Skip to content
New issue

Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.

By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.

Already on GitHub? Sign in to your account

Train with washed up CASIA-WebFace #119

Closed
bamos opened this issue Mar 31, 2016 · 47 comments
Closed

Train with washed up CASIA-WebFace #119

bamos opened this issue Mar 31, 2016 · 47 comments
Labels

Comments

@bamos
Copy link
Collaborator

bamos commented Mar 31, 2016

https://groups.google.com/forum/#!topic/cmu-openface/Xue_D4_mxDQ

@charleshu
Copy link

Hi Basos,
Have you finally trained on this cleaned Casia ?
Unfortunately the link on Baidu of the Washed up list is gone. I am wondering if you can post the list somewhere? or please kindly send me a copy to charleshutx@gmail.com ? I am keen to know how much difference they make. Thank you very much.
Changbo

@allenxcp
Copy link

链接: http://pan.baidu.com/s/1jIqBIcu 密码: eb7h this is the cleaned CASIA-WebFace

@charleshu
Copy link

Thank you very much, Allen. I am downloading. Very slow, though and often being interrupted .  I thought it was a files list; Appear to be images.
I thought VGG triplet tuning must be on their cleaned dataset.  I tried tuning on VGG raw data,   triplet training error going down but test error on LFW keeps going up after a few epochs. It seems not overfitting since only one additional FC layers added;  The only reason I can think of is that the data too noisy.  VGG paper mentioned training on noisy data even better than cleaned data. I guess they referred to classifier training. 
I also tried triplet selection  beta < |n|-|p| <alpha. (following FaceNet). A little bit help but curves remain the same tendency.
Any comments, please let me know.
Thanks,Changbo

On Tuesday, July 19, 2016 2:17 AM, allenxcp <notifications@github.com> wrote:

链接: http://pan.baidu.com/s/1jIqBIcu 密码: eb7h this is the cleaned CASIA-WebFace—
You are receiving this because you commented.
Reply to this email directly, view it on GitHub, or mute the thread.

@xingdi1990
Copy link

Hi, Bamos

Thanks for sharing this cleaned dataset. During the training, I have one question: how do you splits the data into training and validating parts?

@allenxcp
Copy link

allenxcp commented Feb 27, 2017 via email

@sidgan
Copy link

sidgan commented Feb 28, 2017

Is there any working link for the washed CASIA-Webface dataset? All the links mentioned above do not work.

@Honzys
Copy link

Honzys commented Mar 30, 2017

Is there any working link for the washed CASIA-Webface dataset? All the links mentioned above do not work.

@sidgan
Hi, i've found this link: https://pan.baidu.com/s/1kUUP0IN#list/path=%2F, It should work :).

@monktastic
Copy link

monktastic commented Oct 4, 2017

Does anyone know of a link that will work for an English speaker? The Baidu links all seem to require me to download the Baidu downloader, which is exclusively in Chinese.

Edit: The Chrome extension baidudl works.

@stale
Copy link

stale bot commented Dec 3, 2017

This issue has been automatically marked as stale because it has not had recent activity. It will be closed if no further activity occurs. Thank you for your contributions.

@stale stale bot added the stale label Dec 3, 2017
@stale stale bot closed this as completed Dec 10, 2017
@beimingmaster
Copy link

Downloading data from this link(https://pan.baidu.com/s/1kUUP0IN#list/path=%2F), CASIA-maxpy-clean.zip could not decompress correctly, had another link?

@sidgan
Copy link

sidgan commented Dec 20, 2017

See @kihyuks comment on happynear/FaceVerification#30

$ zip -F CASIA-maxpy-clean.zip --out CASIA-maxpy-clean_fix.zip
$ unzip CASIA-maxpy-clean_fix.zip

@beimingmaster
Copy link

beimingmaster commented Dec 21, 2017 via email

@mrwen00
Copy link

mrwen00 commented Dec 24, 2017

thanks very much for solving my problem.

@timho102003
Copy link

@charleshu all the download link here seems dead, can u provide a valid link for Cleaned CASIA Dataset?
Thanks!!!!

@markyugen
Copy link

Are there ANY valid links for the CASIA dataset anymore, dirty or clean?

@Yangel-hide
Copy link

I need CASIA dataset's download link, too!
Thanks!!

@Yangel-hide
Copy link

@sumbara
Copy link

sumbara commented Jul 25, 2018

@Yangel-hide That download link seems dead,too...
I nedd CASIA dataset's downaload link.. Are there valid links for the CASIA? :D

@markyugen
Copy link

I couldn't get that latest link to work for me either. It kept asking for a code and wouldn't accept the correct answer.

@timho102003
Copy link

@sumbara @markyugen You should check if ur country has been block from the website since most of the valid links are from baidu cloud

@Yangel-hide
Copy link

@sumbara @markyugen
the link will be invalid after one week,please download it as soon as possible.
happynear/AMSoftmax#18

@markyugen
Copy link

When I click on download button I get error Sorry, currently only extract files smaller than 4G

@markyugen
Copy link

The other option is to download some kind of app, but when I do so and click on it after it is installed, nothing happens.

