ArcFace baseline model used in SlerpFace paper

[donghan22]

Hi,

Thanks for your great work.

As stated in paper:
""We compare SlerpFace with an unprotected baseline and five FTP methods: ArcFace as the baseline; ""

Could you guide me where can I find this pre-trained model of ArcFace model that you used as baseline model in evaluation? @zzzweakman

[zzzweakman]

Dear @DongHan9722,

Thank you for your interest in our work. I have uploaded the ArcFace pretrain model to the following link. You can access the model weights and training logs there. I hope you find this helpful.

Best regards,
The SlerpFace Team

[donghan22]

Hi, Thanks for your quick replying. May I ask what FR model you were using to test other FTP methods? Sent from Outlook for iOS<https://aka.ms/o0ukef>

…

________________________________ From: Zhizhou Zhong ***@***.***> Sent: Monday, September 1, 2025 7:33:43 PM To: Tencent/TFace ***@***.***> Cc: Clyde ***@***.***>; Mention ***@***.***> Subject: Re: [Tencent/TFace] ArcFace baseline model used in SlerpFace paper (Issue #157) [https://avatars.githubusercontent.com/u/62113708?s=20&v=4]zzzweakman left a comment (Tencent/TFace#157)<#157 (comment)> Dear @DongHan9722<https://github.com/DongHan9722>, Thank you for your interest in our work. I have uploaded the ArcFace pretrain model to the following link<https://drive.google.com/drive/folders/1h-yO2h3dyc8DIIZAA9rdgwZ8qcPjDy7P>. You can access the model weights and training logs there. I hope you find this helpful. Best regards, The SlerpFace Team — Reply to this email directly, view it on GitHub<#157 (comment)>, or unsubscribe<https://github.com/notifications/unsubscribe-auth/ATETRAOGC7II3MLVJEMP7DL3QR7PPAVCNFSM6AAAAACFKG2K3CVHI2DSMVQWIX3LMV43OSLTON2WKQ3PNVWWK3TUHMZTENBSHE4DMMBVGQ>. You are receiving this because you were mentioned.Message ID: ***@***.***>

[zzzweakman]

Hi,
Thanks for your quick replying.
May I ask what FR model you were using to test other FTP methods?
Sent from Outlook for iOShttps://aka.ms/o0ukef
…

Thank you for your message. I used the model trained with the ArcFace loss on the MS1M dataset. This is the same model that I have just uploaded. By the way, you should use TFace to run inference to ensure the SlerpFace and baseline models can be loaded correctly.

[donghan22]

Hi,
Thanks for your quick replying.
May I ask what FR model you were using to test other FTP methods?
Sent from Outlook for iOShttps://aka.ms/o0ukef
…

Thank you for your message. I used the model trained with the ArcFace loss on the MS1M dataset. This is the same model that I have just uploaded. By the way, you should use TFace to run inference to ensure the SlerpFace and baseline models can be loaded correctly.

Hi, thanks for replying.
In terms of PolyProtect, may I ask what value of overlap, m, and range of C,E that you were used?