 0 and sure enough I get the same complaint about 0 23 2 we welcome contribution by user will you be able to update submit a pr use the doc style guide to fix the doc issue yes will do a pr shortly |
tensorflowtensorflow | improve description for tf test be gpu available | Bug | as a part of tensorflow doc sprint from munich a small explanation for add warning cc dynamicwebpaige |
tensorflowtensorflow | tf 2 0 0a0 tf function raise valueerror when compute gradient | Bug | please make sure that this be a bug as per our github policy we only address code doc bug performance issue feature request and build installation issue on github tag bug template 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 ubuntu 16 04 tensorflow version use command below pip install tensorflow gpu 2 0 0a0 python version 3 6 describe the current behavior the code execute normally but raise valueerror when compute gradient tape gradient if I decorate the training function with tf function the traceback be as follow valueerror traceback most recent call last workspace fgenl run py in 80 for batch I d in range num batch each epoch 81 batch datum datum generator get datum v2 82 loss output train one step batch datum v2 83 loss output input sess run opt op loss output batch datum 84 if loss metric be none pyenv version anaconda3 5 2 0 envs tf lib python3 6 site package tensorflow python eager def function py in call self args kwd 424 this be the first call of call so we have to initialize 425 initializer map 426 self initialize args kwd add initializer to initializer map 427 if self create variable 428 try pyenv version anaconda3 5 2 0 envs tf lib python3 6 site package tensorflow python eager def function py in initialize self args kwd add initializer to 368 self concrete stateful fn 369 self stateful fn get concrete function internal garbage collect pylint disable protect access 370 args kwd 371 372 def invalid creator scope unused args unused kwd pyenv version anaconda3 5 2 0 envs tf lib python3 6 site package tensorflow python eager function py in get concrete function internal garbage collect self args kwargs 1311 if self input signature 1312 args kwargs none none 1313 graph function self maybe define function args kwargs 1314 return graph function 1315 pyenv version anaconda3 5 2 0 envs tf lib python3 6 site package tensorflow python eager function py in maybe define function self args kwargs 1578 or call context key not in self function cache miss 1579 self function cache miss add call context key 1580 graph function self create graph function args kwargs 1581 self function cache primary cache key graph function 1582 return graph function args kwargs pyenv version anaconda3 5 2 0 envs tf lib python3 6 site package tensorflow python eager function py in create graph function self args kwargs override flat arg shape 1510 arg name arg name 1511 override flat arg shape override flat arg shape 1512 capture by value self capture by value 1513 self function attribute 1514 pyenv version anaconda3 5 2 0 envs tf lib python3 6 site package tensorflow python framework func graph py in func graph from py func name python func args kwargs signature func graph autograph autograph option add control dependency arg name op return value collection capture by value override flat arg shape 692 convert func 693 694 func output python func func args func kwargs 695 696 invariant func output contain only tensor indexedslice pyenv version anaconda3 5 2 0 envs tf lib python3 6 site package tensorflow python eager def function py in wrap fn args kwd 315 wrap allow autograph to swap in a converted function we give 316 the function a weak reference to itself to avoid a reference cycle 317 return weak wrap fn wrap args kwd 318 weak wrap fn weakref ref wrap fn 319 pyenv version anaconda3 5 2 0 envs tf lib python3 6 site package tensorflow python framework func graph py in wrapper args kwargs 684 optional feature autograph option 685 force conversion true 686 args kwargs 687 688 wrap around a decorator allow check like tf inspect getargspec pyenv version anaconda3 5 2 0 envs tf lib python3 6 site package tensorflow python autograph impl api py in convert call f owner option args kwargs 390 return call unconverted f args kwargs 391 392 result convert f effective args kwargs 393 394 the converted function s closure be simply insert into the function s tmp tmpx0xgcbu3 py in tf train one step batch datum 6 output ag convert call model none ag conversionoption recursive true verbose 0 strip decorator tf function defun ag convert ag do not convert ag convert call force conversion false optional feature internal convert user code true batch datum 7 loss info ag convert call calculate loss loss object ag conversionoption recursive true verbose 0 strip decorator tf function defun 1 ag convert ag do not convert ag convert call force conversion false optional feature internal convert user code true output batch datum 8 gradient ag convert call gradient tape ag conversionoption recursive true verbose 0 strip decorator tf function defun 2 ag convert ag do not convert ag convert call force conversion false optional feature internal convert user code true loss model trainable variable 9 update list grad var for grad var in ag convert call zip none ag conversionoption recursive true verbose 0 strip decorator tf function defun 3 ag convert ag do not convert ag convert call force conversion false optional feature internal convert user code true gradient model trainable variable if grad be not none 10 ag convert call apply gradient optimizer ag conversionoption recursive true verbose 0 strip decorator tf function defun 4 ag convert ag do not convert ag convert call force conversion false optional feature internal convert user code true update list pyenv version anaconda3 5 2 0 envs tf lib python3 6 site package tensorflow python autograph impl api py in convert call f owner option args kwargs 265 266 if not option force conversion and conversion be whiteliste for graph f 267 return call unconverted f args kwargs 268 269 internal convert user code be for example turn off when issue a dynamic pyenv version anaconda3 5 2 0 envs tf lib python3 6 site package tensorflow python autograph impl api py in call unconverted f args kwargs 186 return f self call args kwargs 187 188 return f args kwargs 189 190 pyenv version anaconda3 5 2 0 envs tf lib python3 6 site package tensorflow python eager backprop py in gradient self target source output gradient unconnected gradient 954 flat source 955 output gradient output gradient 956 unconnected gradient unconnected gradient 957 958 if not self persistent pyenv version anaconda3 5 2 0 envs tf lib python3 6 site package tensorflow python eager imperative grad py in imperative grad tape target source output gradient unconnected gradient 70 source 71 output gradient 72 compat as str unconnected gradient value pyenv version anaconda3 5 2 0 envs tf lib python3 6 site package tensorflow python eager backprop py in aggregate grad gradient 565 index slice op indexedslice 566 grad 567 math op range grad shape 0 568 constant op constant grad shape as list 569 index slice list append index slice pyenv version anaconda3 5 2 0 envs tf lib python3 6 site package tensorflow python op math ops py in range start limit delta dtype name 1258 with op name scope name range start limit delta as name 1259 start op convert to tensor start dtype dtype name start 1260 limit op convert to tensor limit dtype dtype name limit 1261 delta op convert to tensor delta dtype dtype name delta 1262 pyenv version anaconda3 5 2 0 envs tf lib python3 6 site package tensorflow python framework op py in convert to tensor value dtype name prefer dtype dtype hint 1048 prefer dtype deprecation deprecate argument lookup 1049 dtype hint dtype hint prefer dtype prefer dtype 1050 return convert to tensor v2 value dtype prefer dtype name 1051 1052 pyenv version anaconda3 5 2 0 envs tf lib python3 6 site package tensorflow python framework op py in convert to tensor v2 value dtype dtype hint name 1106 name name 1107 prefer dtype dtype hint 1108 as ref false 1109 1110 pyenv version anaconda3 5 2 0 envs tf lib python3 6 site package tensorflow python framework op py in internal convert to tensor value dtype name as ref prefer dtype ctx accept symbolic tensor 1184 1185 if ret be none 1186 ret conversion func value dtype dtype name name as ref as ref 1187 1188 if ret be notimplemente pyenv version anaconda3 5 2 0 envs tf lib python3 6 site package tensorflow python framework constant op py in constant tensor conversion function v dtype name as ref 302 as ref false 303 as ref 304 return constant v dtype dtype name name 305 306 pyenv version anaconda3 5 2 0 envs tf lib python3 6 site package tensorflow python framework constant op py in constant value dtype shape name 243 244 return constant impl value dtype shape name verify shape false 245 allow broadcast true 246 247 pyenv version anaconda3 5 2 0 envs tf lib python3 6 site package tensorflow python framework constant op py in constant impl value dtype shape name verify shape allow broadcast 281 tensor util make tensor proto 282 value dtype dtype shape shape verify shape verify shape 283 allow broadcast allow broadcast 284 dtype value attr value pb2 attrvalue type tensor value tensor dtype 285 const tensor g create op pyenv version anaconda3 5 2 0 envs tf lib python3 6 site package tensorflow python framework tensor util py in make tensor proto value dtype shape verify shape allow broadcast 453 else 454 if value be none 455 raise valueerror none value not support 456 if dtype be provide force numpy array to be the type 457 provide if possible valueerror none value not support describe the expect behavior the code should also execute normally when use tf function code to reproduce the issue sorry I do not have a simple snippet to reproduce this issue but could you find something in the traceback see below please |
tensorflowtensorflow | add oxford comma in readme md | Bug | add an oxford comma to line 28 to be consistent with the rest of the file |
tensorflowtensorflow | tf contrib label tensor incorrectly handle or fail with none dimension | Bug | system information have I write custom code as oppose to use a stock example script provide in tensorflow yes os platform and distribution macos sierra tensorflow instal from source or binary binary tensorflow version use command below 1 12 python version 2 7 label tensor mul fail with none dimension when tf multiply doesn t ipdb x ipdb y ipdb x y valueerror mismatch a axis on input tensor axis a dimension 1 and axis a dimension none ipdb lt mul x y valueerror mismatch a axis on input tensor axis a dimension 1 and axis a dimension none ipdb tf multiply x y |
