@frosty3907 Yeah but I'm stuck on LowMem, a good lowmem is still shittier than a bad SAE/H128 :/ I don't want to be that guy who everyone hates because he releases daily and his fakes aren'tgood so they saturate the site
Guests - Sign up to view links and videos on the forums. Please check your spam mail. (February 26, 2018)
- New and improved dark forum theme!
- Guests can now comment on videos on the tube.
Shoutbox archive
there's always gonna be people on the net telling you that your work sucks. i'm struggling to get good results with the new models even with a 2080ti
So anyway, I guess I'm stuck on the classic lowmem trainer and people have been telling me my fakes look like shit so I'm thinking of just stopping soon because of all these brand new models coming out where are apparently far superior, and I cannot use them on my quadro p2000. I'd rather not muddy the waters with shitty deepfakes anyway when people are making better higher res stuff because they have better cards, I've had people tell me my fakes are good but I can see the progression to higher quality models and at the same time people are telling me my work isn't great so I think theres probably a correlation there. Anyway Just something I'm thinking about If I can't find a way to train higher resolution models. I don't want to keep disappointing people.
How are the people on the sheet not running out of memory? I've tried copying the settings there and using far lesser and still running out on a 2080ti with 32Gb ram.
Anyoe think private videos should not be shown in the recent/top played? Kinda defeats the purpoe if you can't watch it and nobody is gunna wait for a frined request to go through to bust a nut
Please do not spam the shoutbox with requests. Please post your requests in the request section of these forums. Thank you.
Anyone able to help answer my recent thread? Best possible GPU configuration for DFL? I want to stress test it.
