3DV-Action icon indicating copy to clipboard operation
3DV-Action copied to clipboard

NTU 60 results

Open JinyangGuo opened this issue 3 years ago • 6 comments

Hi,

Thanks for your interesting work. I use the python code to train the 3DV-PointNet++ on NTU 60. However, I only get 93.4 for cross_view and 85.6 for cross_subject. I first generate the 3dv data and change the data path in train.py. Then I run the train.py file in your code. Do I make a mistake at some points?

JinyangGuo avatar Jul 19 '21 13:07 JinyangGuo

Hi,

Thanks for your interesting work. I use the python code to train the 3DV-PointNet++ on NTU 60. However, I only get 93.4 for cross_view and 85.6 for cross_subject. I first generate the 3dv data and change the data path in train.py. Then I run the train.py file in your code. Do I make a mistake at some points?

Hi Jinyang,

Do you know what's the right way to run the codes provided here? Thanks.

NeilCui6 avatar Sep 21 '21 00:09 NeilCui6

Hi,

I still don’t know the right way to run the code.

Regards, Jinyang

发件人: @.> 发送时间: 2021年9月21日 10:42 收件人: @.> 抄送: Jinyang @.>; @.> 主题: Re: [3huo/3DV-Action] NTU 60 results (#7)

Hi,

Thanks for your interesting work. I use the python code to train the 3DV-PointNet++ on NTU 60. However, I only get 93.4 for cross_view and 85.6 for cross_subject. I first generate the 3dv data and change the data path in train.py. Then I run the train.py file in your code. Do I make a mistake at some points?

Hi Jinyang,

Do you know what's the right way to run the codes provided here? Thanks.

― You are receiving this because you authored the thread. Reply to this email directly, view it on GitHubhttps://github.com/3huo/3DV-Action/issues/7#issuecomment-923484798, or unsubscribehttps://github.com/notifications/unsubscribe-auth/AJBBPAFRXMREKYGXBCUPEDTUC7IHNANCNFSM5ATVBFUQ. 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.

JinyangGuo avatar Sep 21 '21 00:09 JinyangGuo

Hi @JinyangGuo,

I think you have misunderstood my previous message. I am not the owner of the code, but just a new user as well....

I was trying to run the code but got some issues. Can I ask you a few questions for help?

  1. Do we need to firstly download the dataset? If so, where can I find it?
  2. After getting the dataset, do we need to run the codes in "3DV_construction" to transfer the images to the points data?
  3. Then split the points data into training and testing, and feed training points data to train file in "NTU120_Net" for training. Feeding testing points data to test.py in "NTU120_Net" for testing?

Is the procedure above right?

Thanks a lot, Jinyang.

NeilCui6 avatar Sep 21 '21 01:09 NeilCui6

Hi,

I think your procedure is correct. You can find the original dataset from the link provided in the repo. The other steps are correct.

Regards, Jinyang

发件人: @.> 发送时间: 2021年9月21日 11:12 收件人: @.> 抄送: Jinyang @.>; @.> 主题: Re: [3huo/3DV-Action] NTU 60 results (#7)

Hi @JinyangGuohttps://github.com/JinyangGuo,

I think you have misunderstood my previous message. I am not the owner of the code, but just a new user as well....

I was trying to run the code but got some issues. Can I ask you a few questions for help?

  1. Do we need to firstly download the dataset? If so, where can I find it?
  2. After getting the dataset, do we need to run the codes in "3DV_construction" to transfer the images to the points data?
  3. Then split the points data into training and testing, and feed training points data to train file in "NTU120_Net" for training. Feeding testing points data to test.py in "NTU120_Net" for testing?

Is the procedure above right?

Thanks a lot, Jinyang.

― You are receiving this because you were mentioned. Reply to this email directly, view it on GitHubhttps://github.com/3huo/3DV-Action/issues/7#issuecomment-923496422, or unsubscribehttps://github.com/notifications/unsubscribe-auth/AJBBPAHGBFT3FVZBYVP7F3TUC7LWLANCNFSM5ATVBFUQ. 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.

JinyangGuo avatar Sep 21 '21 02:09 JinyangGuo

Thanks, Jinyang.

NeilCui6 avatar Sep 21 '21 02:09 NeilCui6

Hi, I think your procedure is correct. You can find the original dataset from the link provided in the repo. The other steps are correct. Regards, Jinyang 发件人: @.> 发送时间: 2021年9月21日 11:12 收件人: @.> 抄送: Jinyang @.>; @.> 主题: Re: [3huo/3DV-Action] NTU 60 results (#7) Hi @JinyangGuohttps://github.com/JinyangGuo, I think you have misunderstood my previous message. I am not the owner of the code, but just a new user as well.... I was trying to run the code but got some issues. Can I ask you a few questions for help? 1. Do we need to firstly download the dataset? If so, where can I find it? 2. After getting the dataset, do we need to run the codes in "3DV_construction" to transfer the images to the points data? 3. Then split the points data into training and testing, and feed training points data to train file in "NTU120_Net" for training. Feeding testing points data to test.py in "NTU120_Net" for testing? Is the procedure above right? Thanks a lot, Jinyang. ― You are receiving this because you were mentioned. Reply to this email directly, view it on GitHub<#7 (comment)>, or unsubscribehttps://github.com/notifications/unsubscribe-auth/AJBBPAHGBFT3FVZBYVP7F3TUC7LWLANCNFSM5ATVBFUQ. 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.

Hi Jinyang, I still got some issues when running the "ntu120_3dv_pre.py" file within 3DV_construction folder. If it's convenient for you, can I ask you a few more questions? Which dataset were you using via the link provided by the repo below? Did you do any adjustments to the original codes?

https://pan.baidu.com/s/1TtApXBQx4sNi4ZmTOL6lSw code:k0jt

Many thanks!

NeilCui6 avatar Sep 22 '21 06:09 NeilCui6