questions about the reproduction of M3DM, thanks!

[ZHE-SAPI]

Dear author, I have met a problem as shown in nomewang/M3DM#30, I noticed that in your paper, you sucessfully reproduced the M3DM on eyecan dataset, could you pls point out the problem mentiond in nomewang/M3DM#30, thanks very much!!!

[alex-costanzino]

Hey! We were able to reproduce the result on both MVTec 3D-AD and Eyecandies by simply deactivating the UFF module, either by using the provided checkpoints from the authors or by re-training from scratch (without UFF module as well).

Hope this helps :)

[ZHE-SAPI]

thanks a lot!

…

---- Replied Message ---- | From | Alex ***@***.***> | | Date | 06/12/2025 14:46 | | To | CVLAB-Unibo/crossmodal-feature-mapping ***@***.***> | | Cc | zhang ***@***.***>, Author ***@***.***> | | Subject | Re: [CVLAB-Unibo/crossmodal-feature-mapping] questions about the reproduction of M3DM, thanks! (Issue #10) | alex-costanzino left a comment (CVLAB-Unibo/crossmodal-feature-mapping#10) Hey! We were able to reproduce the result on both MVTec 3D-AD and Eyecandies by simply deactivating the UFF module, either by using the provided checkpoints from the authors or by re-training from scratch (without UFF module as well). Hope this helps :) — Reply to this email directly, view it on GitHub, or unsubscribe. You are receiving this because you authored the thread.Message ID: ***@***.***>