Add Axon.stop_grad/2 - #677
Open
georgeguimaraes wants to merge 1 commit into
Open
georgeguimaraes wants to merge 1 commit into
georgeguimaraes wants to merge 1 commit into
Conversation
georgeguimaraes
force-pushed
the
stop-grad-layer
branch
2 times, most recently
from
September 17, 2026 14:59
0dc74c6 to
74df1ca
Compare
A layer that passes its input through unchanged and stops gradients from flowing back through it, for sharing a component between two paths while training it through one of them, or freezing a branch.
georgeguimaraes
force-pushed
the
stop-grad-layer
branch
from
September 17, 2026 14:59
74df1ca to
05592b4
Compare
This file contains hidden or 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
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
1 participant
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Outdated suggestions cannot be applied.This suggestion has been applied or marked resolved.Suggestions cannot be applied from pending reviews.Suggestions cannot be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.
A layer that passes its input through and stops gradients from flowing back through it, the graph-level counterpart of
Nx.Defn.Kernel.stop_grad/1. Useful when a component feeds two paths and should learn from one of them only: soothsayer's trend also scales its multiplicative seasonality and we detach it there withAxon.nx(&Nx.Defn.Kernel.stop_grad/1), which this replaces with a named layer.