When will this code be open source ??? I'm waiting all the time...thank you very much

[plshuaishuai2]

Your work is so great and I am eagerly awaiting the opportunity to explore your work in greater detail.Could you please provide an update on when the code might be available? Your work is of great interest to me and every researcher, and having access to the code would be incredibly beneficial.
Thank you for your time and consideration.

[tzjtatata]

I am sorry for that. But I can not give a precise date due to some uncontrollable issues. If you want to know more details about Myriad, please feel free to contact me via ***@***.***).

…

------------------ 原始邮件 ------------------ 发件人: "tzjtatata/Myriad" ***@***.***>; 发送时间: 2024年1月15日(星期一) 晚上8:58 ***@***.***>; ***@***.***>; 主题: [tzjtatata/Myriad] When will this code be open source ??? I'm waiting all the time...thank you very much (Issue #5) Your work is so great and I am eagerly awaiting the opportunity to explore your work in greater detail.Could you please provide an update on when the code might be available? Your work is of great interest to me and every researcher, and having access to the code would be incredibly beneficial. Thank you for your time and consideration. — Reply to this email directly, view it on GitHub, or unsubscribe. You are receiving this because you are subscribed to this thread.Message ID: ***@***.***>

[EddieEduardo]

Hello, great work!

I'd like to ask how the model is trained. For example, is the model trained on the training set of MVTec-AD, with images and corresponding text as input? How is the text defined?

Thanks for reply！

[tzjtatata]

Thank you for your comment. We evaluate our model by cross evaluation. 1. We train our model on MVtec-AD and evaluate it on VisA. And the model trained on VisA is tested on MVtec-AD. Note that we only use normal data and simulated anomaly data by NSA(just like cut-paste).  2. Our model is trained with the simple text: 'Yes, there are defects.' and 'No, there is no defect.' due to the simulated data. By the way, we are working at the IAD data with more details. 3. The code will be public recently. Thanks for waiting.

…

------------------ 原始邮件 ------------------ 发件人: ***@***.***>; 发送时间: 2024年3月12日(星期二) 晚上8:38 收件人: ***@***.***>; 抄送: ***@***.***>; ***@***.***>; 主题: Re: [tzjtatata/Myriad] When will this code be open source ??? I'm waiting all the time...thank you very much (Issue #5) Hello, great work! I'd like to ask how the model is trained. For example, is the model trained on the training set of MVTec-AD, with images and corresponding text as input? How is the text defined? Thanks for reply！ — Reply to this email directly, view it on GitHub, or unsubscribe. You are receiving this because you commented.Message ID: ***@***.***>

[EddieEduardo]

Understood! Thank you for your response.