tensorflowtensorflow | tf 2 0 0 alpha0 tf keras model reinitialize set weight | Bug | system information have I write custom code yes os platform and distribution e g linux ubuntu 16 04 linux ubuntu 16 04 tensorflow instal from source or binary pip install tensorflow 2 0 0 alpha0 tensorflow version use command below 2 0 0 alpha0 python version 3 7 1 use only cpu describe the current behavior when I initialize a convolution layer with some weight instead of the weight be set it will be randomly re initialize print output return at each run a different weight code to reproduce the issue import tensorflow kera import numpy as np weight np ndarray shape 1 1 1 1 dtype float32 weight 0 0 0 0 0 8888 bias np ndarray shape 1 dtype float32 bias 0 0 weight weight bias input tensorflow keras layer input shape 1 1 1 name input output tensorflow keras layer conv2d filter 1 kernel size 1 stride 1 1 kernel regularizer tensorflow keras regularizer l2 0 0005 weight weight use bias true activation none pad same input model tensorflow keras model inputs input output output for layer in model layer for weight in layer get weight print weight wrong output 1 6804 0 describe the expect behavior with keras version 2 1 5 use tensorflow backend version 1 6 0 when I initialize a convolution layer with some weight the weight will be set and stick print output return as expect 0 8888 0 this be an useful feature especially when one be convert from one model format to another for example when convert darknet model cfg model weight to keras model h5 code to reproduce import keras import numpy as np weight np ndarray shape 1 1 1 1 dtype float32 weight 0 0 0 0 0 8888 bias np ndarray shape 1 dtype float32 bias 0 0 weight weight bias input keras layer input shape 1 1 1 name input output keras layer conv2d filter 1 kernel size 1 stride 1 1 kernel regularizer keras regularizer l2 0 0005 weight weight use bias true activation none pad same input model keras model inputs input output output for layer in model layer for weight in layer get weight print weight correct output 0 8888 0 question be this intend or how can this be disable how can I help you solve this issue thank for take your time look into this |
tensorflowtensorflow | filesystem destructor be not be call when do | Bug | system information have I write custom code as oppose to use a stock example script provide in tensorflow no os platform and distribution e g linux ubuntu 16 04 cento 7 mobile device e g iphone 8 pixel 2 samsung galaxy if the issue happen on mobile device tensorflow instal from source or binary binary tensorflow version use command below v1 13 1 python version python 3 6 3 bazel version if compile from source gcc compiler version if compile from source cuda cudnn version gpu model and memory when a filesystem be register it be allocate in the heap but not destroy on exit when use a python script the filesystem destructor filesystem must be call on exit code to reproduce the issue opt rh rh python36 root usr lib python3 6 site package tensorflow include tensorflow core platform file system h 2019 04 05 10 05 27 253673315 0200 test tensorflow tensorflow core platform file system h 2019 04 05 10 22 15 999269084 0200 225 7 225 7 filesystem printf create filesystem n virtual filesystem virtual filesystem printf destroy filesystem n then we run python and load our filesystem plugin alexey workstation test python c import tensorflow as tf tf load library build lib libfilesystem tensorflow so create filesystem alexey workstation test other info log it be necessary for our custom filesystem plugin to call some function on destruction it seem that c style pointer be use when register a filesystem l450 l457 l336 the scheme be actually map to a unique pointer to the class l64 but I suspect that move could be the issue l114 |
tensorflowtensorflow | what s the license of pre train tensorflow js model | Bug | I want to use tfjs model in a commercial application while source code in the tfjs model repository be license under apache 2 0 I do not find any info about the license of the model host on storage googleapis com for example body pix be load its model from storage googleapis com tfjs model e g posenet mobilenet 025 partmap model json posenet mobilenet 025 partmap tensorflowjs model p posenet mobilenet 025 partmap weight manifest json be the file on storage googleapis com tfjs model license under apache 2 0 as well if yes it would be great to document it and I m willing to submit a pr |
tensorflowtensorflow | duplicate layer name when use a b and add c d in the same keras model | 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 macosx 10 13 6 mobile device e g iphone 8 pixel 2 samsung galaxy if the issue happen on mobile device no tensorflow instal from source or binary binary tensorflow version use command below tf version version 2 0 0 dev20190404 tf version git version v1 12 0 11729 g98c3cfbf74 python version 3 6 8 bazel version if compile from source n a gcc compiler version if compile from source n a cuda cudnn version n a gpu model and memory n a describe the current behavior when use a b then keras layer add in the same keras model I get an exception say the name add be use twice the problem go away if I use keras layer add first or if I use twice the same operation either twice or keras layer add twice describe the expect behavior I don t expect any name conflict code to reproduce the issue the follow code raise a valueerror exception full stacktrace below python from tensorflow import kera input keras layer input shape 2 add1 input input add2 keras layer add input input model keras model model input input output add1 add2 other info log here be the full stacktrace python valueerror traceback most recent call last in 4 add1 input input 5 add2 keras layer add input input 6 model keras model model input input output add1 add2 tensorflow python keras engine training py in init self args kwargs 121 122 def init self args kwargs 123 super model self init args kwargs 124 initialize distribution strategy here since it be possible to call 125 predict on a model without compile it tensorflow python keras engine network py in init self args kwargs 137 input in kwargs and output in kwargs 138 graph network 139 self init graph network args kwargs 140 else 141 subclasse network tensorflow python training tracking base py in method wrapper self args kwargs 456 self setattr track false pylint disable protect access 457 try 458 result method self args kwargs 459 finally 460 self setattr track previous value pylint disable protect access tensorflow python keras engine network py in init graph network self input output name 284 keep track of the network s node and layer 285 node node by depth layer layer by depth map graph network 286 self input self output 287 self network nodes nod 288 self node by depth node by depth tensorflow python keras engine network py in map graph network input output 1916 if all name count name 1 1917 raise valueerror the name name be use 1918 str all name count name time in the model 1919 all layer name should be unique 1920 return network nodes node by depth layer layer by depth valueerror the name add be use 2 time in the model all layer name should be unique |
tensorflowtensorflow | tf 2 0 0 alpha0 memory leak with tf keras model load model | Bug | system information have I write custom code yes os platform and distribution e g linux ubuntu 16 04 macos mojave 10 14 3 mobile device e g iphone 8 pixel 2 samsung galaxy if the issue happen on mobile device tensorflow instal from source or binary docker image tensorflow tensorflow 2 0 0a0 py3 jupyter tensorflow version 2 0 0a0 py3 jupyter python version 3 5 use only cpu describe the current behavior I have a memory leak when I load several kera model with load model function previously save with model save here be what I obtain when I load 20 time a model the same model for the example the output be obtain with the memory profiler library see code below first iteration line mem usage increment line content 14 229 5 mib 229 5 mib profile 15 def load model keras model dir 16 232 5 mib 2 9 mib k clear session 17 252 9 mib 20 4 mib model load model model dir 18 252 9 mib 0 0 mib return model after 20 iteration line mem usage increment line content 14 539 9 mib 539 9 mib profile 15 def load model keras model dir 16 539 9 mib 0 0 mib k clear session 17 566 4 mib 26 5 mib model load model model dir 18 566 4 mib 0 0 mib return model describe the expect behavior I do not have this problem with tensorflow 1 13 1 here be what I obtain with this version and the behavior I expect to have first iteration line mem usage increment line content 14 210 7 mib 210 7 mib profile 15 def load model keras model dir 16 214 2 mib 3 4 mib k clear session 17 239 8 mib 25 6 mib model load model model dir 18 239 8 mib 0 0 mib return model after 20 iteration line mem usage increment line content 14 257 9 mib 257 9 mib profile 15 def load model keras model dir 16 257 9 mib 0 0 mib k clear session 17 259 0 mib 1 1 mib model load model model dir 18 259 0 mib 0 0 mib return model code to reproduce the issue import os import tensorflow as tf from tensorflow keras model import load model import tensorflow kera backend as k from memory profiler import profile model dir model path h5 profile def load model keras model dir k clear session model load model model dir return model for I in range 100 print I load model keras model dir other info log I also try to put within the function model none del model gc collect import gc with no effect |
tensorflowtensorflow | optimizer v2 apply gradient surprising behavior not explicitly document | Bug | system information tensorflow version 2 0 0 aplha0 doc link apply gradient I be implement a dd policy gradient for reinforcement learning and have a bug where my agent would minimize the reward instead of maximize it it turn out optimizer apply gradient follow negative of the gradient I give to it this be a surprising behavior when use this function separately outside of minimize context documentation do not mention anywhere that negative gradient be follow by this method have to use a unit test to clarify the behavior opt tf optimizer adam x tf variable 1 dtype tf float32 dx tf one 1 dtype tf float32 opt apply gradient dx x assert x numpy 0 1 there be also no flag to invert this behavior and follow positive gradient I have to pass negative gradient of the expect reward to follow it in positive direction def train actor self sar obs1 action reward obs2 sar with tf gradienttape as tape would do action self actor obs1 score tf reduce mean self critic observation obs1 action would do action invert score tf optimizer follow negative of the provide gradient for this reason we provide negative gradient of the score it will result in positive gradient be follow grad tape gradient invert self actor trainable weight self optimizer apply gradient zip grad self actor trainable weight |
tensorflowtensorflow | tf reduce sum with multiple negative axis and tf raggedtensor bug | 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 ubuntu 16 04 tensorflow instal from source or binary pip tensorflow version use command below python version v1 13 1 0 g6612da8951 1 13 1 gpu model and memory quadro k620 2 gb current behaviour multiple axis pass to tf reduce sum with first argument be a ragged tensor result in incorrect behaviour python import tensorflow as tf x value tf random normal shape 100 5 6 x row length tf constant 20 30 50 dtype tf int64 x ragged tf raggedtensor from row length x value x row length print x ragged shape 3 5 6 wrong shape print tf reduce sum x rag axis 2 3 shape 50 6 positive axis work print tf reduce sum x rag axis 1 2 shape 3 6 separate reduction work print tf reduce sum tf reduce sum x rag axis 3 axis 2 shape 3 6 expect behaviour same result as correspond positive index separate reduction |