@Yangel-hide
Copy link

You should install Baidu cloud client first, and use Baidu cloud client download it.
http://pan.baidu.com/download
I don‘t have google drive link.

@sumbara
Copy link

sumbara commented Jul 25, 2018 via email

@markyugen
Copy link

The Baidu cloud app is waiting for me to type something in. What do I type? I don't read Chinese.

@sumbara
Copy link

sumbara commented Jul 25, 2018

@markyugen You must sing up the Baidu. You can use Chrome and translate that site

@markyugen
Copy link

It won't let me sign up, it says system error try again later!

@Yangel-hide
Copy link

I have uploaded it to google drive, you can download it now

https://drive.google.com/open?id=1Of_EVz-yHV7QVWQGihYfvtny9Ne8qXVz

@markyugen
Copy link

That seems to be working Thank you for all your assistance! I really appreciate it!

@cllyldz
Copy link

cllyldz commented Jul 30, 2018

When I run the command src/align/align_dataset_mtcnn.py , I get this error "No such file or directory: '/home/pi/datasets/casia/CASIA-maxpy-clean/'". How can I figure out this problem?

@cliffworkman
Copy link

cliffworkman commented Sep 14, 2018

try running this: sudo mkdir /home/pi/datasets/casia/CASIA-maxpy-clean/
... and then running the python script.

@CharlesNord
Copy link

I have uploaded it to google drive, you can download it now

https://drive.google.com/open?id=1Of_EVz-yHV7QVWQGihYfvtny9Ne8qXVz

I think this is the original database, not the washed up one

@cliffworkman
Copy link

thanks for this! I managed to track down a copy of the cleaned database in some dark corner of the web -- any future visitors to this page should feel free to contact me should they need a copy of this. thanks again, charles,

@mauxam
Copy link

mauxam commented Jan 6, 2019

Thank you so much!!! CharlesNord

@chenmingjiong
Copy link

chenmingjiong commented Jan 21, 2019

I have uploaded the washed dataset to OneDrive:
[https://1drv.ms/u/s!AjMP8wpvXdkfhMhnmNDEdj5oor7M_A]https://1drv.ms/u/s!AjMP8wpvXdkfhMhnmNDEdj5oor7M_A)
Change 1drv.ms to 1drv.ws if you can't access this link.
(I downloaded it from the baidu link in the first comment.)

@zylo117
Copy link

zylo117 commented Feb 28, 2019

I have uploaded the washed dataset to OneDrive:
https://1drv.ms/u/s!AjMP8wpvXdkfhMhnbzw1noJ6QkTg5A
Change 1drv.ms to 1drv.ws if you can't access this link.
(I downloaded it from the baidu link in the first comment.)

Great Thanks! Life Saver!

@kevinisbest
Copy link

@chenmingjiongjiong
Thanks!!!

@kobaisi
Copy link

kobaisi commented May 17, 2019

I have uploaded the washed dataset to OneDrive:
https://1drv.ms/u/s!AjMP8wpvXdkfhMhnbzw1noJ6QkTg5A
Change 1drv.ms to 1drv.ws if you can't access this link.
(I downloaded it from the baidu link in the first comment.)

The link is dead. Please, can you upload it again? Thanks

@chenmingjiong
Copy link

I have uploaded the washed dataset to OneDrive:
https://1drv.ms/u/s!AjMP8wpvXdkfhMhnbzw1noJ6QkTg5A
Change 1drv.ms to 1drv.ws if you can't access this link.
(I downloaded it from the baidu link in the first comment.)

The link is dead. Please, can you upload it again? Thanks

OK. I have modified the link.

@Coderx7
Copy link

Coderx7 commented Jan 11, 2020

@chenmingjiongjiong The links dead it seems . could you kindly reupload?

@chenmingjiong
Copy link

@chenmingjiongjiong The links dead it seems . could you kindly reupload?

I can open it. Maybe you clicked the wrong link.
https://1drv.ms/u/s!AjMP8wpvXdkfhMhnmNDEdj5oor7M_A

@Coderx7
Copy link

Coderx7 commented Jan 12, 2020

@chenmingjiongjiong Thanks a lot man. yeah it seems I clicked the wrong link. this very link worked just fine :)

@wencoast
Copy link

wencoast commented May 8, 2020

链接: http://pan.baidu.com/s/1jIqBIcu 密码: eb7h this is the cleaned CASIA-WebFace

This link is not available anymore.

@Boatsure
Copy link

Boatsure commented Jun 1, 2020

I have uploaded the washed dataset to OneDrive:
[https://1drv.ms/u/s!AjMP8wpvXdkfhMhnmNDEdj5oor7M_A]https://1drv.ms/u/s!AjMP8wpvXdkfhMhnmNDEdj5oor7M_A)
Change 1drv.ms to 1drv.ws if you can't access this link.
(I downloaded it from the baidu link in the first comment.)

The year 2020, this onedrive link works for me!

@markyugen
Copy link

markyugen commented Jun 1, 2020 via email

@bjzhb666
Copy link

Thanks for your sharing. I am new to casia-webface dataset. I wonder our casia-maxpy-clean dataset. I wonder is our dataset independent of lfw dataset? Thanks!

Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment
Labels
Projects
None yet
Development

No branches or pull requests