About the [Iteration: 29/ 30][Train Loss: 279.2204]

[jingzhang697]

Hello,
Recently, I did an experiment on remote sensing, but I didn't modify any parameters during running this code, but the loss is very large and the map=0.0642. I am very confused about this problem, and I hope to get your answer.

article：“Asymmetric Hash Code Learning for Remote Sensing Image Retrieval”

[weiweisong415]

Hi,   Thanks for your concerns on our work.    I have downloaded the code from this link (https://github.com/weiweisong415/Demo_AHCL_for_TGRS2022) and carefully re-implemented the experiments without any changes except for the WHURS19 dataset should  be upzipped in "./data/WHURS-19/". I have not found any problems about loss and map as shown in the following figures. You could use the attached dataset and try it again.    Please contact me again if you have any questions. Good luck. Weiwei Song Peng Cheng Laboratory    

…

------------------ Original ------------------ From: ***@***.***>; Date:  Fri, Mar 31, 2023 12:20 PM To: ***@***.***>; Cc: ***@***.***>; Subject:  [weiweisong415/Demo_AHCL_for_TGRS2022] About the [Iteration: 29/ 30][Train Loss: 279.2204] (Issue #1)   Hello, Recently, I did an experiment on remote sensing, but I didn't modify any parameters during running this code, but the loss is very large and the map=0.0642. I am very confused about this problem, and I hope to get your answer. article：“Asymmetric Hash Code Learning for Remote Sensing Image Retrieval” — Reply to this email directly, view it on GitHub, or unsubscribe. You are receiving this because you are subscribed to this thread.Message ID: ***@***.***> 从腾讯企业邮箱发来的超大附件 WHURS19.zip (117.45M, 2023年04月30日 12:51 到期)进入下载页面：https://exmail.qq.com/cgi-bin/ftnExs_download?k=22656534dbfdccc0a96d921f4634530a4c505c56500c020a574857505200490c02525d19035701014e570457550053095301035251346f38342d306632055d16190c15345c&t=exs_ftn_download&code=cee4a4d8&s=email&fid=72/59b18f24-2d34-4a78-bce9-2ac44710dff0

[scripted07]

hey bro i am also doing the same. can you please help me out .

[jingzhang697]

I solved this problem about loss through retraining, and the retrieval accuracy matches the original text.

…

------------------ 原始邮件 ------------------ 发件人: "weiweisong415/Demo_AHCL_for_TGRS2022" ***@***.***>; 发送时间: 2023年4月24日(星期一) 凌晨3:44 ***@***.***>; ***@***.******@***.***>; 主题: Re: [weiweisong415/Demo_AHCL_for_TGRS2022] About the [Iteration: 29/ 30][Train Loss: 279.2204] (Issue #1) hey bro i am also doing the same. can you please help me out . — Reply to this email directly, view it on GitHub, or unsubscribe. You are receiving this because you authored the thread.Message ID: ***@***.***>