tensorflowtensorflow | can not set dynamic property on custom layer | Bug | system information have I write custom code as oppose to use a stock example script provide in tensorflow no os platform and distribution e g linux ubuntu 16 04 colab environment tensorflow version use command below 2 0 alpha I be try to write a custom layer in tf 2 0 and have trouble make it run eagerly for example if I open a documentation page and try to use the follow piece of code from that page to create custom layer python class mysimplelayer tf keras layers layer def init self output unit super mysimplelayer self init self output unit output unit self dynamic true def build self input shape the build method get call the first time your layer be use create variable on build allow you to make their shape depend on the input shape and hence remove the need for the user to specify full shape it be possible to create variable during init if you already know their full shape self kernel self add variable kernel input shape 1 self output unit def call self input override call instead of call so we can perform some bookkeeping return tf matmul input self kernel and then try to create an instance for example like this python mysimplelayer 10 I get the follow error message python attributeerror traceback most recent call last in 1 mysimplelayer 10 in init self output unit 3 super mysimplelayer self init 4 self output unit output unit 5 self dynamic true 6 7 def build self input shape usr local lib python3 6 dist package tensorflow python keras engine base layer py in setattr self name value 1778 exclude property setter from track 1779 hasattr self class name 1780 super layer self setattr name value 1781 return 1782 attributeerror can t set attribute this error appear in colab as well as on my machine window 8 be I do something wrong here |
tensorflowtensorflow | to build with g option | Bug | please make sure that this be a build installation issue as per our github policy we only address code doc bug performance issue feature request and build installation issue on github tag build template system information os platform and distribution e g linux ubuntu 16 04 linux ubuntu 16 04 mobile device e g iphone 8 pixel 2 samsung galaxy if the issue happen on mobile device tensorflow instal from source or binary source tensorflow version 2eebcc63a6cd3aad483bf7c1cb25df2b8780ef67 python version python 3 5 2 instal use virtualenv pip conda virtualenv bazel version if compile from source bazel 0 24 0 installer linux x86 64 sh gcc compiler version if compile from source gcc version 5 4 0 20160609 ubuntu 5 4 0 6ubuntu1 16 04 10 cuda cudnn version gpu model and memory describe the problem my purpose be to build tensorflow c library libtensorflow framework so and libtensorflow so to be step into within gdb via source code I try two method 1 add g o0 option during configure image and then run the below command to build the c librarie bazel test c opt tensorflow tool lib package libtensorflow test bazel build c opt tensorflow tool lib package libtensorflow but gdb show that it be unable to read the symbol as below 0x00007fffedea2400 0x00007ffff41f73ef yes usr local lib libtensorflow so 0x00007fffe9245bc0 0x00007fffea040ea0 yes usr local lib libtensorflow framework so 2 I try dbg version with bazel test c dbg tensorflow tool lib package libtensorflow test and it never stop with info build option compilation mode have change discard analysis cache info analyse target tensorflow tool lib package libtensorflow test 0 package load 15440 target configure info find 1 test target 7 070 7 073 link tensorflow libtensorflow so 14876s local c bazel catch interrupt signal shut down and the build result be weird file bazel out k8 dbg bin tensorflow libtensorflow so bazel out k8 dbg bin tensorflow libtensorflow so datum btw my build command come from |
tensorflowtensorflow | tf print the example doesn t work with tf 2 0 | Bug | please make sure that this be a documentation issue as per our github policy we only address code doc bug performance issue feature request and build installation issue on github tag doc template system information tensorflow version 2 0 doc link describe the documentation issue the example doesn t work with tf 2 0 python tf enable eager execution tf contrib eager defun def f tensor tf range 10 tf print tensor output stream sys stderr return tensor range tensor f python sess tf session with sess as default tensor tf range 10 print op tf print tensor tensor 2 tensor 2 output stream sys stdout with tf control dependency print op triple tensor tensor 3 sess run triple tensor we welcome contribution by user will you be able to update submit a pr use the doc style guide to fix the doc issue |
tensorflowtensorflow | doc 1 x save keras model require eager mode | Bug | system information tensorflow version 1 13 1 doc link describe the documentation issue I follow this tutorial and this api reference page to save a train keras model as a savedmodel which in my use case be require to upload it to gcp ml engine but I encounter a non document assertionerror specify I should use eager mode this seem relate to this commit do I understand something wrong or should this compatibility requirement also be document for save keras model at least with serve only set to true |
tensorflowtensorflow | tf 2 0 api docs tf keras sequential | Bug | system information tensorflow version 2 0 alpha doc link describe the documentation issue description the order of property and method be alphabetical probably by tf doc design with a large doc for an arguably popular module such as tf keras sequential navigate around it may be confusing to the user perhaps we should start with the most important one e g compile and fit under method since tf keras sequential like many other module class expand on keras sequential it would still be logical to list certain property method etc in the beginning at the top of the page in order of importance for well ux see as a good example where these be not in alphabetical order also in the beginning after a short intro inherit from model a linear stack of layer a link to build a simple model with tf keras sequential in tensorflow would be cool for those who be new to tf keras tf sequential model also the official kera io doc include a url in to get start with the keras sequential model example there already be an example right in the beginning maybe add add a separate link to or copy an example from this sequential notebook would be cool too source sequential model parameter inconsistent some reformatte may be need e g input shape input shape also both argument and args be use throughout the doc return raise sometimes exist sometimes not create a ux issue because of inconsistency visual a simple one similar to would be appreciate we welcome contribution by user will you be able to update submit a pr use the doc style guide to fix the doc issue for sure |
tensorflowtensorflow | tf 2 0 api docs tf keras model | Bug | system information tensorflow version 2 0 alpha doc link describe the documentation issue link a link to keras functional api in tensorflow would be appreciate for those who be new to tf kera and tensorflow 1 x and 2 0 e g see this keras model class api doc model class api description take into account this be a heavy module it may not be easy to write good doc for tf keras model since tf keras be a crucial api to tf 2 0 so let s make the doc a delight to read the current documentation for tf keras model be a bit incomprehensible for a novice or experienced user in term of user experience for someone who be new or not new to tf maybe the description should be improve and follow the keras io doc more closely for instance argument under compile be apparently copy paste from keras model doc however you have to scroll all the way down to see they here compile update maybe because method etc be list in alphabetical order which be not intuitive the description at keras io be neater and more organized imo it start with a short description and jump to args from compile I d suggest we follow the same structure also as in we should include a link to the guide to keras functional api in tensorflow which be quite well write example not enough example they be mention here and there see ux issue above under description recommend to rewrite it to follow the original keras model module model class api along with example parameter return raise ux issue see description above visual recommend to add visual from we welcome contribution by user will you be able to update submit a pr use the doc style guide to fix the doc issue yes let s make tf keras docs awesome |
tensorflowtensorflow | tf2 0 gradient problem of use tf nn relu in tf keras model | Bug | system information os platform and distribution linux ubuntu 18 04 tensorflow instal from binary tensorflow version 2 0 0 alpha0 python version 3 6 8 describe the current behavior I build a keras model with only a tf nn relu but the gradient seem to be none after be decorate by tf function code to reproduce the issue 1 tf nn relu tf keras model tf function this be the only case that produce none gradient python import tensorflow as tf z tf keras input h tf nn relu z m tf keras model z h tf function def f x with tf function with tf gradienttape as t t watch x z m x 2 return t gradient z x print f tf convert to tensor 10 0 none 1 2 tf nn relu tf keras model without tf function python def f x without tf function with tf gradienttape as t t watch x z m x 2 return t gradient z x print f tf convert to tensor 10 0 tf tensor 20 0 shape dtype float32 2 tf keras layers relu tf keras model tf function python import tensorflow as tf z tf keras input h tf keras layers relu z m tf keras model z h tf function def f x with tf function with tf gradienttape as t t watch x z m x 2 return t gradient z x print f tf convert to tensor 10 0 tf tensor 20 0 shape dtype float32 2 2 tf keras layers relu tf keras model without tf function python def f x without tf function with tf gradienttape as t t watch x z m x 2 return t gradient z x print f tf convert to tensor 10 0 tf tensor 20 0 shape dtype float32 3 only tf nn relu python import tensorflow as tf m tf nn relu tf function def f x with tf function with tf gradienttape as t t watch x z m x 2 return t gradient z x print f tf convert to tensor 10 0 tf tensor 20 0 shape dtype float32 so I think its the problem between tf nn relu and tf keras model besides tf nn tanh have the same problem |
