PornSites.xxx
PornX
DeepSwap
- Categories:
- Compilation
- Celebrities:
-
Natalie Portman
Natalie Portman800334846241 -
Keira Knightley
Keira Knightley12381474811 -
Scarlett Johansson
Scarlett Johansson12678314622266 -
Anya Taylor-Joy
Anya Taylor-Joy398143318810 -
Kiernan Shipka
Kiernan Shipka20551102464 -
Lili Reinhart
Lili Reinhart534840265
Public version (blurred) of my big model test/report/comparison thingy.
Full version includes:
- DF-UD vs LIAE-UD vs AMP: Fresh vs Reused results (initial training - SRC: Portman, DST: Knightley, reused - SRC: Johansson, DST: Taylor-Joy)
- Additional DF-UD results: Style Power, True Face, GAN Reuse
- Extra fresh run in 2nd SRC/DST scenario for comparison between fresh and reused models (Johansson/Taylor-Joy)
- FF to WF RTM Conversion Concept Test (SRC: Shipka, DST: Reinhart)
- Reverse version of the merges (no training, rtm performance/potential of regular trained models - Portman on Taylor-Joy, Johansson on Knightley)
- Raw merged frames
- Training workflows
- Personal notes and result analysis
- Invite to a telegram group where you can ask me additional questions about model architectures, get personal recommendations regarding the models/architectures/model parameters (dims), review your workflows or get recommendations regarding them, think of this like a limited version of my DFL courses, except really focused on model training workflows, model/architecture choices, parameters, etc. I can also test your models, tell you what's wrong with them, check your RTM datasets or SRC sets, etc.
To get the full report either PM me if you want to pay in crypto (btc, xmr) - 70$, other coins 10% extra or pay in tokens (20% extra so it's 8400 tokens total), then PM me for download link to the 5GB zipped report + invite link to discussion telegram group, if you want the merged frames, you will need premium pixeldrain to download them as they are 140GB or pay extra 10$ and I'll convert them all to videos (you can get up to 3 selected renders for free if you don't need all of them).
Full version includes:
- DF-UD vs LIAE-UD vs AMP: Fresh vs Reused results (initial training - SRC: Portman, DST: Knightley, reused - SRC: Johansson, DST: Taylor-Joy)
- Additional DF-UD results: Style Power, True Face, GAN Reuse
- Extra fresh run in 2nd SRC/DST scenario for comparison between fresh and reused models (Johansson/Taylor-Joy)
- FF to WF RTM Conversion Concept Test (SRC: Shipka, DST: Reinhart)
- Reverse version of the merges (no training, rtm performance/potential of regular trained models - Portman on Taylor-Joy, Johansson on Knightley)
- Raw merged frames
- Training workflows
- Personal notes and result analysis
- Invite to a telegram group where you can ask me additional questions about model architectures, get personal recommendations regarding the models/architectures/model parameters (dims), review your workflows or get recommendations regarding them, think of this like a limited version of my DFL courses, except really focused on model training workflows, model/architecture choices, parameters, etc. I can also test your models, tell you what's wrong with them, check your RTM datasets or SRC sets, etc.
To get the full report either PM me if you want to pay in crypto (btc, xmr) - 70$, other coins 10% extra or pay in tokens (20% extra so it's 8400 tokens total), then PM me for download link to the 5GB zipped report + invite link to discussion telegram group, if you want the merged frames, you will need premium pixeldrain to download them as they are 140GB or pay extra 10$ and I'll convert them all to videos (you can get up to 3 selected renders for free if you don't need all of them).
TMBV
672
2
2468001
![[PMV] [FMV] BlackMamba - Aespa (Karina)](https://mrdeepfakes.com/contents/videos_screenshots/52000/52982/320x180/3.jpg)
![Trans (not) Elyse Willems GFE [Trailer]](https://mrdeepfakes.com/contents/videos_screenshots/42000/42462/320x180/3.jpg)
![(not) Bukkitbrown eats ass & swallows piss [PR][Trailer]](https://mrdeepfakes.com/contents/videos_screenshots/51000/51513/320x180/16.jpg)