hparams: refactor to isolate TF dependencies #2163
Merged
Conversation
This file contains bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
|
It is much easier to review these commits individually. |
nfelt
approved these changes
Apr 29, 2019
wchargin
changed the base branch from
wchargin-hparams-experiment-writing
to
master
Summary: Fixes an unintentional omission from #2130: `api.py` (not just its test) actually does currently depend on TensorFlow. Part of a reorganization to match the structure described here: <#2139 (comment)> wchargin-branch: hparams-api-refactor
Summary: We’ll shortly reinstate `api` as a wrapper over `summary_v2` (in a separate commit to help Git track the moves). Part of a reorganization to match the structure described here: <#2139 (comment)> wchargin-branch: hparams-api-refactor
Summary: We’ll shortly extract the Keras callback out of `summary_v2` such that `summary_v2` does not need to depend on TensorFlow. It will continue to be exposed from `api`. Part of a reorganization to match the structure described here: <#2139 (comment)> wchargin-branch: hparams-api-refactor
Summary: The `summary_v2` module no longer has any fundamental dependency on TensorFlow. Its implementation still depends on `summary`, which does directly reference the TensorFlow proto copies. Part of a reorganization to match the structure described here: <#2139 (comment)> wchargin-branch: hparams-api-refactor
wchargin
force-pushed
the
wchargin-hparams-api-refactor
branch
from
4d46798
to
efe8788
Compare
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
Add this suggestion to a batch that can be applied as a single commit.
This suggestion is invalid because no changes were made to the code.
Suggestions cannot be applied while the pull request is closed.
Suggestions cannot be applied while viewing a subset of changes.
Only one suggestion per line can be applied in a batch.
Add this suggestion to a batch that can be applied as a single commit.
Applying suggestions on deleted lines is not supported.
You must change the existing code in this line in order to create a valid suggestion.
Outdated suggestions cannot be applied.
This suggestion has been applied or marked resolved.
Suggestions cannot be applied from pending reviews.
Suggestions cannot be applied on multi-line comments.
Suggestions cannot be applied while the pull request is queued to merge.
Suggestion cannot be applied right now. Please check back later.
Summary:
This sequence of commits turns the current
apimodule into a wrapperaround two modules,
callback(which depends on TensorFlow and onlyprovides the Keras callback) and
summary_v2(which does not, andprovides everything else). Thus, it is now possible in principle to
write hparams summaries without a TensorFlow dependency, though in
practice
summary_v2is currently implemented in terms ofsummary,which depends on the TensorFlow copy of the
Summaryproto.This is an internal refactor; the demo code is unchanged.
Reorganization to match the structure described here:
#2139 (comment)
Test Plan:
Existing unit tests retained, up to refactoring. The demo still works.
wchargin-branch: hparams-api-refactor