Allow ForwardDiff in BatchNorm's track_stats #2127
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
mcabbott
commented
Nov 27, 2022
| @@ -275,8 +275,8 @@ function _track_stats!( | |||
| μnew = vec(N ∈ reduce_dims ? μ : mean(μ, dims=N)) | |||
| σ²new = vec(N ∈ reduce_dims ? σ² : mean(σ², dims=N)) | |||
|
|
|||
| bn.μ = res_mtm .* bn.μ .+ mtm .* μnew | |||
| bn.σ² = res_mtm .* bn.σ² .+ mtm .* (m / (m - one(V))) .* σ²new | |||
| bn.μ .= value.(res_mtm .* bn.μ .+ mtm .* μnew) | |||
There was a problem hiding this comment.
Also updates bn.μ in-place, rather than replacing the field. We know it's mutable as the constructor made it so.
There was a problem hiding this comment.
can we leave a comment explaining why value is there?
mcabbott
force-pushed
the
fwd_track_stats
branch
from
2a06bcb
to
2dca7ec
Compare
|
Bump? |
ToucheSir
reviewed
Dec 8, 2022
ToucheSir
approved these changes
Dec 8, 2022
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Outdated suggestions cannot be applied.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.
Fixes... well, fixes the error seen in #2122 in the narrowest way, by explicitly removing duals. Does not cause ForwardDiff to put the layer into training mode.
Has a test, which ought to survive even if some more elaborate solution is found.