Giter Site home page Giter Site logo

Comments (42)

wyy123de avatar wyy123de commented on July 29, 2024 1

image
我的损失比图像损失还小

from anomalyclip.

vvthai10 avatar vvthai10 commented on July 29, 2024

from anomalyclip.

vvthai10 avatar vvthai10 commented on July 29, 2024

and how can you show the table after train 15 epochs, I not find way to show this

from anomalyclip.

wyy123de avatar wyy123de commented on July 29, 2024

image

from anomalyclip.

wyy123de avatar wyy123de commented on July 29, 2024

这个你要自己保存到results中,我用的Head-CT数据集
还可以可视化呢

from anomalyclip.

wyy123de avatar wyy123de commented on July 29, 2024

image

from anomalyclip.

vvthai10 avatar vvthai10 commented on July 29, 2024

You train by Brain dataset, not mvtec

from anomalyclip.

vvthai10 avatar vvthai10 commented on July 29, 2024

Can u share me code you visualize graphwhen train model

from anomalyclip.

wyy123de avatar wyy123de commented on July 29, 2024

对啊,我是医学影响异常检测

from anomalyclip.

wyy123de avatar wyy123de commented on July 29, 2024

sure

from anomalyclip.

vvthai10 avatar vvthai10 commented on July 29, 2024

I'm also studying the medical part like you, learnign_rate u still 0.001 ?

from anomalyclip.

wyy123de avatar wyy123de commented on July 29, 2024

yes

from anomalyclip.

vvthai10 avatar vvthai10 commented on July 29, 2024

can u share me link of dataset u use,

from anomalyclip.

vvthai10 avatar vvthai10 commented on July 29, 2024

I try train with mvtec, after 15 epoch, loss, and image_loss is 3.4, 0.3 but when test result still okay

from anomalyclip.

wyy123de avatar wyy123de commented on July 29, 2024

嗯嗯HeadCT

from anomalyclip.

wyy123de avatar wyy123de commented on July 29, 2024

我感觉我训练有BUG

from anomalyclip.

wyy123de avatar wyy123de commented on July 29, 2024

原文效果没那么好,哈哈哈哈

from anomalyclip.

wyy123de avatar wyy123de commented on July 29, 2024

HeadCT

from anomalyclip.

wyy123de avatar wyy123de commented on July 29, 2024

parser.add_argument("--train_data_path", type=str, default="/root/Downloads/Untitled Folder/AnomalyCLIP-main", help="train dataset path")
parser.add_argument("--save_path", type=str, default='/root/Downloads/Untitled Folder/AnomalyCLIP-main/results', help='path to save results')
这个是保存代码,其实作者给的代码很详细,你要保存到位置就可以看到

from anomalyclip.

vvthai10 avatar vvthai10 commented on July 29, 2024

And now i want to add adapter to image encoder, u want try with me

from anomalyclip.

vvthai10 avatar vvthai10 commented on July 29, 2024

"The original text is not that good" I don't understand what u say.

from anomalyclip.

wyy123de avatar wyy123de commented on July 29, 2024

image
训练出现了BUG

from anomalyclip.

wyy123de avatar wyy123de commented on July 29, 2024

导致效果好

from anomalyclip.

wyy123de avatar wyy123de commented on July 29, 2024

image

from anomalyclip.

vvthai10 avatar vvthai10 commented on July 29, 2024

你的意思是你的结果比文章高很多吧?

from anomalyclip.

wyy123de avatar wyy123de commented on July 29, 2024

from anomalyclip.

vvthai10 avatar vvthai10 commented on July 29, 2024

看起来他们在文章中使用 mvtec 进行训练并在医疗设备上进行评估,但如果你使用医疗包进行训练,我认为结果会有所不同。

from anomalyclip.

wyy123de avatar wyy123de commented on July 29, 2024

image
跑了100个epoch

from anomalyclip.

vvthai10 avatar vvthai10 commented on July 29, 2024

为什么训练了 100 个 epoch,结果却低于 15 个 epoch?

from anomalyclip.

vvthai10 avatar vvthai10 commented on July 29, 2024

在数据集文件中,我没有看到的generate_class_info函数返回大脑数据集的obj_list

from anomalyclip.

wyy123de avatar wyy123de commented on July 29, 2024

100个epoch,训练的多,我学习率调到0.000001,15个epoch学习率0.001

from anomalyclip.

vvthai10 avatar vvthai10 commented on July 29, 2024

我发现在这种类型的练习中,他们仅训练 15 或 50 个 epoch,lr 为 0.001 或 0.0001。 因为训练数据很小,训练多了就怕过拟合

from anomalyclip.

wyy123de avatar wyy123de commented on July 29, 2024

yes

from anomalyclip.

vvthai10 avatar vvthai10 commented on July 29, 2024

但是你的损失确实很低,我不明白为什么,因为当我用mvtec文件训练时,它是3.x,哈哈哈

from anomalyclip.

vvthai10 avatar vvthai10 commented on July 29, 2024

您想与您的团队就这个主题进行合作吗? 我正在尝试更新图像编码器部分以将其合并到本文中。

from anomalyclip.

wyy123de avatar wyy123de commented on July 29, 2024

可能你才是对的,不过损失不下降,可能是欠拟合,你可以尝试调大epoch,或者学习率

但是你的损失确实很低,我不明白为什么,因为当我用mvtec文件训练时,它是3.x,哈哈哈

from anomalyclip.

wyy123de avatar wyy123de commented on July 29, 2024

from anomalyclip.

vvthai10 avatar vvthai10 commented on July 29, 2024

from anomalyclip.

vvthai10 avatar vvthai10 commented on July 29, 2024

这里没有显示我的电子邮件地址。我的 Gmail 是 vuvanthai1410。

from anomalyclip.

wyy123de avatar wyy123de commented on July 29, 2024

ok

from anomalyclip.

vvthai10 avatar vvthai10 commented on July 29, 2024

请给我发电子邮件,我会详细告诉你我在做什么:)

from anomalyclip.

vvthai10 avatar vvthai10 commented on July 29, 2024

yeah, i ssee

from anomalyclip.

Related Issues (20)

Recommend Projects

  • React photo React

    A declarative, efficient, and flexible JavaScript library for building user interfaces.

  • Vue.js photo Vue.js

    🖖 Vue.js is a progressive, incrementally-adoptable JavaScript framework for building UI on the web.

  • Typescript photo Typescript

    TypeScript is a superset of JavaScript that compiles to clean JavaScript output.

  • TensorFlow photo TensorFlow

    An Open Source Machine Learning Framework for Everyone

  • Django photo Django

    The Web framework for perfectionists with deadlines.

  • D3 photo D3

    Bring data to life with SVG, Canvas and HTML. 📊📈🎉

Recommend Topics

  • javascript

    JavaScript (JS) is a lightweight interpreted programming language with first-class functions.

  • web

    Some thing interesting about web. New door for the world.

  • server

    A server is a program made to process requests and deliver data to clients.

  • Machine learning

    Machine learning is a way of modeling and interpreting data that allows a piece of software to respond intelligently.

  • Game

    Some thing interesting about game, make everyone happy.

Recommend Org

  • Facebook photo Facebook

    We are working to build community through open source technology. NB: members must have two-factor auth.

  • Microsoft photo Microsoft

    Open source projects and samples from Microsoft.

  • Google photo Google

    Google ❤️ Open Source for everyone.

  • D3 photo D3

    Data-Driven Documents codes.