About the pretrain model

[luchen828]

hi,Can the pre-training model for fitness tracking be provided?

[guillaumebrg]

Hi @luchen828, the fitness tracking model should already be provided. Have you followed the README instructions and downloaded the model weights from here?

[luchen828]

are you sure ?i mean the calorise model

[guillaumebrg]

Yes, I think the calorie model should be there as well. Could you run ls -R from your resources folder and paste the output here?

[luchen828]

sien_lite_met_converter.ckpt simn_lite_met_converter.ckpt
sien_met_converter.ckpt simn_met_converter.ckpt
but the size of sien_met_converter.ckpt is only 514kB, is right?

[guillaumebrg]

Yes, that's right. It's small because those files only contain the weights of the calorie "head" network. Weights of lower layers should be in the backbone folder.

I think you ran ls -R from resources/calorie_estimation. Could you do the same from resources so that we see everything?

[luchen828]

gesture_recognition fitness_activity_recognition backbone
fitness_rep_counting calorie_estimation action_recognition

[guillaumebrg]

What do you have in resources/backbone?

[luchen828]

strided_inflated_efficientnet.ckpt strided_inflated_mobilenet.ckpt

[guillaumebrg]

Looks good to me. Are you on the latest master branch? You might have to git pull.
What does the error look like when you run the calorie estimation script?

[luchen828]

when run the calorie estimation script
the error is " please report a bug to PyTorch. Attempted to read a PyTorch file with version 3, but the maximum supported version for reading is 2. Your PyTorch installation may be too old"
it's like the pytorch version you train the model is too high

[guillaumebrg]

it's like the pytorch version you train the model is too high

... or that your local PyTorch version is too old ;)

PyTorch changed its serialization format not so long ago. Can you try to upgrade your PyTorch and run it again?

[luchen828]

yes,my torch is 1.4 version.it's to low comparied with 1.7 .thank you very much!

[JonyJiang123]

@luchen828 Could you share pre-trained weights which downloaded from https://20bn.com/licensing/sdk/evaluation ? The web was broken

[luchen828]

------------------ 原始邮件 ------------------ 发件人: "quic/sense" ***@***.***>; 发送时间: 2021年11月10日(星期三) 中午11:17 ***@***.***>; ***@***.******@***.***>; 主题: Re: [quic/sense] About the pretrain model (#178) @luchen828 Could you share pre-trained weights which downloaded from https://20bn.com/licensing/sdk/evaluation ? The web was broken — You are receiving this because you were mentioned. Reply to this email directly, view it on GitHub, or unsubscribe. Triage notifications on the go with GitHub Mobile for iOS or Android.

[JonyJiang123]

------------------ 原始邮件 ------------------ 发件人: "quic/sense" @.>; 发送时间: 2021年11月10日(星期三) 中午11:17 @.>; @.@.>; 主题: Re: [quic/sense] About the pretrain model (#178) @luchen828 Could you share pre-trained weights which downloaded from https://20bn.com/licensing/sdk/evaluation ? The web was broken — You are receiving this because you were mentioned. Reply to this email directly, view it on GitHub, or unsubscribe. Triage notifications on the go with GitHub Mobile for iOS or Android.

what's this, Bro

[luchen828]

模型太大被限制了，发不过去

…

------------------ 原始邮件 ------------------ 发件人: "quic/sense" ***@***.***>; 发送时间: 2021年11月10日(星期三) 中午11:29 ***@***.***>; ***@***.******@***.***>; 主题: Re: [quic/sense] About the pretrain model (#178)

------------------ 原始邮件 ------------------ 发件人: "quic/sense" @.>; 发送时间: 2021年11月10日(星期三) 中午11:17 @.>; @.@.>; 主题: Re: [quic/sense] About the pretrain model (#178) @luchen828 Could you share pre-trained weights which downloaded from https://20bn.com/licensing/sdk/evaluation ? The web was broken — You are receiving this because you were mentioned. Reply to this email directly, view it on GitHub, or unsubscribe. Triage notifications on the go with GitHub Mobile for iOS or Android. what's this, Bro — You are receiving this because you were mentioned. Reply to this email directly, view it on GitHub, or unsubscribe. Triage notifications on the go with GitHub Mobile for iOS or Android.

[JonyJiang123]

模型太大被限制了，发不过去
…
------------------ 原始邮件 ------------------ 发件人: "quic/sense" @.>; 发送时间: 2021年11月10日(星期三) 中午11:29 @.>; @.@.>; 主题: Re: [quic/sense] About the pretrain model (#178)
------------------ 原始邮件 ------------------ 发件人: "quic/sense" @.>; 发送时间: 2021年11月10日(星期三) 中午11:17 @.>; @.@.>; 主题: Re: [quic/sense] About the pretrain model (#178) @luchen828 Could you share pre-trained weights which downloaded from https://20bn.com/licensing/sdk/evaluation ? The web was broken — You are receiving this because you were mentioned. Reply to this email directly, view it on GitHub, or unsubscribe. Triage notifications on the go with GitHub Mobile for iOS or Android. what's this, Bro — You are receiving this because you were mentioned. Reply to this email directly, view it on GitHub, or unsubscribe. Triage notifications on the go with GitHub Mobile for iOS or Android.

Could you send it to my email 13576080222@163.com? or any other ways?

[MoonBunnyZZZ]

yes,my torch is 1.4 version.it's to low comparied with 1.7 .thank you very much!

可以分享一下模型文件吗？比如放到网盘上 @luchen828

[KeonAcute]

@luchen828 麻烦能分享一下权重模型的文件吗？我的邮箱17601026331@163.com，非常感谢！

[Ozone-oo]

@luchen828 可以分享下backbone权重模型文件？邮箱1871805784@qq.com

[koralm]

Anyone can share pre-trained weights ?

[Ozone-oo]

Anyone can share pre-trained weights ?

https://1drv.ms/u/s!AgK8kNkUacE6mKkHlstyUExsKe6N_A?e=frmHdS