tensorflowtensorflow | attributeerror module tensorflow have no attribute enable eager execution in verify installation | Bug | please make sure that this be a build installation issue as per our github policy we only address code doc bug performance issue feature request and build installation issue on github tag build template system information os platform and distribution e g linux ubuntu 16 04 window 10 pro mobile device e g iphone 8 pixel 2 samsung galaxy if the issue happen on mobile device not applicable tensorflow instal from source or binary binary tensorflow version tensorflow gpu 2 0 0a0 python version 3 6 7 instal use virtualenv pip conda virtualenv pip bazel version if compile from source gcc compiler version if compile from source cuda cudnn version 10 0 7 5 0 56 gpu model and memory nvidia gtx 1080 max q describe the problem I ve just follow the installation guide for tensorflow 2 0 see above from software requirement the final step be to verify the install see the pip link above which give I the follow output python c import tensorflow as tf tf enable eager execution print tf reduce sum tf random normal 1000 1000 traceback most recent call last file line 1 in attributeerror module tensorflow have no attribute enable eager execution which I do not expect hope to see provide the exact sequence of command step that you execute before run into the problem follow the guide mention above create all the environment variable path within venv then contain d dev project project venv d dev tool cuda cudnn 10 0 windows10 x64 v7 5 0 56 cuda bin c program files nvidia gpu computing toolkit cuda v10 0 bin c program files nvidia gpu computing toolkit cuda v10 0 extras cupti libx64 c program files nvidia gpu computing toolkit cuda v10 0 include c program files nvidia gpu computing toolkit cuda v10 0 libnvvp c programdata dockerdesktop version bin c program file docker docker resource bin d dev tool oraclexe oraclexe app oracle product 11 2 0 server bin c program file python36 script c program file python36 c program file microsoft mpi bin c program file x86 common file oracle java javapath c programdata oracle java javapath c window system32 c window c window system32 wbem c window system32 windowspowershell v1 0 c program file x86 nvidia corporation physx common c program file dotnet c program file microsoft sql server 130 tool binn d dev tool putty c program file tortoisesvn bin c program files nvidia corporation nvidia nvdlisr c program file nodejs c window system32 c window c window system32 wbem c window system32 windowspowershell v1 0 c window system32 openssh c dev tool python python3 6 7 script c dev tool python python3 6 7 c program file python36 script c program file python36 c user tve21314 appdata local microsoft windowsapps d dev tool maven apache maven 3 5 3 bin c program file java jdk1 8 0 201 bin c program file java jre1 8 0 201 bin c program file git bin c user tve21314 appdata local program microsoft vs code bin c user tve21314 appdata roam npm finally run from within the venv python c import tensorflow as tf tf enable eager execution print tf reduce sum tf random normal 1000 1000 any other info log include any log or source code that would be helpful to diagnose the problem if include traceback please include the full traceback large log and file should be attach the traceback again traceback most recent call last file line 1 in attributeerror module tensorflow have no attribute enable eager execution if I use python c import tensorflow as tf print tf reduce sum tf random normal 1000 1000 the error become traceback most recent call last file line 1 in attributeerror module tensorflow have no attribute random normal |
tensorflowtensorflow | use layer class as attribute throw an exception | Bug | system information have I write custom code yes os platform and distribution ubuntu 16 04 tensorflow instal from binary tensorflow version 2 0 0 dev20190402 python version 2 7 12 describe the current behavior when a layer class be use as attribute the code will throw a typeerror exception when call self gather child attribute it appear that the layer class be track describe the expect behavior only layer instance should be track not class code to reproduce the issue python import tensorflow as tf class layer tf keras layers layer def init self super layer self init self layer fn tf keras layer dense layer layer print layer variable other info log text traceback most recent call last file tf2 class py line 10 in print layer variable file tmp tf2 local lib python2 7 site package tensorflow python keras engine base layer py line 1330 in variable return self weight file tmp tf2 local lib python2 7 site package tensorflow python keras engine base layer py line 708 in weight return self trainable weight self non trainable weight file tmp tf2 local lib python2 7 site package tensorflow python keras engine base layer py line 687 in trainable weight nest self gather child attribute trainable weight file tmp tf2 local lib python2 7 site package tensorflow python keras engine base layer py line 1850 in gather child attribute getattr layer attribute for layer in nest layer typeerror property object be not iterable |
tensorflowtensorflow | valueerror invalid tensor mul be find | Bug | this template be for miscellaneous issue not cover by the other issue category for question on how to work with tensorflow or support for problem that be not verify bug in tensorflow please go to stackoverflow if you be report a vulnerability please use the dedicated reporting process for high level discussion about tensorflow please post to for question about the development or internal working of tensorflow or if you would like to know how to contribute to tensorflow please post to python 2 7 tensorflow 1 9 python script input array mul output array final result converter tf contrib lite tococonverter from frozen graph graph def file input array output array tflite model converter convert open model tflite wb write tflite model it be work fine a month before now it show the follow error file home ubuntu local lib python2 7 site package tensorflow contrib lite python convert save model py line 205 in get tensor from tensor name join invalid tensor valueerror invalid tensor mul be find |
tensorflowtensorflow | tensorflow 2 0 0 multiple gpu output zero for non root gpu | Bug | please make sure that this be a bug as per our github policy we only address code doc bug performance issue feature request and build installation issue on github tag bug template system information have I write custom code as oppose to use a stock example script provide in tensorflow no os platform and distribution e g linux ubuntu 16 04 linux ubuntu 18 04 mobile device e g iphone 8 pixel 2 samsung galaxy if the issue happen on mobile device no tensorflow instal from source or binary source tensorflow version use command below v2 0 0 alpha0 0 g2c319fb415 2 0 0 alpha0 python version 3 7 bazel version if compile from source 0 23 0 gcc compiler version if compile from source gcc 4 8 cuda cudnn version cuda 10 0 cudnn 7 gpu model and memory a simple model can reproduce you can collect some of this information use our environment capture script you can also obtain the tensorflow version with python c import tensorflow as tf print tf git version tf version when I run the tf env collect sh script tensorflow just hang there forever I couldn t even kill it what the hell be go on bash base tfenv sh collect system information 2019 04 02 22 07 13 977020 I tensorflow stream executor platform default dso loader cc 42 successfully open dynamic library libcuda so 1 2019 04 02 22 07 15 731652 I tensorflow compiler xla service service cc 162 xla service 0x56368643d070 execute computation on platform cuda device 2019 04 02 22 07 15 731716 I tensorflow compiler xla service service cc 169 streamexecutor device 0 geforce gtx titan x compute capability 5 2 2019 04 02 22 07 15 731733 I tensorflow compiler xla service service cc 169 streamexecutor device 1 geforce gtx titan x compute capability 5 2 2019 04 02 22 07 15 731751 I tensorflow compiler xla service service cc 169 streamexecutor device 2 titan x pascal compute capability 6 1 2019 04 02 22 07 15 731765 I tensorflow compiler xla service service cc 169 streamexecutor device 3 titan x pascal compute capability 6 1 2019 04 02 22 07 15 731780 I tensorflow compiler xla service service cc 169 streamexecutor device 4 titan x pascal compute capability 6 1 2019 04 02 22 07 15 731795 I tensorflow compiler xla service service cc 169 streamexecutor device 5 geforce gtx titan x compute capability 5 2 2019 04 02 22 07 15 731810 I tensorflow compiler xla service service cc 169 streamexecutor device 6 geforce gtx titan x compute capability 5 2 2019 04 02 22 07 15 757870 I tensorflow core platform profile util cpu util cc 94 cpu frequency 2299805000 hz 2019 04 02 22 07 15 761353 I tensorflow compiler xla service service cc 162 xla service 0x563686573540 execute computation on platform host device 2019 04 02 22 07 15 761414 I tensorflow compiler xla service service cc 169 streamexecutor device 0 2019 04 02 22 07 15 762076 I tensorflow core common runtime gpu gpu device cc 1467 find device 0 with property name geforce gtx titan x major 5 minor 2 memoryclockrate ghz 1 076 pcibusid 0000 08 00 0 totalmemory 11 93gib freememory 11 81gib 2019 04 02 22 07 15 762510 I tensorflow core common runtime gpu gpu device cc 1467 find device 1 with property name geforce gtx titan x major 5 minor 2 memoryclockrate ghz 1 076 pcibusid 0000 09 00 0 totalmemory 11 93gib freememory 11 81gib 2019 04 02 22 07 15 762924 I tensorflow core common runtime gpu gpu device cc 1467 find device 2 with property name titan x pascal major 6 minor 1 memoryclockrate ghz 1 531 pcibusid 0000 82 00 0 totalmemory 11 91gib freememory 11 76gib 2019 04 02 22 07 15 763343 I tensorflow core common runtime gpu gpu device cc 1467 find device 3 with property name titan x pascal major 6 minor 1 memoryclockrate ghz 1 531 pcibusid 0000 85 00 0 totalmemory 11 91gib freememory 11 76gib 2019 04 02 22 07 15 763634 I tensorflow core common runtime gpu gpu device cc 1467 find device 4 with property name titan x pascal major 6 minor 1 memoryclockrate ghz 1 531 pcibusid 0000 86 00 0 totalmemory 11 91gib freememory 5 41gib 2019 04 02 22 07 15 764061 I tensorflow core common runtime gpu gpu device cc 1467 find device 5 with property name geforce gtx titan x major 5 minor 2 memoryclockrate ghz 1 076 pcibusid 0000 89 00 0 totalmemory 11 93gib freememory 11 81gib 2019 04 02 22 07 15 764483 I tensorflow core common runtime gpu gpu device cc 1467 find device 6 with property name geforce gtx titan x major 5 minor 2 memoryclockrate ghz 1 076 pcibusid 0000 8a 00 0 totalmemory 11 93gib freememory 11 81gib 2019 04 02 22 07 15 770500 I tensorflow core common runtime gpu gpu device cc 1546 add visible gpu device 0 1 2 3 4 5 6 2019 04 02 22 07 15 770587 I tensorflow stream executor platform default dso loader cc 42 successfully open dynamic library libcudart so 10 0 describe the current behavior run the test code below for gpu 0 the behavior be normal but for gpu 1 the output become zero describe the expect behavior the output of gpu 1 should be the same as gpu 0 code to reproduce the issue provide a reproducible test case that be the bare minimum necessary to generate the problem run python test py python import tensorflow as tf import numpy as np import os os environ cuda device order pci bus i d os environ cuda visible device 0 1 def linear name input output dim reuse false with tf variable scope name reuse reuse w tf get variable weight shape input get shape 1 output dim initializer tf orthogonal initializer b tf get variable bias output dim initializer tf constant initializer 0 x tf matmul input w b with tf device device cpu 0 x tf print x tf reduce sum tf abs w tf reduce sum tf abs b name weight bias return x def tower x reuse false rec tensor rec name x linear fc1 x 10 reuse rec tensor append x rec name append fc1 x linear fc2 x 1 reuse rec tensor append x rec name append fc2 return x rec tensor rec name x tf placeholder tf float32 none 3 x datum np random rand 5 3 feed dict x x datum ys rec xs rec name for I in range 2 with tf device tf devicespec device type gpu device index I y v n tower x I 0 ys append y rec