MFW when I look up from my GAN at the latest psych paper but then realize it's a Registered Report.pic.twitter.com/7jJw17Ekqk
Можеш да добавиш към туитовете си информация за местоположението (например града ти или точното ти местоположение) през мрежата и чрез приложения на трети страни. Винаги можеш да изтриеш историята на местоположенията на туитовете си. Научи повече
I say they're bad but many are legitimately cool or interesting. Setting psi=1.2 just makes the quality range huge... eg these 4 are striking.
(BTW @EvaMonkey have you been looking at any of these Asuka samples? So lulzy.)pic.twitter.com/BH8MPe9Sn7
More 'Bad Ganime'! Psi=1.2 on the latest face model, 1000 samples + interpolation videos: https://mega.nz/#!2DRDQIjJ!JKQ_DhEXCzeYJXjliUSWRvE-_rfrvWv_cq3pgRuFadw …pic.twitter.com/Nj5Jj65m2v
would it be possible to get the pkl for this, thanks
The Asuka StyleGAN pretrained model:https://mega.nz/#!0JVxHQCD!C7ijBpRWNpcL_gubWFR-GTBDJTW1jXI6ThzSxwaw2aE …
What about 30fps and 3× finer step size?
Hm... It's not clear how 'step size' is done in the video-generating code because it generates a large matrix of latents and then applies a Gaussian filter. Increasing 'smoothing' (filter width?) makes it very blurry. Tried increasing FPS to 69 & decreasing RNG SD to 0.5 in 30s.pic.twitter.com/MtP2jm7WkR
Well, there's an interpolation step used to move among latents, no? Anyhow, your animations look good. I was referring to @roadrunning01's.
There is, but it's a Gaussian filter on a matrix of pre-generated latents which makes them more similar to each other. It's not like 'generate a random z-vector then multiply each element by a random percentage for consecutive frame'. (Seems like an unnecessary optimization TBH).
1D Gaussian blur across the multiple codes?
Twitter може да е претоварен или да изпитва моментно затруднение. Опитай отново или виж Twitter Status за повече информация.