New issue
Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.
By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.
Already on GitHub? Sign in to your account
Make checkpoint name optional so that user can save to h5 format. #3411
Conversation
Hmmm, it looks like unit test failed due to this change:
|
Unit Test Results (with flaky tests) 920 files - 55 920 suites - 55 9h 34m 44s ⏱️ - 30m 19s For more details on these failures, see this check. Results for commit b0cc0fa. ± Comparison against base commit 71e10b4. ♻️ This comment has been updated with latest results. |
Seems the docker do not have h5py installed, which makes the keras.load_model think it is a path. |
Seems h5py is installed by default, base on the if condition |
Signed-off-by: Peng Zhang <pengz@uber.com>
Signed-off-by: Peng Zhang <pengz@uber.com>
7f6f8ac
to
b0cc0fa
Compare
Signed-off-by: Peng Zhang pengz@uber.com
making the checkpoint file name optional to enable user to save model checkpoint in
.h5
format.