xs append v rec name append n config tf configproto config gpu option allow growth true config allow soft placement false sess tf session config config sess run tf global variable initializer print check forward for I in range 2 print check gpu d I for j in range len rec xs i t sess run rec xs I j feed dict 0 l1norm np sum np abs t print s 5f rec name I j l1norm other info log include any log or source code that would be helpful to diagnose the problem if include traceback please include the full traceback large log and file should be attach my run log of test py warn log before flag parsing go to stderr w0402 21 57 00 129247 139765947877184 deprecation py 506 from home atlantix anaconda3 lib python3 7 site package tensorflow python ops variable scope py 883 call orthogonal init from tensorflow python op init op with dtype be deprecate and will be remove in a future version instruction for update call initializer instance with the dtype argument instead of pass it to the constructor w0402 21 57 00 150027 139765947877184 deprecation py 323 from test py 15 print from tensorflow python op log op be deprecate and will be remove after 2018 08 20 instruction for update use tf print instead of tf print note that tf print return a no output operator that directly print the output outside of defun or eager mode this operator will not be execute unless it be directly specify in session run or use as a control dependency for other operator this be only a concern in graph mode below be an example of how to ensure tf print execute in graph mode python sess tf session with sess as default tensor tf range 10 print op tf print tensor with tf control dependency print op out tf add tensor tensor sess run out additionally to use tf print in python 2 7 user must make sure to import the following from future import print function 2019 04 02 21 57 00 195153 I tensorflow stream executor platform default dso loader cc 42 successfully open dynamic library libcuda so 1 2019 04 02 21 57 00 626301 I tensorflow compiler xla service service cc 162 xla service 0x55bf73aa0640 execute computation on platform cuda device 2019 04 02 21 57 00 626354 I tensorflow compiler xla service service cc 169 streamexecutor device 0 geforce gtx titan x compute capability 5 2 2019 04 02 21 57 00 626368 I tensorflow compiler xla service service cc 169 streamexecutor device 1 geforce gtx titan x compute capability 5 2 2019 04 02 21 57 00 649925 I tensorflow core platform profile util cpu util cc 94 cpu frequency 2299805000 hz 2019 04 02 21 57 00 653093 I tensorflow compiler xla service service cc 162 xla service 0x55bf73a9fad0 execute computation on platform host device 2019 04 02 21 57 00 653157 I tensorflow compiler xla service service cc 169 streamexecutor device 0 2019 04 02 21 57 00 653809 I tensorflow core common runtime gpu gpu device cc 1467 find device 0 with property name geforce gtx titan x major 5 minor 2 memoryclockrate ghz 1 076 pcibusid 0000 08 00 0 totalmemory 11 93gib freememory 11 81gib 2019 04 02 21 57 00 654269 I tensorflow core common runtime gpu gpu device cc 1467 find device 1 with property name geforce gtx titan x major 5 minor 2 memoryclockrate ghz 1 076 pcibusid 0000 09 00 0 totalmemory 11 93gib freememory 11 81gib 2019 04 02 21 57 00 654961 I tensorflow core common runtime gpu gpu device cc 1546 add visible gpu device 0 1 2019 04 02 21 57 00 655060 I tensorflow stream executor platform default dso loader cc 42 successfully open dynamic library libcudart so 10 0 2019 04 02 21 57 01 477672 I tensorflow core common runtime gpu gpu device cc 1015 device interconnect streamexecutor with strength 1 edge matrix 2019 04 02 21 57 01 477727 I tensorflow core common runtime gpu gpu device cc 1021 0 1 2019 04 02 21 57 01 477757 I tensorflow core common runtime gpu gpu device cc 1034 0 n y 2019 04 02 21 57 01 477767 I tensorflow core common runtime gpu gpu device cc 1034 1 y n 2019 04 02 21 57 01 478354 I tensorflow core common runtime gpu gpu device cc 1149 create tensorflow device job localhost replica 0 task 0 device gpu 0 with 11422 mb memory physical gpu device 0 name geforce gtx titan x pci bus i d 0000 08 00 0 compute capability 5 2 2019 04 02 21 57 01 478791 I tensorflow core common runtime gpu gpu device cc 1149 create tensorflow device job localhost replica 0 task 0 device gpu 1 with 11422 mb memory physical gpu device 1 name geforce gtx titan x pci bus i d 0000 09 00 0 compute capability 5 2 2019 04 02 21 57 01 487682 I tensorflow stream executor platform default dso loader cc 42 successfully open dynamic library libcudart so 10 0 check forward check gpu 0 2019 04 02 21 57 01 527837 I tensorflow stream executor platform default dso loader cc 42 successfully open dynamic library libcubla so 10 0 fc1 weight bias 8 01475811 0 fc1 1 84453 fc1 weight bias 8 01475811 0 fc2 weight bias 2 51356649 0 fc2 0 19745 check gpu 1 fc1 weight bias 8 01475811 0 fc1 0 00000 fc1 weight bias 8 01475811 0 fc2 weight bias 2 51356649 0 fc2 0 00000 as for my gpu the driver version be 410 104 instal accord to official instruction hope it help could anybody help I asap I be almost crazy with tf s multiple gpu thank you very much |
tensorflowtensorflow | unable to use featurecolumn with keras functional api | Bug | system information have I write custom code as oppose to use a stock example script provide in tensorflow y os platform and distribution e g linux ubuntu 16 04 ubuntu 16 04 mobile device e g iphone 8 pixel 2 samsung galaxy if the issue happen on mobile device n a tensorflow instal from source or binary source tensorflow version use command below 2 0 0 alpha0 python version 3 7 bazel version if compile from source n a gcc compiler version if compile from source n a cuda cudnn version n a gpu model and memory n a describe the current behavior I follow the guide classify structured datum the author use tf feature column and tf datum alongside a sequential model which work out just fine then I ve try implement the same use keras functional api but be greet by the error message below attributeerror traceback most recent call last in 1 input layer input tensor feature layer name feature 2 x layer dense 128 activation relu input 3 x layer dense 64 activation relu x 4 5 baggage pre layer dense 1 activation sigmoid name baggage x local lib python3 7 site package tensorflow python keras engine input layer py in input shape batch size name dtype sparse tensor kwargs 231 dtype dtype 232 sparse sparse 233 input tensor tensor 234 return tensor include keras history 235 note that in this case train output and test output be the same pointer local lib python3 7 site package tensorflow python keras engine input layer py in init self input shape batch size dtype input tensor sparse name kwargs 77 dtype backend floatx 78 else 79 dtype backend dtype input tensor 80 elif input tensor be not none and input tensor dtype dtype 81 raise valueerror input tensor dtype differ from dtype s vs s local lib python3 7 site package tensorflow python keras backend py in dtype x 1025 1026 1027 return x dtype base dtype name 1028 1029 attributeerror str object have no attribute base dtype describe the expect behavior if I understand correctly tf keras be suppose to be interoperable with feature column and the way to achieve that be to wrap a list of feature column with tf keras layer densefeature and parse it to the input layer as a tensor like so feature layer tf keras layer densefeature feature column input layer input tensor feature layer name feature code to reproduce the issue here s the code to reproduce the error from future import absolute import division print function import numpy as np import panda as pd pip install tensorflow 2 0 0 alpha0 import tensorflow as tf from tensorflow import feature column from tensorflow import kera from tensorflow keras import layer from sklearn model selection import train test split url dataframe pd read csv url dataframe head train test train test split dataframe test size 0 2 train val train test split train test size 0 2 print len train train example print len val validation example print len test test example a utility method to create a tf datum dataset from a panda dataframe def df to dataset dataframe shuffle true batch size 32 dataframe dataframe copy label dataframe pop target ds tf datum dataset from tensor slice dict dataframe label if shuffle ds ds shuffle buffer size len dataframe ds ds batch batch size return ds batch size 5 a small batch sized be use for demonstration purpose train ds df to dataset train batch size batch size val ds df to dataset val shuffle false batch size batch size test ds df to dataset test shuffle false batch size batch size age feature column numeric column age age bucket feature column bucketize column age boundary 18 25 30 35 40 45 50 55 60 65 thal feature column categorical column with vocabulary list thal fix normal reversible thal one hot feature column indicator column thal thal embed feature column embed column thal dimension 8 thal hash feature column categorical column with hash bucket thal hash bucket size 1000 cross feature feature column cross column age bucket thal hash bucket size 1000 feature column numeric col for header in age trestbps chol thalach oldpeak slope can feature column append feature column numeric column header bucketize col age bucket feature column bucketize column age boundary 18 25 30 35 40 45 50 55 60 65 feature column append age bucket indicator col thal feature column categorical column with vocabulary list thal fix normal reversible thal one hot feature column indicator column thal feature column append thal one hot embed col thal embed feature column embed column thal dimension 8 feature column append thal embed cross col cross feature feature column cross column age bucket thal hash bucket size 1000 cross feature feature column indicator column cross feature feature column append cross feature batch size 32 train ds df to dataset train batch size batch size val ds df to dataset val shuffle false batch size batch size test ds df to dataset test shuffle false batch size batch size feature layer tf keras layer densefeature feature column input layer input tensor feature layer name feature x layer dense 128 activation relu input x layer dense 64 activation relu x baggage pre layer dense 1 activation sigmoid x model keras model inputs input output baggage pre model compile optimizer adam loss binary crossentropy metric accuracy other info log n a |
