Open ImSuMyatNoe opened 2 years ago
ImSuMyatNoe
commented
2 years ago I have not found train_mars.py in the git repository. As you mentioned, to make the Re-ID for the own dataset, train_mars.py file is used to train the data set in your explanation paper. Could you please let me know where can I find this train_mars.py.
ImSuMyatNoe
commented
2 years ago When you train the cosine-metric learning from the original deep sort repository, did you use tensor flow 1 .x ? Is it compatible ?
John1liu
commented
2 years ago Firstly, Thank you very much for sharing and your implementation is amazing!
As I am currently new to the Deep -SORT, May I know about the step by step process of how to Re train Re-ID Model on your custom dataset? I would like also know about if ID switch case occur in your dataset ? If so, could you please tell me how to solve this issue.
Sorry, I still busy on my research and I am not focus on CV now. So I reply you so late. As for the first question, I suggest you learn how to re-train the deep-sort model from https://github.com/nwojke/cosine_metric_learning. It is a good repo for learning it. And of course, there are many ID swith cases in my project, because my datasets is too small. I think you can use a bigger enough datasets to train the deep-sort model as original datasets in original paper. As we all know the ID switch can be caused by underfitting, big datasets may solve this problem. As for the second problem, you can find this file in aforementioned repo. I am sorry my paper is in Chinese and I am amazing that you can understand Chinese. If you are interested in my datasets, please give me your email, I will send it to you. The last question, I still remember I use tensorflow 2. I guess tensorflow 1.0+ also can be useful but I am not sure, I suggest you can test it in Google Colab. I am sorry for my late again. And I will not update the Deep-Sort part because I am not focus on CV area now. I am study in Graph Neural Network now, if you are interested in this field, we can have dicussion. Sorry again.
Best Regards, Zhuoyang LIU
ImSuMyatNoe
commented
2 years ago Dear Zhuoyang LIU,
Thank you very much for your kind explanation. I am also interested about your datasets and if you don't mind could you please send it to my email contact.sumyatnoe@gmail.com. I am also interested about Graph Neural Network also. Let's have discussion more.
Best , Su Myat Noe
John1liu
commented
2 years ago Dear Su Myat Noe,
It is too large to transfer from email. You can find it from this link. https://drive.google.com/file/d/1lushuv4QMTmfFwURU1Ug6mezkzQAMi0S/view?usp=sharing
Best Regards, Zhuoyang LIU
发件人: Su Myat Noe @.> 发送时间: 2021年11月12日 10:40 收件人: John1liu/YOLOV5-DeepSORT-Vehicle-Tracking-Master @.> 抄送: John1liu @.>; Comment @.> 主题: Re: [John1liu/YOLOV5-DeepSORT-Vehicle-Tracking-Master] How to Retrain ReID Model on your custom dataset? (Issue #3)
Dear Zhuoyang LIU,
Thank you very much for your kind explanation. I am also interested about your datasets and if you don't mind could you please send it to my email @.**@.>. I am also interested about Graph Neural Network also. Let's have discussion more.
Best , Su Myat Noe
― You are receiving this because you commented. Reply to this email directly, view it on GitHubhttps://github.com/John1liu/YOLOV5-DeepSORT-Vehicle-Tracking-Master/issues/3#issuecomment-966773916, or unsubscribehttps://github.com/notifications/unsubscribe-auth/APKUIZV33UZ7HHO75YJH3A3ULR5ATANCNFSM5GIHR67A. Triage notifications on the go with GitHub Mobile for iOShttps://apps.apple.com/app/apple-store/id1477376905?ct=notification-email&mt=8&pt=524675 or Androidhttps://play.google.com/store/apps/details?id=com.github.android&referrer=utm_campaign%3Dnotification-email%26utm_medium%3Demail%26utm_source%3Dgithub.
ImSuMyatNoe
commented
2 years ago Dear Zhuoyang LIU,
Thank you very much for your dataset.
Best, Su Myat Noe
John1liu
commented
2 years ago Nothing. But to be honestly, I do not think my research has any value for furthermore research. It will be a good example for engineering application, but not for research. I think you should optimize the algorithm from basement of math.
Get Outlook for Androidhttps://aka.ms/AAb9ysg
发件人: Su Myat Noe @.***> 发送时间: 2021年11月12日星期五 下午5:06 收件人: John1liu/YOLOV5-DeepSORT-Vehicle-Tracking-Master 抄送: John1liu; Comment 主题: Re: [John1liu/YOLOV5-DeepSORT-Vehicle-Tracking-Master] How to Retrain ReID Model on your custom dataset? (Issue #3)
Dear Zhuoyang LIU,
Thank you very much for your dataset.
Best, Su Myat Noe
― You are receiving this because you commented. Reply to this email directly, view it on GitHubhttps://github.com/John1liu/YOLOV5-DeepSORT-Vehicle-Tracking-Master/issues/3#issuecomment-966938307, or unsubscribehttps://github.com/notifications/unsubscribe-auth/APKUIZUI4K4ESAKBLI5IMV3ULTKJ5ANCNFSM5GIHR67A. Triage notifications on the go with GitHub Mobile for iOShttps://apps.apple.com/app/apple-store/id1477376905?ct=notification-email&mt=8&pt=524675 or Androidhttps://play.google.com/store/apps/details?id=com.github.android&referrer=utm_campaign%3Dnotification-email%26utm_medium%3Demail%26utm_source%3Dgithub.
Firstly, Thank you very much for sharing and your implementation is amazing!
As I am currently new to the Deep -SORT, May I know about the step by step process of how to Re train Re-ID Model on your custom dataset? I would like also know about if ID switch case occur in your dataset ? If so, could you please tell me how to solve this issue.