Request for Softmax-Only Zip File

[colin19950703]

Hello,

Thank you for sharing your remarkable research with us.

I am looking to conduct further experiments based on your research, and I require probability values obtained solely through the application of softmax.

Is there a separate zip file available that contains results obtained using only softmax, without hard targets or soft targets?

Thank you.

[szq0214]

Hi @colin19950703, thanks for your interest in our work. Are you referring to the full softmax labels with a 1000-dimension? I'm uploading it and will let you know once completed. Kindly be aware that the file size is approximately 2.3T.

[colin19950703]

Thank you for your kindness. I’ll be waiting for your contact.

…

-----Original Message----- From: Zhiqiang Shen ***@***.***> To: szq0214/FKD ***@***.***> Cc: Jaeung Lee ***@***.***>, Mention ***@***.***> Sent: 2023-10-28 (토) 17:39:26 (GMT+09:00) Subject: Re: [szq0214/FKD] Request for Softmax-Only Zip File (Issue #7) Hi @colin19950703, thanks for your interest in our work. Are you referring to the full softmax labels with a 1000-dimension? I'm uploading it and will let you know once completed. Kindly be aware that the file size is approximately 2.3T. — Reply to this email directly, view it on GitHub, or unsubscribe. You are receiving this because you were mentioned.Message ID: ***@***.***>

[szq0214]

Hi @colin19950703, please find the full soft labels at: https://mbzuaiac-my.sharepoint.com/:f:/g/personal/zhiqiang_shen_mbzuai_ac_ae/Em7k11ojvnFDmcG5HQB9qrwBzsKH71_6auBnbx3MoTEw5w?e=lntPzT.

Also please use the following command to unzip: cat efft_full_soft_label.tar.gz* | tar -zxv