Repository navigation
check compatibility with Mooncake on CPU #640
Description
Activity
Hi @CarloLucibello ,
I have done some progress on this: I've set up the test suite for the spatial convolutional layers comparing Mooncake and Zygote.
Results: 16 / 20 conv layers are compatible (935 passed tests).
However, there are errors with these two layers:- DConv (3 errors): Fails during reverse-mode differentiation (might be related to Wrong implementation DConv #592).
- ChebConv (1 error): Crashes during gradient cache preparation before the test executes.
2 layers were excluded: CGConv and GMMConv, same as existing tests.
How would you like me to proceed? Should I dig into finding workarounds for DConv and ChebConv, or should I move on to testing the other layer categories (pool.jl, heteroconv.jl, etc.)? I can also open a PR if you want to see the code.
Hi @Parvm1102, thanks.
I suggest you implement something similar to the test infrastructure of Flux (see thetest_gradientsfunction)
https://github.com/FluxML/Flux.jl/blob/master/test/test_utils.jl#L82Then open a PR with just the conv layers tested and we will discuss there.
@CarloLucibello Hey, Can I work on this issue I have theoretical knowledge of Neural Networks and a little hands on experience as well, we were taught Neural Networks and fuzzy logic as a separate course in our college. So should I work on this or is there any issue that will be more suitable for me. I would really appreciate your suggestions.
Hi, I'm willing to have a look. I got a few basis in Julia and got a good understanding of Neural Networks, as well as Graphs/Networks more broadly on the theoretical side.
@Parvm1102 this issue should be updated with the current status
- changed the title
[-]check compatibility with Mooncake[/-][+]check compatibility with Mooncake on CPU[/+]on Jul 31, 2026 let's keep track of CUDA compatibility in #702 and leave this issue for CPU
- I've taken good note of that.…On Fri, Jul 31, 2026 at 2:12 AM Carlo Lucibello ***@***.***> wrote: *CarloLucibello* left a comment (JuliaGraphs/GraphNeuralNetworks.jl#640) <#640 (comment)> let's keep track of CUDA compatibility in #702 <#702> and leave this issue for CPU — Reply to this email directly, view it on GitHub <#640?email_source=notifications&email_token=AIQFCPNCHHVKPVXX2GVKXM35HQ2GLA5CNFSNUABFM5UWIORPF5TWS5BNNB2WEL2JONZXKZKDN5WW2ZLOOQXTKMJTHE4TAMZVHE32M4TFMFZW63VHMNXW23LFNZ2KKZLWMVXHJLDGN5XXIZLSL5RWY2LDNM#issuecomment-5139903597>, or unsubscribe <https://github.com/notifications/unsubscribe-auth/AIQFCPMQK4UNKHRGJKGG74T5HQ2GLAVCNFSNUABFKJSXA33TNF2G64TZHM2DAMJSGI3TQOJYHNEXG43VMU5TGOJXGE3DGNBXGE2KC5QC> . You are receiving this because you commented.Message ID: ***@***.***>
Check that all our layers and operators are compatible with Mooncake.jl AD