Closed huguesb closed 6 years ago
It seem I was affected by tensorflow/tensorflow#12491
After applying the solution mentioned there, namely pip uninstall enum && apt-get python-enum34
, I am down to 14 failing tests, which again seem to have a common root cause:
cat /root/.cache/bazel/_bazel_root/3b4c7ccb85580bc382ce4a52e9580003/execroot/__main__/bazel-out/local-opt/testlogs/dragnn/tools/model_trainer_test/test.log
exec ${PAGER:-/usr/bin/less} "$0" || exit 1
-----------------------------------------------------------------------------
2017-10-25 23:11:11.840113: F external/org_tensorflow/tensorflow/core/framework/variant_op_registry.cc:104] Check failed: existing == nullptr (0x26352c8 vs. nullptr)Unary VariantDecodeFn for type_name: tensorflow::Tensor already registered
This looks like a dup of #2355 which seems to have a hacky workaround. Any chance of getting an actual fix?
Same here, issue #2677
Confirmed dup of issue #2355.
System information
models/research/syntaxnet
90267567
(which is the commit that the remote points to inside this repo at commit538f89c4
)bazel test ...
(after following the instructions at https://github.com/tensorflow/models/tree/master/research/syntaxnet )Describe the problem
22 tests are failing, specifically:
Most of the test errors seem to have the same root cause: