tensorflow-aarch64
tensorflow-aarch64 copied to clipboard
KumaTea
build tf v2.5.0 falied
According to your method, I compiled tensorflow_aarch64 on tf2.5, but it reported an error:
ERROR: /root/tensorflow/tensorflow/python/BUILD:2531:29: Linking of rule '//tensorflow/python:gen_candidate_sampling_ops_py_wrappers_cc' failed (Exit 1): gcc failed: error executing command /usr/bin/gcc @bazel-out/host/bin/tensorflow/python/gen_candidate_sampling_ops_py_wrappers_cc-2.params bazel-out/host/bin/_solib_aarch64/_U_S_Stensorflow_Spython_Cgen_Ucandidate_Usampling_Uops_Upy_Uwrappers_Ucc___Utensorflow/libtensorflow_framework.so.2: error: undefined reference to 'std::allocator<absl::lts_2020_09_23::string_view>::allocator()' collect2: error: ld returned 1 exit status Target //tensorflow/tools/pip_package:build_pip_package failed to build
build command
BAZEL_LINKLIBS=-l%:libstdc++.a bazel build --config=noaws --config=nogcp --config=nohdfs --config=nonccl --jobs=64 //tensorflow/tools/pip_package:build_pip_package
But I built whl success at tf v2.4.1， And There is tf2.5.x version in the whl package you provided, is this a bug of 2.5.0?
I build v2.4.1 success , but build v2.5.0/v2.5.1/v2.5.2/v2.5.3 failed, I feel so confused =.=
Hi!
I also failed to build v2.5 about a year ago, and I think it might be a bug of this version...
But I saw that you have provided the whl of v2.5.x, and I am curious about how this is generated. After downloading, I can pip install it on my personal arm machine
you have provided the whl of v2.5.x
That's just a placeholder...
A year ago linaro did release 2.5.x wheels but they are now removed from their server and I didn't backup them. I also have no idea how they did that, since there are no instructions shared by them.
But I just found a repo PINTO0309/Tensorflow-bin, which listed building commands for every version. You may give it a try.