Open mauropk opened 1 year ago
Hello, thanks for your interest in our work, regarding your questions:
Thank you very much for your clarification. It would be awesome if you could send the weights trained on FiveBillionPixels to my e-mail address: @.***
Thank you again for your time and for your work,
Mauro
Inviato da Outlook per Androidhttps://aka.ms/AAb9ysg
From: x-ytong @.> Sent: Thursday, November 9, 2023 8:28:27 PM To: x-ytong/DPA @.> Cc: Mauro Sylos Labini @.>; Author @.> Subject: Re: [x-ytong/DPA] Info on training time (Issue #4)
Hello, thanks for your interest in our work, regarding your questions:
— Reply to this email directly, view it on GitHubhttps://github.com/x-ytong/DPA/issues/4#issuecomment-1804485081, or unsubscribehttps://github.com/notifications/unsubscribe-auth/BDWYUS4WGRS5MZ5WRSDMSKLYDUVFXAVCNFSM6AAAAAA64HBMT2VHI2DSMVQWIX3LMV43OSLTON2WKQ3PNVWWK3TUHMYTQMBUGQ4DKMBYGE. You are receiving this because you authored the thread.Message ID: @.***>
Thanks for the info, but it looks like github doesn't allow the exchange of email addresses, because from what I can see here, your email is blocked. Could you please send me an email first? My email address can be found on the homepage of the Five-Billion-Pixels dataset.
I need the model trained on Five-Billion-Pixels, too. It would be awesome if you could send the weights trained on FiveBillionPixels to my e-mail address: chz_study@outlook.com.
Thank you very much for sharing. It helped me a lot. Could you send me the model trained on Five-Billion-Pixels as well? My email address is 1098902310@qq.com, thank you again for your help!
Hello, I would like to reproduce your code. It would be awesome if you could send the weights trained on FiveBillionPixels to my email.My email address is 1351674318@qq.com, thank you very much
Hello and thank you for your work! I was thinking to reproduce your findings but i would like some more info on hardware used and training time.
1) How long did it take to train the baseline U-net on the FBP dataset? on what hardware? 2) The model weights published on on the repo are the result of the domain adaptation for each megacity is that correct? Any chance that the weights for Unet trained on the FBP dataset could be released?