Changing IDAup out_dim raised RuntimeError when loading a checkpoint

[sunxiaoyang01]

It is a very excellent job! But I get an issue. I am currently encountering a dimension mismatch problem when setting the upsampling dimensions to (3,3) in the DLA model. Despite my efforts, I have not been able to resolve this issue.

[Alexandre-Delplanque]

Hi @sunxiaoyang01 ,
Thanks for opening this issue.
Could you please provide the code you ran that caused the error?

[sunxiaoyang01]

Of course I can.

[Alexandre-Delplanque]

@sunxiaoyang01 ,
The error has occurred because you have tried to load a checkpoint whose parameters do not match the architecture of your model.
Try setting pretrained=False when instantiating the model.

[sunxiaoyang01]

Yes, I have pre trained=False, but this issue still occurs

[Alexandre-Delplanque]

@sunxiaoyang01 ,
Thank you for providing the code snippet. Next time, please provide me with the entire code you ran when opening the issue, so that I can better understand the cause of the error.

The error still appears because you're loading parameters from a checkpoint, which again doesn't match the architecture you've modified:
model.load_state_dict(checkpoint(['model_state_dict'])

[sunxiaoyang01]

@Alexandre-Delplanque Thank you for your answer. I ran the tool/infer. py file in the file and used the model you provided, but then the above error occurred.

…

------------------ 原始邮件 ------------------ 发件人: "Alexandre-Delplanque/HerdNet" ***@***.***>; 发送时间: 2024年6月19日(星期三) 下午3:17 ***@***.***>; ***@***.******@***.***>; 主题: Re: [Alexandre-Delplanque/HerdNet] Changing IDAup out_dim raised RuntimeError (Issue #7) @sunxiaoyang01 , Thank you for providing the code snippet. Next time, please provide me with the entire code you ran when opening the issue, so that I can better understand the cause of the error. The error still appears because you're loading parameters from a checkpoint, which again doesn't match the architecture you've modified: model.load_state_dict(checkpoint(['model_state_dict']) — Reply to this email directly, view it on GitHub, or unsubscribe. You are receiving this because you were mentioned.Message ID: ***@***.***> 从QQ邮箱发来的超大附件 HerdNet-main.zip (500.46M, 2024年07月19日 15:30 到期)进入下载页面：https://mail.qq.com/cgi-bin/ftnExs_download?k=2630316473afddd0766ffc331e37571f1b4402525a5607054e075452081a045605011c565a00011d5202075c0155040002060651382765780642552a5d43485d02595f4a425e15305e&t=exs_ftn_download&code=c01d87e0

[Alexandre-Delplanque]

@sunxiaoyang01 ,
I don't see any code or screenshot, please provide enough details for me to understand the error.

[sunxiaoyang01]

@Alexandre-Delplanque I have attached the file to the email.

…

------------------ 原始邮件 ------------------ 发件人: "Alexandre-Delplanque/HerdNet" ***@***.***>; 发送时间: 2024年6月20日(星期四) 下午4:30 ***@***.***>; ***@***.******@***.***>; 主题: Re: [Alexandre-Delplanque/HerdNet] Changing IDAup out_dim raised RuntimeError when loading a checkpoint (Issue #7) @sunxiaoyang01 , I don't see any code or screenshot, please provide enough details for me to understand the error. — Reply to this email directly, view it on GitHub, or unsubscribe. You are receiving this because you were mentioned.Message ID: ***@***.***> 从QQ邮箱发来的超大附件 HerdNet-main.zip (500.46M, 2024年07月20日 16:41 到期)进入下载页面：https://mail.qq.com/cgi-bin/ftnExs_download?k=2437396266c68dd57468f4351264071a19435f56570704064c520a035549060c500f145756515618530309500055530653050c543474357d04455d2c51101858005e574c4e0d45355c&t=exs_ftn_download&code=a79b4d55

[Alexandre-Delplanque]

@sunxiaoyang01 ,
Please put it directly on Github

[Alexandre-Delplanque]

Hi @sunxiaoyang01,
Problem solved?