Skip to content

Test: Replaced raw Mooncake with Flux implementation and checked Mooncake compatibility for more layers#642

Open
Parvm1102 wants to merge 1 commit intoJuliaGraphs:masterfrom
Parvm1102:mooncake-ad-testing
Open

Test: Replaced raw Mooncake with Flux implementation and checked Mooncake compatibility for more layers#642
Parvm1102 wants to merge 1 commit intoJuliaGraphs:masterfrom
Parvm1102:mooncake-ad-testing

Conversation

@Parvm1102
Copy link
Contributor

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:

  • GlobalPool: PASS
  • GlobalAttentionPool: PASS
  • TGCNCell / TGCN: PASS
  • EvolveGCNO: PASS
  • GNNChain: Fail (MooncakeRuleCompilationError, complex chain with anonymous fns, Dropout, BatchNorm)
  • GConvLSTMCell / GConvLSTM: Fail (Uses ChebConv internally)
  • GConvGRUCell / GConvGRU: Fail (Uses ChebConv internally)
  • DCGRUCell / DCGRU: Fail (Uses DConv internally)
  • EvolveGCNOCell: Fail ("with state" call fails during Mooncake rule compilation)

…cake compatibility for more layers

Signed-off-by: Parvm1102 <parvmittal31757@gmail.com>
@Parvm1102 Parvm1102 force-pushed the mooncake-ad-testing branch from 2d8ad69 to b5130f0 Compare March 7, 2026 17:37
@Parvm1102
Copy link
Contributor Author

Parvm1102 commented Mar 7, 2026

@CarloLucibello could you please review this?

@Parvm1102
Copy link
Contributor Author

Interesting observation: TransformerConv and GATv2Conv were failing in the CI earlier, but now they pass with mooncake enabled.

Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment

Labels

None yet

Projects

None yet

Development

Successfully merging this pull request may close these issues.

1 participant