Fake ASMR Fakes - Fakes of ASMRtists
- Thread starter BlueStripes
- Start date
-
Welcome to the Fakes / AI / Deepfakes category
When posting AI generated content please include prompts used where possible.
Another Heather image I wanted to add. Didn't try to get this one perfect, just best one from the first batch
uhmmmmm, I tried 
it's Kinda complicated and without many resources you kinda just have to throw shit until it sticks.
I imagine if you load this bitch up with max samples, you could probably get something very decent. But these are just some attempts and they kinda get there but definitely not convincing at all atm. Its going to take a lot more samples and tampering with the settings.
it's Kinda complicated and without many resources you kinda just have to throw shit until it sticks.I imagine if you load this bitch up with max samples, you could probably get something very decent. But these are just some attempts and they kinda get there but definitely not convincing at all atm. Its going to take a lot more samples and tampering with the settings.
My gibi model's ok and it only uses 4 samples, but that's probably because they're around 20 mins each
yeah yours is superb, the complications with Sharon is her accent. When she speaks regularly she is extremely french, but her whispered voice is very subtle. The ai only sometimes whispers depending on the stability and the french accent is the same, so most the time its her whispered voice spoken in a regular tone which is incorrect. Ive only got 5000 words left on my quota and this is the best ive gotten so far...
Not to mention because of the low stability needed to pull something this close of, each revelation can sound completely different, thus it burns through the allotted word quota and ive barely used it. If I had more time to cycle through this I think you can achieve something really good, with more precise samples and fine tuned settings.
yeah I got a resubscription with an extra email, though the generator does not recognise the context, ignores all """ and proceeds to say she whispered instead of well, yknow whispering.
EDIT:
This is an okay attempt, it really comes down to the samples added, I used less samples for this one and concentrated it onto 2 single videos, it turned out as well as you can expect with such a small sample size. I'm probably done with this now until some more documentation comes out since it feels like random chance to get something worth while for me atleast. Maybe someone smarter/committed can take a shot.
Last edited:
It's called Stable Diffusion. There are a couple forums on this site explaining setup and use
Maimy:
Goodnight Moon/Fresh Blush ASMR/Erin Timony
