Skip to content

Conversation

karmel
Copy link
Contributor

@karmel karmel commented Mar 29, 2018

No description provided.

@karmel karmel requested review from nealwu and robieta March 29, 2018 23:33
@karmel karmel requested a review from k-w-w as a code owner March 29, 2018 23:33
@karmel
Copy link
Contributor Author

karmel commented Mar 30, 2018

The files here seem to be giving the wrong accuracy; closing this while I sort out what happened.

@karmel karmel closed this Mar 30, 2018
@karmel karmel reopened this Mar 31, 2018
@karmel
Copy link
Contributor Author

karmel commented Mar 31, 2018

Reopening, as the checkpoints seem to work with master + tf-nightly.

@karmel
Copy link
Contributor Author

karmel commented Apr 2, 2018

@robieta , in the interest of minimizing the time during which we are wrong, how do you feel about merging this, and then updating with the checkpoints/savedmodels from #3687 when they are ready and that is in? You can ditch the Frozen Graphs at that point.

cc @qlzh727 - it would be nice if the benchmarks stuck their checkpoints and savedmodels somewhere were we could just link to a directory, and tell people to take the most recent ones, rather than manually updating.

@robieta
Copy link
Contributor

robieta commented Apr 2, 2018

Agreed. That seems like the most sensible approach.

@karmel karmel merged commit 59c7683 into master Apr 2, 2018
@karmel karmel deleted the fix/checkpoints branch April 2, 2018 21:44
omegafragger pushed a commit to omegafragger/models that referenced this pull request May 15, 2018
* Adding new checkpoints

* Adding links to other files as well
Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment
Labels
Projects
None yet
Development

Successfully merging this pull request may close these issues.

3 participants