Test: Replaced raw Mooncake with Flux implementation and checked Mooncake compatibility for more layers#642
Merged
CarloLucibello merged 3 commits intoJuliaGraphs:masterfrom Mar 15, 2026
Conversation
…cake compatibility for more layers Signed-off-by: Parvm1102 <parvmittal31757@gmail.com>
Parvm1102
force-pushed
the
mooncake-ad-testing
branch
from
2d8ad69 to
b5130f0
Compare
Contributor
Author
|
@CarloLucibello could you please review this? |
Contributor
Author
|
Interesting observation: |
Contributor
Author
|
@CarloLucibello Please review this when you have time |
Member
It would be good to have a MWE for this and report it upstream |
Parvm1102
commented
Mar 15, 2026
| g_mc_result = result.grad # Extract gradients tuple | ||
| y_mc, g_mc = Flux.withgradient(f -> loss(f, graph, xs...), Flux.AutoMooncake(), f) | ||
| @assert isapprox(y, y_mc; rtol, atol) | ||
| check_equal_leaves(g, g_mc_result; rtol, atol) |
Contributor
Author
There was a problem hiding this comment.
Suggested change
| check_equal_leaves(g, g_mc_result; rtol, atol) | |
| check_equal_leaves(g, g_mc; rtol, atol) |
@CarloLucibello The g_mc_result should also be changed to g_mc, otherwise the tests are failing. Should I open a new pr fixing the mistake?
Member
There was a problem hiding this comment.
Ops, sorry, yes please fix
Parvm1102
added a commit
to Parvm1102/GraphNeuralNetworks.jl
that referenced
this pull request
Mar 15, 2026
Signed-off-by: Parvm1102 <parvmittal31757@gmail.com>
Merged
CarloLucibello
pushed a commit
that referenced
this pull request
Mar 16, 2026
Contributor
Author
Yeah I wil do it soon! |
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
2 participants
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Outdated suggestions cannot be applied.This suggestion has been applied or marked resolved.Suggestions cannot be applied from pending reviews.Suggestions cannot be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.
I had reported the friendly_tangents= true issue to Mooncake.jl and they have resolved it in Mooncake v0.5.17.
I have changed the implementation with the preferred Flux.withgradient().
Also I have also checked few more layers in basic.jl, pool.jl and temporal.jl.
Observations: