search

kaiwang960112 / Self-Cure-Network

This is a novel and easy method for annotation uncertainties.
409 stars 96 forks source link

您好我想请问一下您是如何在ferplus根据投票标签提取硬标签的呢，是直接取最大值还是取超过投票数半数的作为标签啊。谢谢 #48

Closed muse1998 closed 9 months ago

kaiwang960112 commented 3 years ago

max

Diy_pooling @.***> 於 2021年4月28日週三 下午1:20寫道：

— You are receiving this because you are subscribed to this thread. Reply to this email directly, view it on GitHub https://github.com/kaiwang960112/Self-Cure-Network/issues/48, or unsubscribe https://github.com/notifications/unsubscribe-auth/AI6LK4EUGSWT3LHAXN2USDDTK6LI3ANCNFSM43WKFRTA .

muse1998 commented 3 years ago

max Diy_pooling @.***> 於 2021年4月28日週三 下午1:20寫道： … — You are receiving this because you are subscribed to this thread. Reply to this email directly, view it on GitHub <#48>, or unsubscribe https://github.com/notifications/unsubscribe-auth/AI6LK4EUGSWT3LHAXN2USDDTK6LI3ANCNFSM43WKFRTA .

那您是将灰度图像直接复制成三通道的输入到SCN中进行处理吗，和您提供的RAF-DB的代码在预处理部分有什么区别呢。

kaiwang960112 commented 3 years ago

是复制成三通道的，然后size是224*224哈，具体可以refer论文里哈

Diy_pooling @.***> 於 2021年4月28日週三 下午2:50寫道：

max Diy_pooling @.***> 於 2021年4月28日週三 下午1:20寫道： … <#m8387694905756452998> — You are receiving this because you are subscribed to this thread. Reply to this email directly, view it on GitHub <#48 https://github.com/kaiwang960112/Self-Cure-Network/issues/48>, or unsubscribe https://github.com/notifications/unsubscribe-auth/AI6LK4EUGSWT3LHAXN2USDDTK6LI3ANCNFSM43WKFRTA .

那您是将灰度图像直接复制成三通道的输入到SCN中进行处理吗，和您提供的RAF-DB的代码在预处理部分有什么区别呢。

— You are receiving this because you commented. Reply to this email directly, view it on GitHub https://github.com/kaiwang960112/Self-Cure-Network/issues/48#issuecomment-828194131, or unsubscribe https://github.com/notifications/unsubscribe-auth/AI6LK4HFKAZDXBPTY77FMNLTK6V4DANCNFSM43WKFRTA .

yangjian1218 commented 2 years ago

fer图片比较小,很模糊,训练出来的准确度才0.7.作者怎么达到的89%?

kaiwang960112 commented 2 years ago

杨健 @.***>于2021年10月28日 周四下午2:04写道：

fer图片比较小,很模糊,训练出来的准确度才0.7.作者怎么达到的89%?

— You are receiving this because you commented. Reply to this email directly, view it on GitHub https://github.com/kaiwang960112/Self-Cure-Network/issues/48#issuecomment-953528842, or unsubscribe https://github.com/notifications/unsubscribe-auth/AI6LK4ANFNCHICUDSZZVW3LUJDRWHANCNFSM43WKFRTA . Triage notifications on the go with GitHub Mobile for iOS https://apps.apple.com/app/apple-store/id1477376905?ct=notification-email&mt=8&pt=524675 or Android https://play.google.com/store/apps/details?id=com.github.android&referrer=utm_campaign%3Dnotification-email%26utm_medium%3Demail%26utm_source%3Dgithub.

你莫非是没resize

yangjian1218 commented 2 years ago

@kaiwang960112 在ransforms.Compose就有resize了,没有单独设置, 我是针对fer2013的数据跟自己制作标签,并没有针对ferplus的标签来训练. 另外问一下,如果batchsize设置128或者256 , margin_1还是设置为0.07好么?还是可以增加. 我增加了好像没效果

qiang333 commented 2 years ago

是复制成三通道的，然后size是224*224哈，具体可以refer论文里哈 Diy_pooling @.> 於 2021年4月28日週三 下午2:50寫道： … max Diy_pooling @.> 於 2021年4月28日週三 下午1:20寫道： … <#m8387694905756452998> — You are receiving this because you are subscribed to this thread. Reply to this email directly, view it on GitHub <#48 <#48>>, or unsubscribe https://github.com/notifications/unsubscribe-auth/AI6LK4EUGSWT3LHAXN2USDDTK6LI3ANCNFSM43WKFRTA . 那您是将灰度图像直接复制成三通道的输入到SCN中进行处理吗，和您提供的RAF-DB的代码在预处理部分有什么区别呢。 — You are receiving this because you commented. Reply to this email directly, view it on GitHub <#48 (comment)>, or unsubscribe https://github.com/notifications/unsubscribe-auth/AI6LK4HFKAZDXBPTY77FMNLTK6V4DANCNFSM43WKFRTA .

作者您好，fer plus数据集里面有一些not unknown or non-face 的标签，有这些标签的样本是不参与训练对嘛。我在对fer plus处理数据集之后训练集大概是 25000，而原始的训练集是28709。而且我看好多论文的描述都说fer plus训练集是28709，所以就很疑惑。

kaiwang960112 commented 2 years ago

那些不知道和没有人脸的，我是直接去除的。标签取最大的。

发自我的iPhone

在 2022年9月14日，上午11:03，qiang333 @.***> 写道：

 是复制成三通道的，然后size是224224哈，具体可以refer论文里哈 Diy_pooling @.> 於 2021年4月28日週三 下午2:50寫道： … max Diy_pooling @.*> 於 2021年4月28日週三 下午1:20寫道： … <#m8387694905756452998> — You are receiving this because you are subscribed to this thread. Reply to this email directly, view it on GitHub <#48 <#48>>, or unsubscribe https://github.com/notifications/unsubscribe-auth/AI6LK4EUGSWT3LHAXN2USDDTK6LI3ANCNFSM43WKFRTA . 那您是将灰度图像直接复制成三通道的输入到SCN中进行处理吗，和您提供的RAF-DB的代码在预处理部分有什么区别呢。 — You are receiving this because you commented. Reply to this email directly, view it on GitHub <#48 (comment)>, or unsubscribe https://github.com/notifications/unsubscribe-auth/AI6LK4HFKAZDXBPTY77FMNLTK6V4DANCNFSM43WKFRTA .

作者您好，fer plus数据集里面有一些not unknown or non-face 的标签，有这些标签的样本是不参与训练对嘛。我在对fer plus处理数据集之后训练集大概是 25000，而原始的训练集是28709。而且我看好多论文的描述都说fer plus训练集是28709，所以就很疑惑。

— Reply to this email directly, view it on GitHub, or unsubscribe. You are receiving this because you were mentioned.

zerooooone commented 12 months ago

@kaiwang960112 请问只要有1个人投票是unknow和没有人脸的就进行去除吗，还是需要达到某个投票数才去除？