You signed in with another tab or window. Reload to refresh your session.You signed out in another tab or window. Reload to refresh your session.You switched accounts on another tab or window. Reload to refresh your session.Dismiss alert
{{ message }}
This repository was archived by the owner on Jul 7, 2023. It is now read-only.
Bump T2T version to 1.15.5
With this change Travis will turn green.
When we import compat.v1, we shouldn't do tf.compat.v1 anymore.
Also the following to unclog Travis:
six>=1.12.0 is required by tf-hub 0.8.0, one of the core dependencies seems to
be installing tf-hub, so pinning tf-hub to 0.7.0 (which is in the extras
section) doesn't seem to help.
Locally the previous setup.py installs fine, but only errors on Travis for some
reason.
PiperOrigin-RevId: 307203166
Changes needed for T2T's problems to be imported on TF 2.0 -- this do…
…esn't mean we are using these problems.
This is only so that we can `import trax` on these environments.
PiperOrigin-RevId: 282108974
Move from tf flags to absl flags, since they seem to be gone in tf 1.15
NOTE: This is only for libraries (and not binaries) in data_generators/
Also bump up version to 1.15.1
Also add a description to show on PyPI
PiperOrigin-RevId: 282056584
pip installing tensorflow serving seems to be uninstalling the installed
version of tensorflow (1.15.0) and replacing with 2.0, leading to errors of
missing contrib -- so install a pinned version.
Make an error in all_problems.py log earlier.
Reverting the earlier change about tf flags not being shipped with TF, because of this TF 2.0 subterfuge.
PiperOrigin-RevId: 281893644
Execute `jupyter nbconvert` on python3 instead of python2.
python2 fails with:
`RuntimeError: Kernel died before replying to kernel_info`
python3 seems to pass.
PiperOrigin-RevId: 272718077