tensorflowtensorflow | invalid argument conv2dslowbackpropinput input and out backprop must have the same batch size | Bug | I want to use tensorflow to train my model firstly I use conv2d transpose in order to realize upsampling then I use batch normalization after each conv2d transpose layer and then I use conv2d and maxpool but when I train my model I meet a problem such like this invalid argument conv2dslowbackpropinput input and out backprop must have the same batch size input batch 64 outbackprop batch 56 batch dim 0 what make I more puzzled be the training can run some step my batch size 64 when the step arrive at 101 step the training stop here be my deconv function python def deconv2d input output shape k h 5 k w 5 d h 2 d w 2 stddev 0 02 name deconv2d with w false with tf variable scope name w tf get variable w k h k w output shape 1 input get shape 1 initializer tf random normal initializer stddev stddev deconv tf nn conv2d transpose input w output shape output shape stride 1 d h d w 1 bias tf get variable bias output shape 1 initializer tf constant initializer 0 0 deconv tf reshape tf nn bias add deconv bias deconv get shape if with w return deconv w bias else return deconv and this be how I use this deconv2d fucntion python h1 deconv2d h0 batch size s h8 s w8 f dim 4 name deconv1 h1 tf nn relu d bn 1 h1 and this be my batch normalization fuction python class batchnorm object def init self epsilon 1e 5 momentum 0 9 name batch norm with tf variable scope name self epsilon epsilon self momentum momentum self name name def call self x train true return tf contrib layers batch norm x decay self momentum update collection none epsilon self epsilon scale true be train train scope self name I have try some method such as use tf reshape right after the conv2d transpose but it didn t work I hope anyone else can help I resolve this problem thank you very much |
tensorflowtensorflow | read a tensor from file in python which be save use c | Bug | I m have trouble read interpret a tensor from a file use read file operation in python the tensor be save use save op in tensorflow cc api I can read the underlie byte encode tensor but find no method to convert that back to float original content of the tensor I have try decode raw in python but it complaint about of not be of proper multiple of 4 the above tensor comprise of a 3 channel nd array with float content would appreciate any help in decipher the content more precisely would really appreciate on some detail on how save op in tensorflow cc api work |
tensorflowtensorflow | transpose can be very slow on cpu | 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 ubuntu 18 04 wsl mobile device e g iphone 8 pixel 2 samsung galaxy if the issue happen on mobile device tensorflow instal from source or binary binary mkl pip install intel tensorflow tensorflow version use command below b v1 13 1 0 g6612da8 1 13 1 python version 3 6 8 anaconda describe the current behavior on cpu tf transpose can be 10 time slow than np tranpose copy depend on the dimension of the tensor in my example case the speed can be fix by sandwich the tf transpose in reshape that reduce the dimensionality for the transpose describe the expect behavior tf tranpose should be fast without the need to resort to reshape trick which will not always apply anyway note also that some transpose be implicit perform say by tf tensordot code to reproduce the issue python import time import numpy as np import tensorflow as tf tf enable v2 behavior d 2 n 22 note tf start get a lot slow than numpy at n 9 it seem try d 5 n 9 vs d 7 n 8 create a n dimensional ndarray psi np random randn d for n in range n astype np float64 psi np psi copy for in range 5 numpy s transpose be a metadata operation copy actually carry out the reordering of the datum in memory t0 time time psi np np transpose psi np list range 2 n 0 1 copy print time time t0 around 0 07 on my system psi tf tf convert to tensor psi tf contrib eager defun make this a graph just in case def f p return tf transpose p list range 2 n 0 1 for in range 5 t0 time time psi tf f psi tf print time time t0 around 0 5s on my system almost 10 time slow check result be the same print np linalg norm psi tf numpy ravel psi np ravel should print 0 0 note that the speed depend on the shape we now do an equivalent tranpose but sandwich by reshape to reduce the number of dim psi tf tf convert to tensor psi tf contrib eager defun def f2 p rephrase the high dimensional transpose as a matrix transpose pr tf reshape p d 2 d n 2 prt tf transpose pr 1 0 return tf reshape prt p shape for in range 5 t0 time time psi tf f2 psi tf print time time t0 around 0 008 on my system check result be the same print np linalg norm psi tf numpy ravel psi np ravel should print 0 0 other info log compare to 15697 I be actually time the equivalent op in numpy transpose then copy also this happen on mkl build as well as non mkl build of tf 1 13 1 |
tensorflowtensorflow | docs load really slow | Bug | please make sure that this be a documentation issue as per our github policy we only address code doc bug performance issue feature request and build installation issue on github tag doc template system information chrome internet speed 120 mbps tensorflow version all doc link all page it take a really long time to load the doc navigate to another page can take 10 second to load a new page |
tensorflowtensorflow | sample code be have syntax error in tensorflow nodejs code | Bug | system information tensorflow version doc link syntax error in tensorflow nodejs example see sample code for node js usage extra semicolon be cause the syntax error onepochend epoch log console log epoch epoch loss log loss in the above line semi colon in the end should be remove |
tensorflowtensorflow | dataset c extend documentation be outdate for tf 2 0 datasetv2 | Bug | system information tensorflow version 2 0 doc link describe the documentation issue the c code to extend dataset especially since datasetv2 be outdate here be a work version of file need in the documentation note the only part I be not really sure about be this one l76 basically let it as it be but I don t see the point there be also and external test run from python and bazel file not sure where or if I even can do a pull request to change the doc |
tensorflowtensorflow | break link | Bug | please make sure that this be a documentation issue as per our github policy we only address code doc bug performance issue feature request and build installation issue on github tag doc template system information tensorflow version api r1 13 doc link describe the documentation issue a 404 page not find error be display when click on the link in the paragraph below additional documentation can be find on the keras site we welcome contribution by user will you be able to update submit a pr use the doc style guide to fix the doc issue |
tensorflowtensorflow | tensorflow 2 0 batchnorm not work | Bug | system information I write custom code os platform and distribution ubuntu 18 04 tensorflow instal from pip tensorflow version 2 0 python version 3 6 cuda cudnn version 10 0 7 5 gpu model and memory gtx 1070 8 gig so I write a simple model use tf keras and I be use batchnorm it be work just fine then sudddenly it stop work log traceback most recent call last file home kislay document deepq dqmodel py line 30 in print dqn callback np random rand 4 210 160 1 random action 4 file home kislay document deepq dqmodel py line 18 in callback frame stack self batchnorm frame stack file home kislay local lib python3 6 site package tensorflow python keras engine base layer py line 660 in call output self call input args kwargs file home kislay local lib python3 6 site package tensorflow python keras layers normalization py line 589 in call output self fuse batch norm input training training file home kislay local lib python3 6 site package tensorflow python keras layers normalization py line 465 in fuse batch norm training fuse batch norm training fuse batch norm inference file home kislay local lib python3 6 site package tensorflow python keras util tf util py line 56 in smart cond pre true fn true fn false fn false fn name name file home kislay local lib python3 6 site package tensorflow python framework smart cond py line 56 in smart cond return false fn file home kislay local lib python3 6 site package tensorflow python keras layers normalization py line 462 in fuse batch norm inference datum format self datum format file home kislay local lib python3 6 site package tensorflow python op nn impl py line 1206 in fuse batch norm name name file home kislay local lib python3 6 site package tensorflow python ops gen nn op py line 3931 in fuse batch norm six raise from core status to exception e code message none file line 3 in raise from tensorflow python framework error impl internalerror could not find valid device for node node node fusedbatchnorm all kernel register for op fusedbatchnorm device xla cpu t in dt float device xla gpu t in dt float device xla cpu jit t in dt float device cpu t in dt float device gpu t in dt float device xla gpu jit |
tensorflowtensorflow | tf2 0 estimatorv2 use non exist export savedmodel method | Bug | dear tensorflower as per the title estimatorv2 s exporter py call a method that have be remove the fix be pretty easy but I m not sure about the implication of the change see more below system information have I write custom code as oppose to use a stock example script provide in tensorflow custom code but use estimator os platform and distribution e g linux ubuntu 16 04 4 19 20 1rodete1 amd64 1 smp debian 4 19 20 1rodete1 tensorflow instal from source or binary pip install upgrade tensorflow 2 0 0alpha0 tensorflow version use command below 2 0 0 alpha0 python version 3 7 1 describe the problem use tf estimator train and evaluate estimator train spec eval spec result in attributeerror estimatorv2 object have no attribute export savedmodel be throw 2 small modification to my local tensorflow estimator python estimator exporter py file fix the issue however before submit a pr I want to clarify that I have to 1 rename the method call to estimator export save model and this pose no problem 2 remove the strip default attrs self strip default attrs keyword argument l120 but I have no idea what this entail I assume that since you can t specify this argument anymore it default to the default policy in tf 1 x of strip the graphdef default attribute if that s the case be there any action require log file usr local google home msteiner pyenv version 3 7 1 lib python3 7 site package tensorflow estimator python estimator training py line 473 in train and evaluate return executor run file usr local google home msteiner pyenv version 3 7 1 lib python3 7 site package tensorflow estimator python estimator training py line 613 in run return self run local file usr local google home msteiner pyenv version 3 7 1 lib python3 7 site package tensorflow estimator python estimator training py line 714 in run local saving listener save listener file usr local google home msteiner pyenv version 3 7 1 lib python3 7 site package tensorflow estimator python estimator estimator py line 359 in train loss self train model input fn hook save listener file usr local google home msteiner pyenv version 3 7 1 lib python3 7 site package tensorflow estimator python estimator estimator py line 1139 in train model return self train model default input fn hook save listener file usr local google home msteiner pyenv version 3 7 1 lib python3 7 site package tensorflow estimator python estimator estimator py line 1173 in train model default save listener file usr local google home msteiner pyenv version 3 7 1 lib python3 7 site package tensorflow estimator python estimator estimator py line 1452 in train with estimator spec any step do true file usr local google home msteiner pyenv version 3 7 1 lib python3 7 site package tensorflow python training monitor session py line 856 in exit self close internal exception type file usr local google home msteiner pyenv version 3 7 1 lib python3 7 site package tensorflow python training monitor session py line 889 in close internal h end self coordinated creator tf sess file usr local google home msteiner pyenv version 3 7 1 lib python3 7 site package tensorflow python training basic session run hook py line 588 in end self save session last step file usr local google home msteiner pyenv version 3 7 1 lib python3 7 site package tensorflow python training basic session run hook py line 607 in save if l after save session step file usr local google home msteiner pyenv version 3 7 1 lib python3 7 site package tensorflow estimator python estimator training py line 519 in after save self evaluate global step value update self eval result file usr local google home msteiner pyenv version 3 7 1 lib python3 7 site package tensorflow estimator python estimator training py line 539 in evaluate self evaluator evaluate and export file usr local google home msteiner pyenv version 3 7 1 lib python3 7 site package tensorflow estimator python estimator training py line 927 in evaluate and export be the final export file usr local google home msteiner pyenv version 3 7 1 lib python3 7 site package tensorflow estimator python estimator training py line 960 in export eval result be the final export be the final export file usr local google home msteiner pyenv version 3 7 1 lib python3 7 site package tensorflow estimator python estimator exporter py line 420 in export be the final export file usr local google home msteiner pyenv version 3 7 1 lib python3 7 site package tensorflow estimator python estimator exporter py line 120 in export export result estimator export savedmodel attributeerror estimatorv2 object have no attribute export savedmodel code I m try to adapt the cmle template to tf 2 0 I avoid put the input fns and model fns since the model train and eval correctly the only problem be during the export python train input fn model input generate input fn file name pattern flag train file mode tf estimator modekeys train num epoch flag num epoch batch size flag train batch size eval input fn model input generate input fn file name pattern flag eval file mode tf estimator modekeys eval batch size flag eval batch size if metadata task type classification estimator model create classifier config run config elif metadata task type regression estimator model create regressor config run config else estimator model create estimator config run config train spec tf estimator trainspec train input fn max step int train step export for the final prediction graph exporter default tf estimator finalexporter estimator model input serve function flag export format as text false eval spec tf estimator evalspec eval input fn step flag eval step exporter exporter default this tell throttle sec flag eval every sec start delay secs 0 main train and evaluate loop tf estimator train and evaluate estimator train spec eval spec happy to provide any additional information |
tensorflowtensorflow | tpu training with keras api raise error in tensorflow 2 0 | Bug | system information have I write custom code as oppose to use a stock example script provide in tensorflow no os platform and distribution e g linux ubuntu 16 04 google colab mobile device e g iphone 8 pixel 2 samsung galaxy if the issue happen on mobile device no tensorflow instal from source or binary no tensorflow version use command below 2 0 0 dev20190330 python version 3 6 bazel version if compile from source gcc compiler version if compile from source cuda cudnn version gpu model and memory google colab colud tpu the current behavior by follow the official guide on tensorflow 2 0 distribute training the model fit method raise an error regard unregistered op type op type not register experimentalrebatchdataset in binary run on n 23c14aca w 0 the expect behavior the expect bahaviour be run code without any exception code to reproduce the issue the code in the google colab python pip install tf nightly 2 0 preview from future import absolute import division print function import tensorflow as tf tf compat v1 disable eager execution resolver tf distribute cluster resolver tpuclusterresolver tf tpu experimental initialize tpu system resolver tpu strategy tf distribute experimental tpustrategy resolver with tpu strategy scope model tf keras sequential tf keras layer dense 1 input shape 1 model compile loss mse optimizer sgd batch size per replica 10 batch size batch size per replica tpu strategy num replicas in sync dataset tf datum dataset from tensor 1 1 repeat 1000 batch batch size drop remainder true model fit dataset epoch 2 model evaluate dataset other info log the full exception warn log before flag parsing go to stderr w0331 08 24 17 732600 140554474739584 training util py 1268 expect a shuffle dataset but input dataset x be not shuffle please invoke shuffle on input dataset notfounderror traceback most recent call last usr local lib python3 6 dist package tensorflow python client session py in do call self fn args 1336 try 1337 return fn args 1338 except error operror as e usr local lib python3 6 dist package tensorflow python client session py in run fn feed dict fetch list target list option run metadata 1319 ensure any change to the graph be reflect in the runtime 1320 self extend graph 1321 return self call tf sessionrun usr local lib python3 6 dist package tensorflow python client session py in extend graph self 1354 with self graph session run lock pylint disable protect access 1355 tf session extendsession self session 1356 notfounderror op type not register experimentalrebatchdataset in binary run on n 23c14aca w 0 make sure the op and kernel be register in the binary run in this process note that if you be load a save graph which use op from tf contrib accessing e g tf contrib resampler should be do before import the graph as contrib op be lazily register when the module be first access during handling of the above exception another exception occur notfounderror traceback most recent call last in 2 batch size batch size per replica tpu strategy num replicas in sync 3 dataset tf datum dataset from tensor 1 1 repeat 1000 batch batch size drop remainder true 4 model fit dataset epoch 2 5 model evaluate dataset usr local lib python3 6 dist package tensorflow python keras engine training py in fit self x y batch size epoch verbose callback validation split validation datum shuffle class weight sample weight initial epoch step per epoch validation step validation freq max queue size worker use multiprocesse kwargs 745 step per epoch step per epoch 746 validation step validation step 747 validation freq validation freq 748 749 batch size self validate or infer batch size usr local lib python3 6 dist package tensorflow python keras engine training distribute py in fit distribute model x y batch size epoch verbose callback validation split validation datum shuffle class weight sample weight initial epoch step per epoch validation step validation freq 117 step per epoch step per epoch 118 validation step validation step 119 validation freq validation freq 120 else 121 return training array fit loop usr local lib python3 6 dist package tensorflow python keras engine training distribute py in experimental tpu fit loop model dataset epoch verbose callback initial epoch step per epoch val dataset validation step validation freq 310 todo fchollet add support for step per epoch none in tpu loop 311 current strategy model distribution strategy 312 iterator distribute training util get iterator dataset current strategy 313 step per epoch training util infer step for dataset 314 dataset step per epoch epoch step name step per epoch usr local lib python3 6 dist package tensorflow python keras engine distribute training util py in get iterator dataset distribution strategy 519 with distribution strategy scope 520 iterator distribution strategy make dataset iterator dataset 521 initialize iterator iterator distribution strategy 522 return iterator 523 usr local lib python3 6 dist package tensorflow python keras engine distribute training util py in initialize iterator iterator distribution strategy 527 init op control flow op group iterator initialize 528 if not context execute eagerly 529 k get session init op run init op 530 531 usr local lib python3 6 dist package tensorflow python client session py in run self fetch feed dict option run metadata 930 try 931 result self run none fetch feed dict option ptr 932 run metadata ptr 933 if run metadata 934 proto datum tf session tf getbuffer run metadata ptr usr local lib python3 6 dist package tensorflow python client session py in run self handle fetch feed dict option run metadata 1153 if final fetch or final target or handle and feed dict tensor 1154 result self do run handle final target final fetch 1155 feed dict tensor option run metadata 1156 else 1157 result usr local lib python3 6 dist package tensorflow python client session py in do run self handle target list fetch list feed dict option run metadata 1329 if handle be none 1330 return self do call run fn feed fetch target option 1331 run metadata 1332 else 1333 return self do call prun fn handle feed fetch usr local lib python3 6 dist package tensorflow python client session py in do call self fn args 1349 pass 1350 message error interpolation interpolate message self graph 1351 raise type e node def op message 1352 1353 def extend graph self notfounderror op type not register experimentalrebatchdataset in binary run on n 23c14aca w 0 make sure the op and kernel be register in the binary run in this process note that if you be load a save graph which use op from tf contrib accessing e g tf contrib resampler should be do before import the graph as contrib op be lazily register when the module be first access |
tensorflowtensorflow | add example of loading model that have be save with save keras model | Bug | please make sure that this be a documentation issue as per our github policy we only address code doc bug performance issue feature request and build installation issue on github tag doc template system information tensorflow version 1 13 doc link describe the documentation issue the above example provide no example of load the model after it s save I m used to use the old style of freeze optimize optimize for inference lib optimize for inference and save model with tf train write graph output graph def output dir model pb as text false which give I a graph model pb I run prediction on this graph locally on my mobile application with full tensorflow in unity so not use tf lite in my local prediction script I load the graph like this python def load graph model file graph tf graph graph def tf graphdef with open model file rb as f graph def parsefromstring f read with graph as default tf import graph def graph def return graph then grab the graph input and output op and run a prediction session this work pretty well however this new example use tf contrib save model save keras model to export the model which I would love to use because it export all kind of useful information which I d like to use in conjunction with tf serve for certain application and it look to be make it into tf 2 0 core with save keras model I end up with asset label csv save model pb variable when try to use my old graph loading script with this newly generate model with save keras model I get an error like this bash graph def parsefromstring f read google protobuf message decodeerror error parse message this have I puzzle as I expect this pb graph generate with the new api to work the same as the one I generate with tf train write graph browse through their content they look the same but the new api s model be small in file size which make I wonder if it s miss something I find an alternative method to load it as a graph anyways but I run into issue down the road with run it s op so I m not sure I m approach this correctly by load it as a graph or if there s a new keras way of load a model and run prediction maybe I m too far in the weed with try to load this model as a graph and run it s op I d expect there s a high level api to do this so I have a few question be the export save model pb frozen be it optimize if not I can convert the model I have to a graph then back to a model for something like save keras model ultimately can an example be add to this tutorial that demonstrate how to load a save model I write our initial codebase use tf 1 7 a little while back and I haven t change much since then but the api be change a lot and I want to take full advantage of the advancement also I understand if this issue should be move I assume this be appropriate since there be no documentation demonstrating load a save model with the new api thank you we welcome contribution by user will you be able to update submit a pr use the doc style guide to fix the doc issue yes |
