Skip to content
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

SameDiff: Add graph transformation / transfer learning functionality #7199

Closed
AlexDBlack opened this issue Feb 20, 2019 · 1 comment

Comments

Projects
None yet
1 participant
@AlexDBlack
Copy link
Member

commented Feb 20, 2019

Graph transformation functionality:

  • Removing ops
  • Replacing ops
  • Renaming variables
  • Reinitializing parameters
  • Changing connection structure

We already have the ability to convert variables to constants, and back.

There are 2 use cases that having graph transformation functionality would be useful:

  1. For transfer learning ("replace output layer(s)") etc
  2. For import use cases

With regard to import: I've recently run into a model that I can't clean up in TF: I either run into model freezing bugs (in TF 1.12, maybe fixed in 1.13) or end up with training-specific nodes still in the graph that make it useless (no placeholders, but IteratorV2 and similar ops instead)

Aha! Link: https://skymindai.aha.io/features/ND4J-101

@lock

This comment has been minimized.

Copy link

commented Apr 14, 2019

This thread has been automatically locked since there has not been any recent activity after it was closed. Please open a new issue for related bugs.

@lock lock bot locked and limited conversation to collaborators Apr 14, 2019

Sign up for free to subscribe to this conversation on GitHub. Already have an account? Sign in.
You can’t perform that action at this time.