tensorflowtensorflow | unable to convert frozen graph to usable model on iphone | Bug | please make sure that this be a bug as per our github policy we only address code doc bug performance issue feature request and build installation issue on github tag bug template system information have I write custom code as oppose to use a stock example script provide in tensorflow no os platform and distribution e g linux ubuntu 16 04 linux ubuntu 18 04 mobile device e g iphone 8 pixel 2 samsung galaxy if the issue happen on mobile device na tensorflow instal from source or binary binary tensorflow version use command below 1 13 1 python version 2 7 15rc1 bazel version if compile from source na gcc compiler version if compile from source na cuda cudnn version 10 0 gpu model and memory geforce gtx 1080 you can collect some of this information use our environment capture script you can also obtain the tensorflow version with python c import tensorflow as tf print tf git version tf version describe the current behavior I be fine tune the pretraine ssd mobilenetv1 model use the model research object detection sample config file for detect the bound box of object in an image python model main py logtostderr model dir training pipeline config path training ssd mobilenet v1 pet config after train I be freeze the model use python export inference graph py input type image tensor pipeline config path training ssd mobilenet v1 pet config output directory inference graph train checkpoint prefix training model ckpt 3740 after the frozen inference graph pb be be generate I d like to convert the model to a model format that be compatible with iphone I try the follow approach approach 1 convert to mlmodel python convert to mlmodel py warn root tensorflow version 1 13 1 detect last version know to be fully compatible be 1 12 0 load the tf graph 2019 03 30 18 39 48 956255 I tensorflow core platform cpu feature guard cc 141 your cpu support instruction that this tensorflow binary be not compile to use avx2 fma 2019 03 30 18 39 49 156054 I tensorflow compiler xla service service cc 150 xla service 0x5592d6f7e880 execute computation on platform cuda device 2019 03 30 18 39 49 156083 I tensorflow compiler xla service service cc 158 streamexecutor device 0 geforce gtx 1080 ti compute capability 6 1 2019 03 30 18 39 49 177932 I tensorflow core platform profile util cpu util cc 94 cpu frequency 3398155000 hz 2019 03 30 18 39 49 178509 I tensorflow compiler xla service service cc 150 xla service 0x5592d6fdf420 execute computation on platform host device 2019 03 30 18 39 49 178548 I tensorflow compiler xla service service cc 158 streamexecutor device 0 2019 03 30 18 39 49 178847 I tensorflow core common runtime gpu gpu device cc 1433 find device 0 with property name geforce gtx 1080 ti major 6 minor 1 memoryclockrate ghz 1 582 pcibusid 0000 01 00 0 totalmemory 10 92gib freememory 211 44mib 2019 03 30 18 39 49 178889 I tensorflow core common runtime gpu gpu device cc 1512 add visible gpu device 0 2019 03 30 18 39 49 179849 I tensorflow core common runtime gpu gpu device cc 984 device interconnect streamexecutor with strength 1 edge matrix 2019 03 30 18 39 49 179872 I tensorflow core common runtime gpu gpu device cc 990 0 2019 03 30 18 39 49 179882 I tensorflow core common runtime gpu gpu device cc 1003 0 n 2019 03 30 18 39 49 180081 I tensorflow core common runtime gpu gpu device cc 1115 create tensorflow device job localhost replica 0 task 0 device gpu 0 with 211 mb memory physical gpu device 0 name geforce gtx 1080 ti pci bus i d 0000 01 00 0 compute capability 6 1 graph load traceback most recent call last file convert to mlmodel py line 4 in output feature name softmax 0 file home deepvision local lib python2 7 site package tfcoreml tf coreml converter py line 586 in convert custom conversion function custom conversion function file home deepvision local lib python2 7 site package tfcoreml tf coreml converter py line 167 in convert pb to mlmodel op topological sort op op file home deepvision local lib python2 7 site package tfcoreml tf graph transform py line 194 in topological sort op push stack stack node in stack file home deepvision local lib python2 7 site package tfcoreml tf graph transform py line 38 in push stack raise valueerror graph have cycle valueerror graph have cycle the conversion code look like import tfcoreml as tf converter tf converter convert tf model path frozen inference graph pb mlmodel path my model mlmodel output feature name softmax 0 try the above with fast rcnn resnet101 and ssd mobilenet both give the same error not sure why the output feature name have to be a softmax as its a detection problem approach 2 convert to tflite tflite convert output file output tflite save model dir home deepvision code model research object detection inference graph save model 2019 03 30 19 22 30 354332 I tensorflow core platform cpu feature guard cc 141 your cpu support instruction that this tensorflow binary be not compile to use avx2 fma 2019 03 30 19 22 30 757830 I tensorflow compiler xla service service cc 150 xla service 0x5592ab466420 execute computation on platform cuda device 2019 03 30 19 22 30 757861 I tensorflow compiler xla service service cc 158 streamexecutor device 0 geforce gtx 1080 ti compute capability 6 1 2019 03 30 19 22 30 785890 I tensorflow core platform profile util cpu util cc 94 cpu frequency 3398155000 hz 2019 03 30 19 22 30 791145 I tensorflow compiler xla service service cc 150 xla service 0x5592ab498d20 execute computation on platform host device 2019 03 30 19 22 30 791161 I tensorflow compiler xla service service cc 158 streamexecutor device 0 2019 03 30 19 22 30 791279 I tensorflow core common runtime gpu gpu device cc 1433 find device 0 with property name geforce gtx 1080 ti major 6 minor 1 memoryclockrate ghz 1 582 pcibusid 0000 01 00 0 totalmemory 10 92gib freememory 69 38mib 2019 03 30 19 22 30 791294 I tensorflow core common runtime gpu gpu device cc 1512 add visible gpu device 0 2019 03 30 19 22 30 791745 I tensorflow core common runtime gpu gpu device cc 984 device interconnect streamexecutor with strength 1 edge matrix 2019 03 30 19 22 30 791755 I tensorflow core common runtime gpu gpu device cc 990 0 2019 03 30 19 22 30 791762 I tensorflow core common runtime gpu gpu device cc 1003 0 n 2019 03 30 19 22 30 791844 I tensorflow core common runtime gpu gpu device cc 1115 create tensorflow device job localhost replica 0 task 0 device gpu 0 with 69 mb memory physical gpu device 0 name geforce gtx 1080 ti pci bus i d 0000 01 00 0 compute capability 6 1 warning tensorflow from home deepvision local lib python2 7 site package tensorflow lite python convert save model py 61 load from tensorflow python save model loader impl be deprecate and will be remove in a future version instruction for update this function will only be available through the v1 compatibility library as tf compat v1 save model loader load or tf compat v1 save model load there will be a new function for import savedmodel in tensorflow 2 0 2019 03 30 19 22 31 298973 I tensorflow core common runtime gpu gpu device cc 1512 add visible gpu device 0 2019 03 30 19 22 31 299020 I tensorflow core common runtime gpu gpu device cc 984 device interconnect streamexecutor with strength 1 edge matrix 2019 03 30 19 22 31 299032 I tensorflow core common runtime gpu gpu device cc 990 0 2019 03 30 19 22 31 299040 I tensorflow core common runtime gpu gpu device cc 1003 0 n 2019 03 30 19 22 31 299186 I tensorflow core common runtime gpu gpu device cc 1115 create tensorflow device job localhost replica 0 task 0 device gpu 0 with 69 mb memory physical gpu device 0 name geforce gtx 1080 ti pci bus i d 0000 01 00 0 compute capability 6 1 warning tensorflow from home deepvision local lib python2 7 site package tensorflow lite python convert save model py 275 convert variable to constant from tensorflow python framework graph util impl be deprecate and will be remove in a future version instruction for update use tf compat v1 graph util convert variable to constant warn tensorflow from home deepvision local lib python2 7 site package tensorflow python framework graph util impl py 245 extract sub graph from tensorflow python framework graph util impl be deprecate and will be remove in a future version instruction for update use tf compat v1 graph util extract sub graph traceback most recent call last file home deepvision local bin tflite convert line 11 in sys exit main file home deepvision local lib python2 7 site package tensorflow lite python tflite convert py line 442 in main app run main run main argv sys argv 1 file home deepvision local lib python2 7 site package tensorflow python platform app py line 125 in run sys exit main argv file home deepvision local lib python2 7 site package tensorflow lite python tflite convert py line 438 in run main convert model tflite flag file home deepvision local lib python2 7 site package tensorflow lite python tflite convert py line 191 in convert model output datum converter convert file home deepvision local lib python2 7 site package tensorflow lite python lite py line 411 in convert invalid shape 1 format tensor name tensor shape list valueerror none be only support in the 1st dimension tensor image tensor have invalid shape none none none 3 also try the follow tflite convert output file output tflite graph def file frozen inference graph pb input array input output array but I be not sure what the last layer be as the model be fairly convoluted try visualize with netron but describe the expect behavior seamless conversion to tflite and mlmodel file code to reproduce the issue provide a reproducible test case that be the bare minimum necessary to generate the problem na other info log include any log or source code that would be helpful to diagnose the problem if include traceback please include the full traceback large log and file should be attach na |
Subsets and Splits
No community queries yet
The top public SQL queries from the community will appear here once available.